Professional Documents
Culture Documents
Culture, Leadership, and Organizations The GLOBE Study of 62 Societies by House, Robert J PDF
Culture, Leadership, and Organizations The GLOBE Study of 62 Societies by House, Robert J PDF
62 GLOBE SOCIETIES
1. Albania 32. Japan
2. Argentina 33. Kazakhstan
3. Australia 34. Kuwait
4. Austria 35. Malaysia
5. Bolivia 36. Mexico
6. Brazil 37. Morocco
7. Canada (English speaking) 38. Namibia
8. China 39. The Netherlands
9. Colombia 40. New Zealand
10. Costa Rica 41. Nigeria
11. Czech Republic 42. The Philippines
12. Denmark 43. Poland
13. Ecuador 44. Portugal
14. Egypt 45. Qatar
15. El Salvador 46. Russia
16. England 47. Singapore
17. Finland 48. Slovenia
18. France 49. South Africa (Black Sample)
19. Georgia 50. South Africa (White Sample)
20. Germany-East (former GDR) 51. South Korea
21. Germany-West (former FRG) 52. Spain
22. Greece 53. Sweden
23. Guatemala 54. Switzerland
24. Hong Kong 55. Switzerland (French-Speaking)
25. Hungary 56. Taiwan
26. India 57. Thailand
27. Indonesia 58. Turkey
28. Iran 59. United States
29. Ireland 60. Venezuela
30. Israel 61. Zambia
31. Italy 62. Zimbabwe
FM-House.qxd 3/19/2004 4:05 PM Page iii
Jorbenadse, Revaz, Tbilisi State University (Georgia) Schramm-Nielson, Jette, Copenhagen Business School
Kabasakal, Hayat, Bogazici University (Turkey) (Denmark)
Keating, Mary, University of Dublin—Trinity College Schultz, Majken, Copenhagen Business School (Denmark)
(Ireland) Shi, Zuqi, Fudan University (China)
Keller, Andrea, University of Zurich (Switzerland [French- Sigfrids, Camilla, Swedish School of Economics and
speaking]) Business Administration (Finland)
Kennedy, Jeffrey C., Lincoln University (New Zealand) Song, Kye-Chung, Chungnam National University (South
Kim, Jay S., Ohio State University (South Korea) Korea)
Kipiani, Giorgi, Georgian Academy of Sciences (Georgia) Szabo, Erna, Johannes Kepler University (Austria)
Kipping, Matthias, University of Reading (England) Teo, Albert C., National University of Singapore (Singapore)
Konrad, Edvard, Universiti Ljubjana (Slovenia) Thierry, Henk, University of Tilburg (Netherlands)
Koopman, Paul L., Vrije Univeriteit-Amsterdam Tjakranegara, Jann Hidayat, Institut Technologi Bandung
(Netherlands) (Indonesia)
Kuan, Fuh-Yeong, Shu-Te Institute of Technology (Taiwan) Trimi, Sylvana, University of Nebraska—Lincoln (Albania)
Kurc, Alexandre, University of Nancy 2 (France) Tsui, Anne S., Hong Kong University of Science and
Lacassagne, Marie-Francoise Universite de Bourgogne, Technology (China)
(France) Ubolwanna, Pavakanum, Thammasat University (Thailand)
Lee, Sang M., University of Nebraska—Lincoln (Albania van Wyk, Marius W., University of South Africa (South
and (Kazakhstan) Africa)
Leeds, Christopher, University of Nancy 2 (France) Vondrysova, Marie, University of South Bohemia (Czech
Leguizamon, Francisco, INCAE (Costa Rica) Republic)
Lindell, Martin, Swedish School of Economics and Weibler, Juergen, University of Hagen (Switzerland)
Business Administration (Finland) Wilderom, Celeste, Tilburg University (Netherlands)
Lobell, Jean, AcXEL International (Philippines) Wu, Rongxian, Suzhou University (China)
Luthans, Fred, University of Nebraska—Lincoln (Albania Wunderer, Rolf, University of St. Gallen (Switzerland)
and Kazakhstan) Yakob, Nik Rahiman Nik, Universiti Kebangsaan Malaysia
Maczynski, Jerzy, University of Wroclaw (Poland) (Malaysia)
Mansor, Norma, University of Malaysia (Malaysia) Yang, Yongkang, Fudan University (China)
Martin, Gillian, University of Dublin—Trinity College Yin, Zuoqiu, Fudan University (China)
(Ireland) Yoshida, Michio, Kumamoto University (Japan)
Martin, Michael, University of Nebraska—Lincoln (Albania) Zhou, Jian, Fudan University (China)
Martinez, Sandra M., Widener University (Mexico)
McMillen, Cecilia, University of San Francisco de Quito
(Costa Rica, El Salvador, and Guatemala)
Messallam, Aly, Alexandria University (Egypt) GLOBE SENIOR RESEARCH ASSOCIATES3
Misumi, Emiko, Institute for Group Dynamics (Japan)
Misumi, Jyuji, Institute for Group Dynamics (Japan) Roy-Bhattacharya, Joydeep, Independent Literary Author,
Morsi, Nabil M., Alexandria University (Egypt) USA
Ngin, Phyllisis M., Melbourne Business School (Singapore) Ram Aditya, Florida International University, USA
O’Connell, Jeremiah, Bentley College (Spain) Mary Sully de Luque, Thunderbird University, Arizona, USA
Ogliastri, Enrique, Instituto Centroamericano de Administra- Vipin Gupta, Grand Valley State University, Michigan, USA
cion de Empresas, Costa Rica (Colombia)
Papalexandris, Nancy, Athens University of Economics and
Business (Greece)
Peng, T. K., I-Shou University (Taiwan) END NOTES
Preziosa, Maria Marta, Instituto para el Desrrollo de
Ejecutivos en la Argentina (Argentina) 1. The Globe Coordinating Team served as an advisory
Prieto, Jose M., Complutense University (Spain) group to the Principal Investigator, Robert J. House, with
Rakitsky, Boris V., Institute of Perspectives and Problems respect to policies and procedures for data collection and
of the Country (Russia) administration of GLOBE.
Reber, Gerhard, Johannes Kepler University (Austria) 2. The Country Co-Investigators (CCIs) assisted in the
Rogovsky, Nikolai, International Labor Organization (Russia) formulation of the instruments used in this study, collected
Rozen, Amir, Israel Institute of Technology (Israel) the data relevant to their countries and assisted in interpre-
Sabadin, Argio, Universiti Ljubljana (Slovenia) tation of the data. The countries they represented are in
Sahaba, Majhoub, Groupe EFET (Morocco) parenthesis following their names and their institutional
Siles, Marcelo, Michigan State University (Bolivia) affiliations.
Salgado, Elvira, University of the Andes, Bogota (Colombia) 3. The Senior Research Associates assisted Robert
Salon de Bustamente, Colombia, Universidad de los Andes House in the administration of the GLOBE research
(Venezuela) program, provided advice to Country Co-Investigators,
Santana-Melgoza, Carmen, Smith College (Mexico) maintained GLOBE files and budget, provided status
Sauers, Daniel A, Lincoln University (New Zealand) reports, and performed other administrative functions.
FM-House.qxd 3/19/2004 4:05 PM Page v
FM-House.qxd 3/19/2004 4:05 PM Page vi
All rights reserved. No part of this book may be reproduced or utilized in any form or by any
means, electronic or mechanical, including photocopying, recording, or by any information
storage and retrieval system, without permission in writing from the publisher.
For information:
Sage Publications, Inc.
2455 Teller Road
Thousand Oaks, California 91320
E-mail: [email protected]
Sage Publications Ltd.
1 Oliver’s Yard
55 City Road
London EC1Y 1SP
United Kingdom
Sage Publications India Pvt. Ltd.
B-42, Panchsheel Enclave
Post Box 4109
New Delhi 110 017 India
04 05 06 07 10 9 8 7 6 5 4 3 2 1
CONTENTS
Foreword xv
Harry C. Triandis
Preface xxi
Robert J. House
PART I: Introduction 1
Robert J. House
1. Illustrative Examples of GLOBE Findings 3
Robert J. House
Increased Globalization of Business 4
The Increased Importance of Sensitivity
to Cultural Differences 5
Differing Views and Conceptualizations of Leadership 5
An Illustration of Differences and Cultural Practices 6
Differences in Leadership Practices 7
Plan of This Book 7
2. Overview of GLOBE 9
Robert J. House
Mansour Javidan
Major Questions Concerning the Effects of Cultures
on the Practice of Leadership 9
The Need for Cross-Cultural Leadership Theory and Research 10
The GLOBE Research Program 11
Leader Behavior and Attributes 14
Qualitative Research 15
Construct Definitions of Leadership and Culture 15
GLOBE Operational Definition of Culture 15
The GLOBE Conceptual Model 16
The Integrated Theory 17
Two Fundamental Cross-Cultural Issues 19
The GLOBE Questionnaire Scales 20
Phase 2 Hypotheses 23
Strengths of the GLOBE Research Design 24
Contributions 24
Appendix: Structural Contingency Theory 26
FM-House.qxd 3/19/2004 4:05 PM Page viii
GLOBE Measures 98
Unique Strengths of the GLOBE Research Design 101
FOREWORD
A manifest in act and artifact” (p. vii). This is consistent with the definition used
by the GLOBE research project, which examines culture as practices and
values. Practices are acts or “the way things are done in this culture,” and values are
artifacts because they are human made and, in this specific case, are judgments about
“the way things should be done.”
GLOBE measured practices and values exist at the levels of industry (financial
services, food processing, telecommunications), organization (several in each indus-
try), and society (62 cultures). Thus, the GLOBE researchers measured culture at dif-
ferent levels with both practices and values. Then they asked: How is culture related
to societal, organizational, and leadership effectiveness?
The GLOBE research could be called the Manhattan Project of the study of the rela-
tionship of culture to conceptions of leadership. One hundred and seventy investigators
from 62 cultures worked on this project. Twenty of them participated in writing this
book. They tested 27 hypotheses that linked culture to interesting outcomes, with data
from 17,300 managers in 951 organizations. They measured the variables with cultural
sensitivity, developing instruments in consultation with members of the relevant cul-
tures. By using focus groups, and by heavy dependence on the previous literature, the
investigators developed instruments that tapped local meanings that were appropriate
for each level of the data and also had equivalence across cultures.
Cross-cultural research is tricky and difficult. The researchers took full advantage of
methodologies developed during the past 50 years to do their research in a sophisticated
way. They did pilot tests, they used double translations, they checked the psychometric
characteristics of their instruments, and they eliminated response biases. They used
multiple measurements of the constructs. They checked reliabilities and construct
validity with multitrait multimethod approaches. They started with huge numbers of
items and dropped more than half. They checked their results against the work of
Hofstede (e.g., their Power Distance practices measure correlated with Hofstede’s
Power Distance .61, p < .01, but Power Distance values did not correlate; the Gender
Egalitarianism dimension, on the other hand, had no relation to Hofstede’s Masculinity
dimension), Schwartz (e.g., Uncertainty Avoidance values correlated .74 with Shalom
Schwartz’ Embeddedness), and Inglehart (see below). They measured organizational
and societal culture with different forms of the instrument. They developed unobtrusive
methods for the measurement of some of the constructs. They analyzed the data with
multilevel confirmative factor analysis and hierarchical linear modeling (similar to
regression analysis except that it takes the level—i.e., industry, organization, society—
into account). They addressed the reverse ecological fallacy and identified clusters of
xv
FM-House.qxd 3/19/2004 4:05 PM Page xvi
societies and patterns of outstanding leadership. They then linked the clusters of
societies and the patterns of leadership, and examined how these clusters are related
to interesting health, economic, and social indicators. They also tested how well their
results could be generalized beyond the particular samples of managers, methods, or
kinds of data analyses.
The investigators used an imaginative theoretical framework in which leader accep-
tance and effectiveness were the dependent variables and social culture and organiza-
tional practices were the independent variables. In between these sets of variables they
included variables that may moderate the relationships between the independent and
dependent variables, such as cultural theories of what is a good leader and organiza-
tional contingencies imposed by the fact that the data came from different industries.
The result is an encyclopedia of findings linking culture to societal functioning
and leadership. Table after table give data that contrast cultures on such topics as the
percent of the sample in each culture who dislike democracy to the percentages who
consider a particular leadership attribute desirable. The findings can answer fascinating
questions. In what cultures can a leader violate the societal norms and get away with it?
What leader traits are universally desirable and what traits are undesirable? What is the
ideal of the outstanding leader in Northern Europe or in the Middle East or in eight
other clusters of cultures? What organizations from what countries can be successful
(or unsuccessful) in mergers and acquisitions?
In carrying out this project they broke new ground by developing additional dimen-
sions of societal cultural variation. For instance, they distinguished between institu-
tional and in-group collectivism. The former reflects institutions that encourage
collective rewards, collective distribution of rewards, and collective action.
Institutional collectivism is a cultural pattern that can lead to leadership effectiveness.
In-group collectivism involves pride and loyalty for the organization. It is also related
negatively to divorce rates. They also added new dimensions of organizational culture,
and showed that both practices and values are useful in distinguishing among different
kinds of organizations. They found that organizational cultures reflect societal cultures.
Previous research, such as Hofstede’s 1980 monumental study, had identified
four dimensions of cultural variation—Power Distance, Individualism, Masculinity,
and Uncertainty Avoidance. The GLOBE researchers used Future Orientation, Gender
Equality, Assertiveness, Humane Orientation, In-group Collectivism, Institutional
Collectivism, Performance Orientation, Power Distance, and Uncertainty Avoidance.
Here is a brief description of some of the empirical content of these dimensions.
Performance Orientation is related to high religious diversity and a low concentration
of the largest religion, indicating a culture that is nondogmatic with a creative orienta-
tion. Empirically this attribute of a culture helps economic accomplishments, but people
in such cultures do not live as long as they do in some other cultures.
Assertiveness is related to natural resources such as oil or metals not being important
in the country’s exports. Societies high in this attribute do well in global competitiveness
but exhibit low levels of psychological health.
Future Orientation is related to a large number of trademarks per capita. In countries
high on this attribute people do not visit spontaneously, but call before they visit. Many
people use English as a way to advance in their career. People enjoy economic prosper-
ity, and there is scientific advancement, democracy, gender equality, and social health.
Humane Orientation is related to few retail outlets per capita. People who live in
such cultures tend to use extended, warm greetings. Hospitality is very important.
People show empathy and are very high in satisfaction.
Institutional Collectivism is an attribute that is especially high in Confucian Asia
cultures. These are societies that are less segmented than in other parts of the world.
FM-House.qxd 3/19/2004 4:05 PM Page xvii
Foreword– • –xvii
In-Group Collectivism is related to low divorce rates and poor due process, suggesting
emphasis on the family.
Gender Egalitarianism is related to a high proportion of women earning an income.
Women have access to resources. It is positively correlated with longevity.
Power Distance is related to a limited number of scientists per unit of gross national
product. These are societies in which the rich differ from the poor, and thus economic
growth often results in unemployment and, instead of helping the poor, makes their
position even less satisfactory. Empirically, there is low societal health and little human
development (e.g., education).
Uncertainty Avoidance is related to a high share of home corporations in national
research and development. In such cultures there is an extensive, modern telecommunica-
tion system. There is scientific progress and the government supports economic activities.
One could develop fascinating hypotheses to explain such empirical relationships.
Thousands of doctoral dissertations in the future will start with these findings.
Leadership was studied with six culturally implicit theories of leadership:
Charismatic/Value-based, Team-Oriented, Participative, Autonomous, Humane, and
Self-Protective leadership. In most cultures the first of these is considered most desir-
able. The second is desirable. The other leadership styles, except for the Self-
Protective, are seen as acceptable. But whereas the first is universally desirable (of
course, how could one not like a leader who inspires, is visionary, self-sacrifices, and
is performance oriented?), the other patterns are often culturally contingent. To be
ambitious, for instance, is “good” in some cultures and “bad” in other cultures.
The chapters in this volume contain impressive reviews of the literature corre-
sponding to each of the dimensions of cultural variation. In the case of the collectivism
construct, for instance, the authors traced the development of the concept over the
course of 3,750 years! The findings of the project are presented in each of these dimen-
sion chapters. The findings are too numerous to mention. A few will be selected just to
provide the flavor of what is to come.
The standard literature assumed that societal practices and values are positively
correlated. In this study, often, they were negatively correlated. In short, the managers
indicated that “the way we do things” is negatively related to “what would be the ideal
way of doing things.” This was especially strong in the case of Power Distance. In most
cultures the managers said there was a lot of Power Distance and they would like very
much less Power Distance. If Uncertainty Avoidance was high, the managers indicated
that they would have liked less of it. In most male-oriented cultures many respondents
would have liked less of it. In most cultures there was much collectivism and not
enough emphasis on high performance.
The data suggest that for economic development the managers are correct in want-
ing less Power Distance, less male domination, and more high performance because
that is the pattern that was found in the countries with high gross national product per
capita. The results suggest that movements toward economic and gender equality as
seen in Scandinavia, for instance, can be recommended.
For high prosperity one needs high Performance Orientation, Institutional
Collectivism, and Uncertainty Avoidance, and little Power Distance and In-Group
Collectivism. For human development, such as good health, the culture should be low
in In-Group Collectivism and Power Distance. Life expectancy was positively correlated
with Uncertainty Avoidance practices but negatively correlated with Uncertainty
Avoidance values.
A major focus of the study was the identification of leadership styles associated with
different cultural patterns. Here we find attributes that facilitate (e.g., decisiveness) and
inhibit (e.g., irritable) outstanding leadership. Charismatic/Value-based leadership
FM-House.qxd 3/19/2004 4:05 PM Page xviii
1. We know from the work of Osgood that most concepts are high or low in
evaluation. If we look at the desirable leadership patterns, they consist mostly of
positively evaluated attributes; the undesirable patterns consist mostly of negatively
evaluated attributes.
FM-House.qxd 3/19/2004 4:05 PM Page xix
Foreword– • –xix
2. How can one represent large countries like China, India, or the United States with
a few samples? Would one not need a dozen samples? For India, for instance, would it
not be nice to have samples from Ahmedabad, Mumbai, Delhi, Allahabad, Patna,
Kolkata, Bhubaneswar, Srinagar, Hyderabad, Bangalore, Mysore, and Chennai to pick
up the major cultures and languages of India?
3. Can we believe each statement? For instance, the observation that in Greece
managers have a dim view of “leadership” may depend on the way leadership was
translated. If they used archon, which is an acceptable translation, then the observation
may be due to the fact that this term has the same linguistic root as archondas, a rather
arrogant upper class person.
If one were to go over all the findings with such a fine-toothed comb one could raise
many questions. But one lifetime would not suffice to do that. There is so much here.
In sum, this is a most exciting book for those interested in the relationship of culture
and leadership.
—HARRY C. TRIANDIS
REFERENCES
PREFACE
he idea for GLOBE came to me in the summer of 1991. At that time a substantial
T amount of research had recently been published indicating the validity of the
charismatic leadership paradigm. Leader charisma, as defined by Weber (1947),
House (1976) and Bass (1985), had been found to be predictive of individual satisfaction
in India (Pereira, 1987), mean academic performance scores of educational institutions
in Singapore (Koh, Terborg, & Steers, 1991), performance of U.S. presidents (House,
Spangler, & Woycke, 1991) and profit and cost performance in large grocery stores in the
Netherlands (subsequently published in 1993, Koene, Pennings, & Schreuder, 1993). The
wide variety of cultures in which these studies were conducted suggested to me that
charismatic leader behavior may be universally acceptable and effective.
Consequently, I began thinking about conducting a cross-cultural study based on
data to be collected from approximately 20 culturally diverse societies to test the cross-
cultural generalizability of charismatic leadership in organizations. I began reading the
organizational cultural literature to prepare for such a study. After reading a fair sam-
pling of the organizational cultural literature I concluded that I was learning very little
because the conceptualization and research methodology of the empirical work in this
literature were not very rigorous.
I also read Hofstede’s (1980) book and a small sample of the more general cross-
cultural literature. After reading a number of studies concerning organizational culture
and cross-cultural research, I decided to proceed with the development of the GLOBE
research design without doing an exhaustive literature review. I judged that such a
review would take me at least a year and that my knowledge as a social scientist would
likely be sufficient for a preliminary research design. I also decided that I would
rely on experienced cross-cultural social scientists to advise me concerning unique
methodological problems associated with GLOBE. I spent most of the 1991–1992 aca-
demic year completing outstanding commitments and did some additional reading with
respect to cross-cultural literature.
EVOLUTION OF GLOBE
In the summer of 1992, I was a visiting professor at the Vrie University of Amsterdam
where I worked with Paul Koopman, Henk Thiery, and Celeste Wilderome. I drafted a
preliminary research design and met with these scholars several times. During these
meetings we generated a questionnaire item pool consisting of approximately 300 ques-
tions relevant to Hofstede’s four dimensions of culture plus three additional dimen-
sions: performance and humane and future orientation. During the autumn of 1992, I
was a visiting professor at the European Institute for Business Administration. There
xxi
FM-House.qxd 3/19/2004 4:05 PM Page xxii
I worked with Phillip Podsakoff to further develop the questionnaire item pool. In addition
to items relevant to organizational practices, Phil and I also generated items relevant to
leader attributes and behavior. By the end of that semester, we had developed a total of
approximately 735 items to measure cultural and leadership dimensions.
During the spring semester of 1993, I was a visiting professor at the Departments
of Psychology and Management at the University of Maryland. While there I met Paul
Hanges, who agreed to participate in GLOBE and assume major responsibility for data
storage, maintenance, and analysis. I also met Marcus Dickson, who was to become
Paul’s Research Assistant with respect to GLOBE data maintenance and analysis. Paul
and Marcus were in the Department of Psychology at the University of Maryland.
I also met and recruited Mike Agar and Jeff Thomas. Mike is an anthropologist and
Jeff was a Ph.D. student in management. Jeff served as Mike’s assistant and they jointly
produced guides for anthropological interviewing, media data collection, and content
analysis of interview and media information. These guides were combined to form the
Qualitative Research Manual, which was designed to be a self-instruction guide for
country co-investigators (CCIs). Several CCIs used this manual for guidance when
doing qualitative research.
In the summer of 1993, I was a visiting professor at the University of Victoria, on
Victoria Island in British Columbia, Canada. As of approximately mid-May of 1993,
the GLOBE research design was completed and a questionnaire item pool of more than
700 items had been developed.
In May of 1993, the U.S. Department of Education announced a request for propos-
als under the auspices of the Dwight D. Eisenhower Leadership Education Program.
The Department of Education would fund GLOBE and provided a budget of $300,000.
In 1994 the grant was renewed with an additional funding of $325,000.
In August of 1993, I had begun recruiting CCIs with the objective of obtaining
representation of 20 countries, rather evenly distributed throughout the major regions
of the world.
By October of 1993, scholars from approximately 20 countries had agreed to partic-
ipate. Also, approximately 10 countries had written to me, unsolicited, offering to par-
ticipate. Naturally, I accepted these offers. Encouraged by such enthusiasm for GLOBE
and the grant from the Department of Education, I increased my aspiration to include
40 countries in GLOBE. To ensure representation of least 40 countries in the final sam-
ple, after allowing for attrition, I continued to recruit additional countries until we
reached 65. This number was reached by approximately January 1994.
To provide me with advice I recruited Michael Agar, Marcus Dickson, Paul Hanges,
and Antonio Ruiz Quintanilla. We labeled this group as the U.S. GLOBE Coordinating
Team. The title of the U.S. GLOBE Coordinating Team implied a strong U.S. bias.
Consequently this group and I recruited six additional scholars: Mansour Javidan from
Canada, Peter Dorfman from the U.S., Stephan Akerblom from Sweden, Felix
Brodbeck from Germany, Enrique Ogliastri from Columbia, and Marius Van Wyck
from South Africa. We labeled this group the GLOBE Coordinating Team. This group
met at least once per year. Several problems were discussed and committees were
established to make recommendations with respect to solutions to these problems.
In August of 1994, all CCIs were invited to attend a meeting in Calgary, Canada.
Fifty-four CCIs representing 38 cultures attended the meeting. Problems were dis-
cussed and committee task forces were established to solve these problems. Finally, a
definition of leadership was agreed upon at the Calgary meeting.
Leadership was defined as the ability to motivate, influence, and enable individuals
to contribute to the objectives of organizations of which they are members. Data were
collected from 62 countries during the interval between 1994 and 1997.
FM-House.qxd 3/19/2004 4:05 PM Page xxiii
Preface– • –xxiii
Upon collection of the data from the 62 societies, Paul Hanges and I prepared an
outline for the present book. As the reader can see from the table of contents, the intent
of the book was to be a seamless whole, not an anthology of articles written previously.
It was obvious that writing the book was going to be a huge job, more than two people
could complete in several years. Fortunately, Mansour Javidan, Peter Dorfman with
the aid of his assistant, Renee Brown, and Vipin Gupta volunteered to participate as
coeditors. We formed the GLOBE integrative team consisting of Paul Hanges, Mansour
Javidan, Peter Dorfman, Vipin Gupta, and myself to work closely to create a seamless
book. Over the past three years, the team had several week-long meetings in
Philadelphia to help move things along. We have also had many hour long conference
calls to resolve various issues.
LESSONS LEARNED
One of the major things we learned from the GLOBE experience is that managing a
multination research program of the scope of GLOBE is more than a full-time
endeavor, even with excellent help from office administrators, postdoctoral research
fellows, and research assistants. Throughout 1994–1996, we have had four under-
graduate research assistants who are of tremendous help and fun to be around. In the
summer of 1995, one of the research assistants began counting the number of corre-
spondence items that had been sent by GLOBE to CCIs or others. When she reached
15,000 messages she stopped counting! The 15,000 messages did not include the mes-
sages that were stored on my e-mail hard drive, nor did they include mass mailings or
messages sent since July 1995. Dividing 15,000 by the number of working days it took
to process them yields 33.3 messages per day. On this basis I estimate that since the
beginning of the GLOBE program, we have processed at least 30,000 pieces of corre-
spondence plus mass mailings and network-wide postings on the e-mail net. If we add
the number of CCIs messages sent and include mass mailing and network-wide posting
on the e-mail net, the number would be approximately 46,000 messages in 2 years and
10 months.
We also learned that it is extremely difficult to obtain funding for such a large
program. We have submitted proposals to 27 potential funding agencies or foundations
for Phase 3 of GLOBE and have only recently been successful. It appears that addi-
tional funding will have to be raised with grant proposals with smaller budgets. This is
an extremely time-consuming effort. Each proposal needs to be modified to meet the
requirements of the potential funder.
We also confirmed an earlier conviction: the value of persistence. It was necessary
to be in contact with the CCIs continuously, to monitor their progress, to try to help
them solve problems, to provide encouragement for some, to mildly cajole others. From
1993 through June of 1996, I worked about 12 hours per business day and about 6 hours
on both weekend days when I was not traveling. Some of this time was of course
devoted to teaching and institutional work at Wharton.
Throughout the data collection process it became apparent that several countries
were not able to collect the data. In some cases this was due to lack of interest and in
other cases lack of access to samples. As a consequence, country-investigating teams
from five countries were replaced during the Phase 2 collection period. Three countries
dropped out of GLOBE and were not replaced, bringing the total number of cultures
studied to 62.
Almost all of the teams that dropped out of GLOBE and the teams that were replaced
were headed by consultants or senior professors. The consultants found the research to
FM-House.qxd 3/19/2004 4:05 PM Page xxiv
be too demanding because of their need to continue their normal consulting function.
Several of the senior professors who accepted responsibility for the GLOBE data
collection delegated this responsibility to junior professors. These senior professors
neglected the GLOBE research and consequently their teams were ineffective.
The lesson learned is that it is generally better to recruit less well-established scholars
who have a great deal to gain from participation in such a research project than to recruit
consultants or well-reputed scholars who have less to gain. If senior professors are recruited
it is important to stress the importance of their total involvement in the conduct of the
research and obtain their commitment for such participation as part of the recruiting process.
The original questionnaire items measuring cultural, organizational, and leadership
variables were developed on the basis of Q-sorts, item evaluations, and psychometric
analyses described above. These tasks took longer than expected. Consequently, the
questionnaire scales were not ready for use at the time for which they were initially
scheduled. Further, almost all of the country co-investigators found it much more
difficult to gain access to samples than was originally expected. Consequently the schedule
for data collection was changed several times. As a result, country co-investigators
from two countries had obtained commitment dates from participating organizations for
scale administration that were earlier than the time at which the scales were available.
This caused embarrassment for some of the country co-investigators and concern about
whether GLOBE would be successfully completed. The remaining teams were not
embarrassed because they had not arranged for data collection to take place prior to the
availability of the questionnaires.
The research design also called for the development of scales to measure strategic
organizational contingency variables (size, strategy, technology, and organizational
environment) as well as scales to measure unobtrusive measures and participant obser-
vations relevant to the core GLOBE dimensions. The development of these scales was
not completed until December of 1995, approximately 6 months later than originally
planned. This delay was due to the fact that the scales to measure unobtrusive mea-
sures and participant observations were much more difficult to develop than originally
anticipated. The lesson learned is that cross-cultural research takes much longer than
single-culture research and should be planned accordingly.
Another lesson learned is that it is important to maintain schedule flexibility and to
explain the need for schedule changes clearly to all involved. It is important to develop
among participating investigators a high tolerance for uncertainty because the need for
schedule changes will be inevitable. Tolerating schedule changes was quite difficult for
many of the countries, especially those with low scores on Hofstede’s tolerance of
uncertainty measures. Some CCIs rather strongly urged me to terminate relationships
with CCI teams that were behind schedule. Had we done this there would be no
GLOBE data at this time because only two countries met the original schedule dead-
lines. As will be explained below, this was due to many factors associated with the
administration of GLOBE and not due to delinquency of CCIs.
It was originally planned to have four phases for the entire GLOBE research
program. When it became apparent that we would not be funded for Phase 3, I rec-
ommended to the GLOBE Coordinating Committee that we integrate many of Phase
2 and Phase 3 research questions into Phase 2 in order to have a coherent research
product and to take advantage of what might be accomplished with existing funds,
given the momentum that had developed. The committee agreed. This change
involved collecting some, but not much, additional data relevant to organizational
demography and strategic contingencies. Some CCIs objected to this, reflecting either
a low tolerance for change or a belief that it was unfair to ask them to put forth
additional effort. The additional effort would be only marginally more than what was
FM-House.qxd 3/19/2004 4:05 PM Page xxv
Preface– • –xxv
CONCLUSION
We have a very adequate dataset to replicate Hofstede’s (1980) landmark study and
extend that study to test hypotheses relevant to relationships among societal-level
variables, organizational practices, and leader attributes and behavior. We also have
sufficient data to replicate middle-management perceptions and unobtrusive measures.
The original data collection target was December of l995. Data collection was final-
ized about August of 1997. The data collection would have gone much more quickly,
smoothly, and efficiently had the scales been developed earlier and had all country
co-investigating teams been able to access samples earlier. However, to have accom-
plished such smooth and efficient data collection would have required a much longer
schedule and would have delayed final data collection. Even if we had established a
longer schedule we would certainly have met unanticipated problems.
My final conclusion is that we are in a position to make a major contribution to
the organizational behavior and leadership literature. To date more than 90% of the
organizational behavior literature reflects U.S.-based research and theory. Hopefully
GLOBE will be able to liberate organizational behavior from the U.S. hegemony.
All of the members of the editorial team, including Mary Sully de Luque, contributed
to the final chapter of this book. However, Mansour Javidan wrote the final chapter
that was reviewed by all members of the editorial team. The members of the editorial
team worked long and hard for the last three years. Among other tasks, Peter Dorfman,
with the aid of Renee Brown, performed copyediting duties; checked for continuity of
terminology, data, and content; and ensured data consistency among chapters for
conformity of presentation. Renee Brown and Mary Sully de Luque created the
mechanism to keep track of chapter drafts, no small task given the multiple editing
of each chapter by the editorial team members.
Peter Dorfman was responsible for reading the entire manuscript and ensuring that
the general material on leadership processes was consistent across chapters. He
worked to achieve uniformity and accuracy for data analysis and presentation for
the culturally-endorsed leadership profiles. Peter also worked with Paul Hanges to
develop the template for presentation of statistical analyses regarding the relationship
of societal-cultural dimensions and leadership dimensions. Peter also developed the
format for presenting these relationships in figure form used throughout the text. Both
Renee and Peter regularly conferred with other editorial team members to assure that
the revision of chapters followed a regular sequence to minimize or eliminate errors
regarding text changes, omissions, and deletions.
FM-House.qxd 3/19/2004 4:05 PM Page xxvi
Paul Hanges had the prime responsibility for developing the GLOBE scales. This
involved everything from introducing the frame of reference in the organizational and
societal culture scale items to analyzing the item evaluation reports conducted by CCIs
and reviewing the translation-back translation reports, to planning and conducting the
statistical analyses that were used to verify the psychometric properties of the GLOBE
scales as well as their validity. He also was responsible for building the GLOBE data-
base and for evaluating and determining the highly sophisticated statistical approach
that we would use to test the GLOBE hypotheses. When current statistical procedures,
such as the traditional response bias correction procedure, didn’t fully address our
questions, we relied on Paul to identify or develop new approaches that could more
adequately address the goals of GLOBE. Paul was successful in this regard. See
Appendix B: Response Bias Correction Procedure Used in GLOBE.
The tangible product of his work can be seen in the quality of the GLOBE scales and
the statistical analyses conducted to test the GLOBE hypotheses. While his statistical con-
tributions are very clear, it may not be also clear that his contributions extended beyond
statistics. He helped distribute information among the GLOBE community by proposing,
developing, and maintaining the first GLOBE Web site. The identification of universal
leadership attributes were also a function of Paul’s work on GLOBE and these attributes
were really a side product of his efforts developing the leadership scales, referred to as
CLT scales throughout this book. His contributions also extend to the intellectual discus-
sions that we had regarding the meaning of the GLOBE scales and whether they had
meaning beyond the opinions of the middle managers who completed the scales.
Mansour Javidan was the editor of all the chapters on cultural dimensions. He wrote
the Performance Orientation chapter which became the model for all dimension chapters.
He then worked very closely with the writers of each dimension chapter
to make sure that the conceptualization, literature review, and interpretation of the
findings were comprehensive and consistent across all dimension chapters. He went
through up to five revisions with the writers of these chapters. The result is that Part 4 of
the book, even though written by a large number of authors from different parts of the
world, is a seamless and consistent piece. Mansour read the entire manuscript to make
sure any connections between dimension chapters and others are properly addressed in
all chapters of the book. He further designed the external dataset. He first developed a
series of hypotheses to link external data with GLOBE cultural dimensions and worked
with Markus Hauser to develop a large database of external data from various sources
which were directly relevant to GLOBE. The database was made available to the authors
of dimension chapters who used it to link their specific cultural dimensions to external
data. As a result, we extended our original theoretical model to incorporate the relation-
ship between culture and societal achievement, and performed an extensive empirical
investigation of that relationship. Mansour has also played a helpful managerial role to
ensure the progress of our book. Over the past three years, he helped the editorial team
coordinate its activities and progress. He also has made many presentations and keynote
speeches on GLOBE at a large number of conferences all over the world.
Vipin Gupta’s rich knowledge of international and strategic management, and
valuable research methodology skills have been great assets to the GLOBE Project.
Vipin conducted extensive multidisciplinary literature searches finding theoretical and
empirical evidence regarding GLOBE dimensions. Along with Paul Hanges, he formu-
lated regional and climatic clusters of the societal cultures. He also provided data analy-
sis for the sections on regional and climate clustering and on organizational
culture scores. Vipin reviewed all of the empirical chapters to verify the statistical
analyses and provide additional discussion regarding interpretations of the data. He also
formulated and operationalized, with the help of Mary Sully de Luque, the unobtrusive
FM-House.qxd 3/19/2004 4:05 PM Page xxvii
Preface– • –xxvii
and outcropping validation measures for the GLOBE societal cultural scales. He
painstakingly checked the consistency of reported findings across different chapters,
and between tables and text.
We would like to give special thanks to Mary Sully de Luque, who devoted tireless
effort to this project during her 3 years as a Wharton Senior Fellow, despite being a
loving and caring mother to a new born baby, Devin, who was born three months after
Mary joined GLOBE. Mary, having worked for two U.S. senators, brought to bear a
substantial amount of administrative ability as well as her scholarly contributions.
During her tenure, Mary took charge of assembling the multiple parts of the GLOBE
phase 3 project, corresponding with all Country Co-Investigators, facilitating the back-
translation process, developing significant parts of the databases, analyzing preliminary
data, evaluating appropriate statistical procedures, and generally maintaining the project.
Additionally, and no less importantly, Mary became an intricate part of this book. She
coauthored the measurement validation chapter, working countless hours on the content
analysis required of this type of research. Her contribution as lead author on the
Uncertainty Avoidance chapter was invaluable. In the final year of the project, Mary
became the curator of the book chapters; a role Peter Dorfman suggested to sustain
continuity in the process of multiple editing of the chapters. In retrospect, the success of
any project of this magnitude requires the addition of this type of process. Clearly, Mary
worked long and hard on this book and went far beyond the normal expectations of her
position as a Post Doctoral Research Fellow. Without her participation, both scholarly
and administrative, the completion of this book would not have been possible.
ACKNOWLEDGMENTS
We would like to thank many people who directly or indirectly contributed to devel-
opment and completion of this book. Foremost, we would like to thank our spouses
Tessa House, Carol Shouvlin Hanges, Soheila Yazdanbakhsh, Sharon Dorfman, Bhakti
Gupta, and Edgard Luque for their patience and perseverance with this project. The
support they provided was invaluable, especially through the final few years of the writ-
ing of this book. The first author would like to especially thank Joseph Frank Bernstein
for continued financial support, and Donald Bigelow, of the Dwight D. Eisenhower
Research Program at the U.S. Department of Education, and the National Science
Foundation for grant funding of this research.
Special acknowledgments for organizations that assisted in this project go to
University of Pennsylvania; University of Maryland; the University of Calgary; New
Mexico State University; Grand Valley State University; Thunderbird, the American
Graduate School of International Management; and the Industrial Management
Institute of Iran. Specifically, we would like to thank the Department of Psychology
at the University of Maryland for providing computer support for this project.
A project of this magnitude requires the participation and support of a multitude of
people. We respectfully acknowledge the contribution of the many book coauthors and
country co-investigators and the support of their institutions in providing the resources
to accomplish this research. We would like to thank Michael Bond for providing his
scholarly expertise and Habir Singh for his unwavering academic support. Numerous
people contributed during various phases of this project; of these we would like to
acknowledge the academic assistance of Harry Triandis; Major Paul Bliese; Katherine
Klein; David Saunders, former Dean of Management at the University of Calgary; as
well as the research assistance of James Zale, Gavaskar Balasingam, Louani Bascara,
Bhairvee (Ravi) Shavdia, Karen Toll, Narda Quigley, Sandra Martinez, and Gian
FM-House.qxd 3/19/2004 4:05 PM Page xxviii
Casimir. We would also like to thank Michael Grojean, Ellen Godfrey, Mina Sipe,
Beng-Chong Lim, Lisa Nishii, and William Oakley for helping update and maintain the
GLOBE database over the years. In addition, we would like to thank Marcus Dickson
for all of his assistance during the many years of the GLOBE project. Special thanks go
to Geri Grosso, Renee Brown, and Angela Smith who provided administrative assistance
essential for this project.
We are also grateful to MaryAnn Vail and Diane Foster for their patience and
assistance in bringing this book to fruition.
—ROBERT J. HOUSE
REFERENCES
Bass, B. M. (1985). Leadership and performance beyond expectations. New York: Free Press.
Hofstede, G. (1980). Culture’s consequences: International differences in work-related values.
Beverly Hills, CA: Sage.
House, R. J. (1976). A 1976 theory of charismatic leadership. In J. Hunt & L. Larson (Eds.),
Leadership: The cutting edge (199–272). Carbondale: Southern Illinois University Press.
House, R. J., Spangler, D., & Woycke, J. (1991). Personality and charisma in the U.S. presidency:
A psychological theory of leadership effectiveness. Administrative Science Quarterly, 36,
364–396.
Koene, H., Pennings, H., & Schreuder, M. (1993). Leadership, culture, and organizational effec-
tiveness. In K. E. Clark, M. E. Clark, & D. P. Campbell (Eds.), The impact of leadership
(pp. 230–241). Greensboro, NC: The Center for Creative Leadership.
Koh, W. L., Terborg, J. R., & Steers, R. M. (1991, August). The impact of transformational
leaders on organizational commitment, organizational citizenship behavior, teacher satis-
faction and student performance in Singapore. Paper presented at the meeting of the
Academy of Management, Miami.
Pereira, D. (1987). Factors associated with transformational leadership in an Indian engineering
firm. Paper presented at the meeting of the Administrative Science Association of Canada,
Vancouver.
Weber, M. (1947). The theory of social and economic organizations (T. Parsons, Trans.)
New York: Free Press. (Original work published 1924)
01-House.qxd 3/19/2004 4:06 PM Page 1
PART I
INTRODUCTION
ROBERT J. HOUSE
t the present time there is a greater need extremely positive connotation associated
1
01-House.qxd 3/19/2004 4:06 PM Page 2
1
ILLUSTRATIVE EXAMPLES
OF GLOBE FINDINGS
ROBERT J. HOUSE
he 21st century may very well become in perceptions of effective and ineffective leader
3
01-House.qxd 3/19/2004 4:06 PM Page 4
4– • –INTRODUCTION
6– • –INTRODUCTION
8– • –INTRODUCTION
Gregersen, H. B., Morrison, A. J., & Black, J. S. June 25, 2003 from http://www.oecd.org/pdf/
(1998). Developing leaders for the global M00042000/M00042212.pdf
frontier. Sloan Management Review, 40(1) The Bank of New York. (2002). Depositary receipts
21–32. (ADRs and GDRs): 2001 year-end market review.
House, R. J., Wright, N., & Aditya, R. N. (1997). New York: The Bank of New York.
Cross cultural research on organizational leader- The Bank of New York. (2003, January). The ADR
ship: A critical analysis and a proposed theory. investor. New York: The Bank of New York.
In P. C. Earley & M. Erez (Eds.), New perspec- The world’s 500 largest industrial corporations. (2002,
tives on international industrial/organizational July 22). Fortune.
psychology (pp. 535–625). San Francisco, CA: McFarland, L. J., Senen, S., & Childress, J. R. (1993).
New Lexington Press. Twenty-first-century leadership. New York:
Organisation for Economic Co-operation and Leadership Press.
Development. (2003). Trends and recent devel- World Trade Organization. (2003, June). WTO World
opments in foreign direct investment. Retrieved tourism barometer, 1(1), 1–14.
02-House.qxd 3/19/2004 4:07 PM Page 9
2
OVERVIEW OF GLOBE
ROBERT J. HOUSE
MANSOUR JAVIDAN
9
02-House.qxd 3/19/2004 4:07 PM Page 10
10– • –INTRODUCTION
progress has been made in several areas (see The 21st century may very well become known
Dorfman, 2004; House, Wright, & Aditya, 1997 as the century of the “global world” (McFarland,
for an extensive review of relevant leadership Senen, & Childress, 1993). Although there is
literature). some controversy with respect to the effects
GLOBE is a programmatic research effort of leaders on organizational effectiveness, we
designed to explore the fascinating and complex believe that effective organizational leadership
effects of culture on leadership, organizational is critical to the success of international opera-
effectiveness, economic competitiveness of tions. We also believe that the amount of influ-
societies, and the human condition of members ence, prestige, and privilege given to leaders
of the societies studied. To address these issues varies widely by culture. In some cultures, there
we conducted an extensive quantitative and are severe constraints on what leaders can and
qualitative study of 62 cultures. The specific cannot do. In other cultures, leaders are granted
cultures studied will be described below. a substantial amount of power over followers
Specific objectives of GLOBE include answe- and are given special privileges and high status.
ring the following fundamental questions: Globalization of industrial organizations
presents numerous organizational and leader-
1. Are there leader behaviors, attributes, and orga- ship challenges. They include the design of
nizational practices that are universally accepted multinational organizational structures, the iden-
and effective across cultures? tification and selection of leaders appropriate
to the cultures in which they will be functioning,
2. Are there leader behaviors, attributes, and organi-
the management of organizations with cultur-
zational practices that are accepted and effective
ally diverse employees, as well as cross-border
in only some cultures?
negotiations, sales, and mergers and acquisitions.
3. How do attributes of societal and organiza- Unfortunately, the literature provides little in the
tional cultures influence whether specific leader way of guidance for leaders facing these chal-
behaviors will be accepted and effective? lenges (House, Wright, & Aditya, 1997; House &
Aditya, 1997). Cross-cultural research and the
4. How do attributes of societal and organizational
development of cross-cultural theory are needed
cultures affect selected organizational practices?
to fill this knowledge gap.
5. How do attributes of societal cultures affect the There are compelling reasons for considering
economic, physical and psychological welfare the role of societal and organizational culture
of members of the societies studied? in influencing leadership and organizational
processes. What we need are theories of leader-
6. What is the relationship between societal cultural
ship and organizations that transcend cultures.
variables and international competitiveness of the
However, there are inherent limitations in
societies studied?
applying theories across widely varying cul-
tures. What functions effectively in one culture
may not in another culture. Cross-cultural
THE NEED FOR CROSS-CULTURAL research on leadership and organizations will
LEADERSHIP THEORY AND RESEARCH help us test our knowledge in other cultures,
identify boundary conditions for our theories,
Given the increasing globalization of industrial fine-tune existing theories by incorporating
organizations and the growing interdepen- cultural variables, and identify potentially uni-
dencies among nations, the need for a better versal aspects of leadership (Berry & Dason,
understanding of cultural influences on leader- 1974; Dorfman, 2004).
ship and organizational practices has never Although research literature relevant to the
been greater. Leaders confront situations that above questions has increased substantially in
are highly complex, constantly evolving, and the past 15 years, it is often atheoretical, fraught
difficult to interpret. Managers of global firms with methodological problems, and fragmented
are facing unprecedented fierce and rapidly across a wide variety of publication outlets.
changing competition. More questions than answers persist regarding
02-House.qxd 3/19/2004 4:07 PM Page 11
the culturally contingent aspects of leadership. 100 articles and book chapters have been written
Project GLOBE is designed to contribute theo- based on the GLOBE data that have been
retical developments and empirical findings to collected to date.
fill this knowledge deficiency.
Independent Variables
THE GLOBE RESEARCH PROGRAM The first significant question addressed by
Project GLOBE concerns the differentiating
GLOBE is a worldwide, multiphase, multi- attributes of societal and organizational cultures.
method project intended to answer the questions To address this issue we developed 735 ques-
raised above. The concept of a global research tionnaire items on the basis of prior literature
program concerned with leadership and organi- and our own theorizing. Responses to these ques-
zation practices was conceived in the summer tions by middle managers in two pilot studies
of 1991. In the spring of 1993 a grant proposal were analyzed by conventional psychometric
was written that followed a substantial literature procedures (e.g., item analysis, factor analysis,
review and development of a pool of 735 ques- cluster analysis, generalizability analysis).
tionnaire items. GLOBE was funded in October These analyses resulted in the identification of
1993, and the recruiting of GLOBE country nine major attributes of cultures and six global
co-investigators (referred to hereafter as CCIs) leader behaviors of culturally endorsed implicit
to collect data in 62 societies began.1 One hun- theories of leadership (referred to as CLTs
dred seventy social scientists and management throughout the remainder of this book). These
scholars from 62 cultures representing all major analyses are reported in greater detail in Chapter
regions of the world are currently engaged 8 by Hanges and Dickson.
in this long-term programmatic series of cross- When quantified, these cultural attributes are
cultural studies. Table 2.1 lists the countries referred to as cultural dimensions and serve as
in which cultures were studied as part of the the independent variables of Project GLOBE.
GLOBE research reported in this book. The nine cultural dimensions we identified as
The CCIs, together with the principal inves- independent variables are Uncertainty Avoidance,
tigators and research associates, comprise the Power Distance, Institutional Collectivism, In-
members of the GLOBE community. The CCIs Group Collectivism, Gender Egalitarianism,
are responsible for leadership of the project in Assertiveness, Future Orientation, Performance
a specific culture in which they have expertise. Orientation, and Humane Orientation. We devel-
Their activities include collecting quantitative oped original scales for each of these dimensions.
and qualitative data, ensuring the accuracy of There were two forms of questions for each
questionnaire translations, writing country- dimension. The first of these two forms measured
specific descriptions of their cultures, inter- managerial reports of actual practices in their
preting the results of quantitative data relevant organization and managerial reports of what
to their culture, and contributing insights from shoule be (values) in their organization. The sec-
their unique cultural perspectives to the ongo- ond form measured managerial reports of prac-
ing GLOBE research. In most cases, CCIs are tices and values in their societies. Thus we have
natives of the cultures from which they are col- 18 scales to measure the practices and values
lecting data and reside in that culture. Some of with respect to the core GLOBE dimensions of
the CCIs have extensive experience in more culture. A description of the two forms, and the
than one culture. Most cultures have a research levels of analysis at which they are applied, is
team of between two and five CCIs working on presented below.
the project. Following are the definitions of the core
The initial activities of the project as a whole GLOBE cultural dimensions.
were coordinated by the GLOBE coordinating Uncertainty Avoidance is the extent to which
team (GCT), which consists of members from members of an organization or society strive
a variety of cultures. The members of this team to avoid uncertainty by relying on established
are listed in the front of this book. More than social norms, rituals, and bureaucratic practices.
02-House.qxd 3/19/2004 4:07 PM Page 12
12– • –INTRODUCTION
14– • –INTRODUCTION
Our substitution avoids the confusion and (a) visionary, (b) inspirational, (c) self-sacrifice,
interpretation difficulties of Hofstede’s measure (d) integrity, (e) decisive and (f) performance
of masculinity. oriented.
16– • –INTRODUCTION
shared organizational values, and organizational 1974). The key features of the first three theories
history. For purposes of GLOBE research, cul- are briefly outlined below. The integrated theory
ture is operationally defined by the use of indi- is then described. The acronym for GLOBE
cators reflecting two distinct kinds of cultural Culturally Endorsed Implicit Theory of Leader-
manifestations: (a) the commonality (agreement) ship is CLT and this acronym will be used
among members of collectives with respect to throughout the book. Because tests of structural
the psychological attributes specified above; and contingency theory are not reported in this book,
(b) the commonality of observed and reported this theory is not described here. For a more
practices of entities such as families, schools, detailed description of the integrated theory, see
work organizations, economic and legal systems, House, Wright, and Aditya (1997).
and political institutions.
The core cultural dimensions specified above
Implicit Leadership Theory
are measured in terms of two manifestations of
cultures: modal practices and modal values of According to this theory individuals have
collectives. Modal practices are measured by the implicit beliefs, convictions, and assumptions
responses of middle managers to questionnaire concerning attributes and behaviors that distin-
items concerning “What Is,” or “What Are,” guish leaders from followers, effective leaders
common behaviors, institutional practices, pro- from ineffective leaders, and moral leaders
scriptions, and prescriptions. This approach to from evil leaders. These beliefs, convictions,
the assessment of culture grows out of a psycho- and assumptions are referred to as individual
logical and behavioral tradition in which it is implicit theories of leadership.
assumed that cultures should be studied as they It is believed that implicit leadership theories
are interpreted by their members (Segall, Lonner, held by individuals influence the way they view
& Berry, 1998) and that shared values are the importance of leadership, the values they
enacted in behaviors, policies, and practices. attribute to leadership, and the values they place
Values are expressed in response to ques- on selected leader behaviors and attributes. The
tionnaire items concerning judgments of “What following propositions express the major asser-
Should Be,” which are intended as a measure of tions of implicit leadership theory.
the respondents’ values concerning the practices
reported by the respondents. These may be 1. Leadership qualities are attributed to individuals,
referred to as contextualized values as opposed and those individuals are accepted as leaders on
to more abstract values such as values concern- the basis of the degree of congruence between
ing justice; independence; freedom; and a world the leader behaviors they enact and the implicit
of order, beauty, and peace. Emphasis on values leadership theory held by the attributers.
grows out of an anthropological tradition of
culture assessment (Kluckhohn & Strodtbeck, 2. Implicit leadership theories constrain, moder-
1961). GLOBE researchers tested the assumptions ate, and guide the exercise of leadership, the
presented above and the results are presented in acceptance of leaders, the perception of leaders
Chapters 12 through 19. as influential, acceptable, and effective, and the
degree to which leaders are granted status and
privileges.
THE GLOBE CONCEPTUAL MODEL
Substantial experimental evidence supports
The theory that guides the GLOBE research this theory (Hanges, Braverman, & Rentsch, 1991;
program is an integration of implicit leadership Hanges et al., 1997; Lord & Maher, 1991; Sipe &
theory (Lord & Maher, 1991), value–belief Hanges, 1997). A major part of the GLOBE
theory of culture (Hofstede, 1980; Triandis, research program is designed to capture the CLTs
1995), implicit motivation theory (McClelland, of each society studied. We found that if aggre-
1985), and structural contingency theory of orga- gated to the societal level of analysis, responses
nizational form and effectiveness (Donaldson, to the leadership questionnaire reflect the cultur-
1993; Hickson, Hinings, McMillan, & Schwitter, ally endorsed implicit theory of leadership of the
02-House.qxd 3/19/2004 4:07 PM Page 17
societies studied. The evidence for this is that We will test the assumption that middle
there is a high and significant within-society manager reports of explicit motives, if aggregated
agreement with respect to questions concerning to the societal or organizational level, function in
the effectiveness of leader attributes and behavior. the same way that McClelland asserts implicit
Further, aggregated leadership scores were signif- motives function. Whereas McClelland’s theory
icantly different among the societies studied. is an individual theory of both nonconscious and
Thus, each society studied was found to have a conscious motivation, the GLOBE theory is a
unique profile with respect to the culturally theory of motivation resulting from cultural
endorsed implicit theory of leadership. forces. Cross-cultural tests of some of the major
aspects of the integrated theory are reported in
Chapters 12 through 21 of this volume.
Value–Belief Theory
According to value–belief theory (Hofstede,
2001; Triandis, 1995), the values and beliefs held THE INTEGRATED THEORY
by members of cultures influence the degree to
which the behaviors of individuals, groups, and The central proposition of the integrated theory
institutions within cultures are enacted, and the is that the attributes and entities that differen-
degree to which they are viewed as legitimate, tiate a specified culture are predictive of organi-
acceptable, and effective. Collectively, the core zational practices and leader attributes and
GLOBE cultural dimensions described earlier behaviors that are most frequently enacted and
reflect not only the dimensions of Hofstede’s most effective in that culture.
and Trandis’s theories but also McClelland’s The integrated theory also consists of the
theory of human motivation and economic following propositions, which are shown in the
development (McClelland, 1985). The humane, system diagram in Figure 2.1:
power distance, and performance orientation
of cultures are conceptually analogous to the 1. Societal cultural norms of shared values and
affiliative, power, and achievement motives in practices affect leaders’ behavior. Substantial
McClelland’s theory of human motivation. empirical evidence supports this assertion
Implicit motivation theory is a theory (Dorfman, 2004; House, Wright, & Aditya,
of nonconscious motives originally advanced 1997; Schein, 1992; Schneider, 1987; Schneider,
by McClelland, Atkinson, Clark, and Lowell Goldstein, & Smith, 1995). First, founders and
(1953). In its most general form, the theory original members of organizations are immersed
asserts that the essential nature of long-term in their own societal culture. Consequently their
and complex human motivation can be under- leader behavior and management practices are
stood in terms of three implicit (nonconscious) likely to reflect behavior patterns favored in that
motives: achievement, affiliation, and power culture. Subordinates of the original founder of
(social influence). This theory also identifies an organization and subsequent leaders also use
three explicit (conscious) motives related to management practices that reflect the values
achievement, affiliation, and power that are shared by members of the culture. For example,
predictive of short-term, noncomplex behav- founders and subsequent leaders establish selec-
ior. In contrast to behavioral intentions and tion criteria for hiring and promotions, serve as
conscious values, which are predictive of dis- role models by setting personal examples, and
crete task behaviors for short periods of time socialize organizational members in a manner
under constant situational forces (Ajzen & that reflects the broader culture in which they
Fishbein, 1970), implicit motives are predictive function. Further, dominant cultural norms
of (a) motive arousal in the presence of particu- induce global leader behavior patterns and orga-
lar stimuli; (b) spontaneous behavior in the nizational practices that are expected and viewed
absence of motive-arousal stimuli; and (c) long- as legitimate.
term individual global behavior patterns, such as The attributes and behaviors of leaders are
social relationship patterns, citizenship behavior, also a reflection of prevailing organizational
child-rearing practices, and leadership styles. practices in their industries. Industrial practices
02-House.qxd 3/19/2004 4:07 PM Page 18
18– • –INTRODUCTION
14 Economic Performance
of Societies
15
Physical & Psychological
Well Being of Societies
5
Societal Culture,
1 Culturally Endorsed Implicit
Norms & Practices
6 Leadership Theory (CLT)
3
2
Organizational Form, Leader Attributes (10)
Culture & Practices & Behaviors Leader Acceptance
4
(9) 13 12
8
7 Leader Effectiveness
(11)
Strategic Organizational
Contingencies
are to a nontrivial extent indicative of societal 5,6. Societal culture and organizational culture
cultural dimensions (e.g., Kopelman, Brief, & and practices influence the process by which
Guzzo, 1990). people come to share implicit theories of leader-
ship. Over time, CLT profiles are developed in
2. Leadership affects organizational form, culture, each culture in response to both societal and orga-
and practice. Founders and subsequent leaders nizational culture and practices. CLT profiles are
continue to influence and maintain the organi- composed of the six CLT leadership dimensions.
zational culture (Bass, 1985; Miller & Droge,
1986; Schein, 1992; Thompson & Luthans, 1990; 7. Strategic organizational contingencies (orga-
Yukl, 2002). nizational environment, size, and technology)
affect organizational form, culture, and practices.
3. Societal cultural values and practices also
Organizational practices are largely directed
affect organizational culture and practices.
toward meeting the requirements imposed by
Although societal culture has a direct impact on
organizational contingencies (Burns & Stalker,
organizational culture, over time, organizational
1961; Donaldson, 1993; Lawrence & Lorsch, 1967;
cultures influence the broader societal culture.
Tushman, Newman, & Nadler, 1988).
Collective meaning that results from the dominant
cultural values, beliefs, assumptions, and implicit
8. Strategic organizational contingencies affect
motives endorsed by societal culture results in
leader attributes and behavior. Leaders are
common leadership and implicit organization
selected to meet the requirements of organiza-
theories held by members of the culture (House,
tional contingencies. They will then continue to
Wright, & Aditya, 1997; Lord & Maher, 1991).
adjust behavior on the basis of these contingencies.
4. Organizational culture and practices also affect
leaders’ behavior. Over time, founders and sub- 9. Relationships between strategic organiza-
sequent leaders in organizations respond to the tional contingencies and organizational form,
organizational culture and alter their behaviors culture, and practices will be moderated by
and leader styles accordingly (Lombardo, 1983; cultural forces. For example, in low uncertainty
Schein, 1992; Trice & Beyer, 1984). avoidance cultures we expect that forces toward
02-House.qxd 3/19/2004 4:07 PM Page 19
formalization will be weaker, and therefore 15. Societal cultural practices are related to the
the relationship between such forces and physical and psychological well-being of their
organizational formalization practices will be members. Cultures that are high on power distance
lower. In low power distance cultures, we expect and low on humane orientation practices will
that forces toward centralization of decision have members who are dissatisfied with life
making will be weaker, and therefore the rela- in general. Further, in such societies members
tionship between such forces and decentraliza- will have less social support and less access to
tion and delegation practices will be lower. We medical care.
specify such moderating effects in detail below
when we discuss Phase 3 hypotheses. In summary, the differentiating values and
practices of each culture and the organizational
10. Leader acceptance is a function of the inter- contingencies faced by each organization will be
action between CLTs and leader attributes predictive of the leader attributes and behaviors
and behaviors. Accordingly, leader attributes and organizational practices that are most fre-
and behaviors that are congruent with CLTs will quently perceived as acceptable and most
be more accepted than leader attributes and frequently enacted. Further, selected attributes
behaviors that are not congruent with CLTs. of cultural practices will predict the economic
competitiveness of nations and physical and
11. Leader effectiveness is a function of the psychological well-being of their members.
interaction between strategic organizational The theory is depicted in Figure 2.1 in the
contingencies and leader attributes and behav- form of a systems model. The complexity of
iors. Leader attributes and behaviors that meet the model, however, mitigates being tested
the requirements of strategic organizational comprehensively. Rather, individual linkages
contingencies will result in increased leader or subsets of linkages can be tested, and the
effectiveness. validity of the model can be inferred from such
tests. House, Wright, and Aditya (1997) present
12. Acceptance of the leader by followers facil- a more detailed, fine-grained elaboration of the
itates leader effectiveness. Leaders who are not relationships depicted in Figure 2.1.
accepted by members of their organization will
find it more difficult and arduous to influence
these members than leaders who are accepted. TWO FUNDAMENTAL
CROSS-CULTURAL ISSUES
13. Leader effectiveness, over time, will increase
leader acceptance. Effective leaders will, over Two central aspects of cultures are frequently
time, demonstrate their competence by being discussed in cross-cultural literature: culturally
effective. Demonstration of competence will generalizable and culturally specific aspects.
change attitudes of some of the organizational Culturally generalizable phenomena are com-
members toward the leader and result in mon to all cultures to some extent. A phe-
increased acceptance of the leader. Further, over nomenon is culturally generalizable if all
time, followers who do not accept the leader will cultures can be assessed in terms of a common
either be dismissed or will leave their organiza- metric and cultures can be compared in terms
tion voluntarily. of such phenomena. In contrast, culturally
specific phenomena occur in only a subset of
14. Societal cultural practices are related to cultures, and are not comparable across all
economic competitiveness of nations. Societies cultures.
that are high on power distance and low on uncer- Project GLOBE uses a variety of methods
tainty avoidance and performance orientation to make cross-cultural comparisons. The pri-
will be less competitive internationally, because mary sources of data used to measure the
these dimensions of culture are assumed to core GLOBE dimensions are questionnaire
impede competitive performance. responses of middle managers in three selected
02-House.qxd 3/19/2004 4:07 PM Page 20
20– • –INTRODUCTION
industries: financial services, food processing, each dimension are presented in Chapters 12
and telecommunications. Data were collected through 19.
from approximately 17,000 managers in 951
organizations throughout the world. At least three
cultures in each major region of the world are THE GLOBE QUESTIONNAIRE SCALES
included in the sample (see Chapter 10 by Gupta
and Hanges in this volume, for a description of the In this section, we very briefly describe Phase 1
regions studied). When aggregated to the culture of Project GLOBE. Phase 1 is concerned with
level of analysis, responses to the questionnaire scale development and validation and is
provide measurement of the practices and values described in more detail in Chapter 8 by Hanges
relevant to the nine core GLOBE dimensions. and Dickson. After briefly describing Phase 1,
The sampling strategy for the collection of we will present an overview of the methodology
questionnaire data “controls” for nation, indus- and objectives for GLOBE Phase 2. The remain-
try, occupation broadly defined (managers), and der of the book will then be devoted to literature
organizational level broadly defined (middle reviews, and explication of the methodology of
management). The samples for Phase 2, in which Phase 2, results, discussion, and conclusion.
hypotheses are tested, are described in more
detail in Chapter 6 by House and Hanges.
Sampling from middle managers permits us to
Questionnaire
make generalizations about the subcultures of
Development and Validation
middle managers in the three industries studied.
This sampling strategy increases the internal Two pilot studies were conducted. The first
validity of the study by ensuring that the units was designed to identify appropriate factor
of analysis are well defined and internally homo- structures for the societal level, organizational
geneous and therefore comparable. level, and leadership questionnaires. The second
Strictly speaking, we are studying the study was designed to replicate the first study
cultures of middle managers in 62 cultures. and determine whether the findings are stable.
Thus, we must be cautious when making gener- The two pilot studies and analyses of the data
alizations about national cultures. However, as collected for Phase 2 all demonstrated that the
we demonstrate in Chapter 9 by Gupta, Sully de scales had sound psychometric properties.
Luque, and House, the core GLOBE societal More specifically, the GLOBE scales are
practice and value dimensions are strongly all unidimensional and demonstrate significant
and significantly correlated with unobtrusive and nontrivial within-culture response agree-
measures that reflect the broader society. Also ment, between-culture differences, and respec-
demonstrated in Chapter 9 is that the core table reliability of response consistency.
GLOBE societal value measures are signifi- Generalizability coefficients, which are joint
cantly correlated with independently collected measures of these psychometric properties,
indicators of societal values in the World Value exceed .85 for all scales. These coefficients indi-
Survey (Inglehart, Basanez, & Moreno, 1998). cate that the scales can be meaningfully applied
Thus we can conclude that both the GLOBE to measure leadership variables and cultural
questionnaire responses reflect the broader variables at the societal and organizational levels
culture in which the middle managers are of analysis.
embedded rather than the cultures of middle Statistical analyses reported in Chapter 8
managers alone. by Hanges and Dickson also provide support
The term construct refers to the construction for the construct validity of the culture scales,
of conceptions or ideas by the investigator. A within the context of a nomological network.
construct is a product of the investigator’s Further, correlations between unobtrusive mea-
creativity. The major constructs investigated sures and corresponding societal level organiza-
in the GLOBE research program are the tional practice scales were all significant.
nine dimensions of cultures described previ- Correlations between the societal values scales
ously. Examples of questionnaire items for and independently gathered data relevant to
02-House.qxd 3/19/2004 4:07 PM Page 21
each scale were also found to be significant. and Dickson (Chapter 8, this volume). The
(Inglehart et al., 1998) The findings also demon- relationships between the six global leader
strated strong convergent and discriminate behaviors and the nine cultural dimension
validity of the societal-level scales. These find- scales are described in Chapter 21 by Dorfman,
ings are reported in more detail in Chapter 9 by Hanges, and Brodbeck.
Gupta, Sully de Luque, and House.
Statistical analyses also indicated justifica-
The Societal and
tion for grouping the items into scales relevant
Organizational Culture Questionnaires
to nine core GLOBE values. Analyses also indi-
cated justification for aggregating the scales to Questionnaire items for the nine core
the organizational and societal levels of analy- GLOBE dimensions were written to elicit
sis. We demonstrate that for all culture dimen- reports of current societal and organizational
sion scales measuring values, and for most practices and respondents’ values with respect
cultural dimension scales measuring practices, to these practices. Questionnaire items were
there are significant and substantial relation- derived from a review of relevant literature,
ships between organizational practices and the interviews, and focus groups conducted in sev-
societies in which they are embedded. eral cultures, as well as extant organizational
Finally, the generalizability coefficient and culture theory. These items were designed
which is a joint measure of the above properties to obtain data about both the societal and the
is .85 or above for all scales. This coefficient organizational cultural variables. Respondents
is well above the minimum standard for such rated the items on a 7-point Likert-type scale.
research. The development and validation of For some scales, the response indicators ranged
these dimensions is described in more detail in from 1, indicating high agreement, to 7, indicat-
Chapter 8 by Hanges and Dickson. ing high disagreement. For other scales, the
verbal anchors in the 7-point scale reflected the
end points on a continuum (e.g., 1 = assertive,
The Leadership Questionnaire
7 = nonassertive). All culture scales, however,
The GLOBE Leader Attributes and Behavior were 7-point scales. The items were written as
Questionnaire includes 112 leader attributes and “quartets” having isomorphic structures across
behaviors items. Leader attributes were rated the two levels of analysis (societal and organi-
1 through 7 with 1 indicating “This behavior zational) and across the two culture manifesta-
or characteristic greatly inhibits a person from tions (As Is and Should Be).
being an outstanding leader” to a high of 7 indi- Table 2.3 contains an example of a quartet of
cating “This behavior or characteristic con- parallel culture items showing essentially the
tributes greatly to a person being an outstanding same question in four forms: organization and
leader.” Examples of the Leader Behavior society practices (questions with As Is response
Questionnaire items are presented Table 2.2. format) and organizational and societal values
These items are based on a review of the (questions with Should Be response format).
leadership literature as well as findings relevant Responses to As Is questions reveal the percep-
to leadership resulting from focus groups, inter- tions of middle managers concerning current
views, and analysis of media. Factor analysis practices in their societies and their organiza-
yielded 21 leadership subscales. A second order tions. Responses to Should Be questions reveal
factor analysis of the 21 scales yielded four managers’ values with respect to what they
factors. Two of the factors were subdivided believe should be the practices in their societies
into two subscales each, thus yielding six or organizations.
global leader behavior dimensions. We will To eliminate the problem of common
report statistical analysis relevant to these source bias respondents were divided into two
dimensions after describing the societal and groups within each organization studied. The
organizational culture questionnaire. The 21 Alpha group responded to questions designed
leadership subscales and the six global leader to capture the respondents’ reports of the
behavior dimensions are presented by Hanges core GLOBE dimensions with respect to their
02-House.qxd 3/19/2004 4:07 PM Page 22
22– • –INTRODUCTION
Table 2.2 Sample Items and Response Alternatives From the Culturally Endorsed Leadership Theory
(CLT) Questionnaire
Organization As Is
1 2 3 4 5 6 7
Organization Should Be
In this organization, the pay and bonus system should be designed to maximize:
1 2 3 4 5 6 7
Society As Is
1 2 3 4 5 6 7
Society Should Be
I believe that the economic system in this society should be designed to maximize:
1 2 3 4 5 6 7
24– • –INTRODUCTION
new procedure to estimate and remove response These divisions preserved conceptual clarity
bias for each country with respect to the core while providing dimensions related to current
GLOBE dimensions and CLTs (see Chapter 8 leadership theory and prior empirical studies.
by Hanges & Dickson, Chapter 21 by Dorfman The 21 subscales are grouped into six higher-
et al., and Appendix C which describes in more order leader behavior–attribute dimensions,
detail how we treated response bias). which are presented in Chapter 8. As previously
stated, we refer to the higher-order dimen-
Phase 2: Universal and Culturally sions as global CLT dimensions because
they represent classes of leader behavior rather
Contingent CLT dimensions
than specific leader behaviors. We refer to the
One of the objectives of GLOBE is to deter- 21 first-order factors as primary CLT subscales.
mine whether there are CLT dimensions that are These subscales measure more-specific leader
universally endorsed and dimensions that are dif- attributes and behaviors. Composite profiles
ferentially endorsed across cultures. Recall that of the six CLT dimensions represent what is
CLTs are culturally endorsed dimensions com- generally referred to as leadership styles.
posed of perceived effective or ineffective leader
attributes or behaviors about which members
within each culture agree. Profiles of CLT STRENGTHS OF THE
dimensions reflect what is commonly referred to GLOBE RESEARCH DESIGN
as “leadership styles” in the leadership literature.
Shaw (1990) suggests that much of the Project GLOBE differs from previous cross-
cross-cultural literature indicates differences in cultural research in several ways. The primary
managerial beliefs, values, and style. He also strength of this research is that we have not made
suggests that these differences are a reflection assumptions about how best to measure cultural
of societal cultural differences. Dorfman (2004) phenomena. Rather, we use multiple measure-
and House, Wright, and Aditya (1997) have ment methods to empirically test which methods
reviewed the relevant cross-cultural literature are most meaningful. This is most evident in the
extensively. They find support for Shaw’s development of four sets of measures assessing
suggestion in studies by O’Connell, Lord, and culture: (a) those based on shared values of orga-
O’Connell (1990) and Gerstner and Day (1994). nizational or society members, (b) those based on
Using the means of Phase 2 leader attribute reported current organizational and societal prac-
questionnaire subscales from 61 countries, we tices, (c) unobtrusive measures, and (d) those based
performed a multilevel confirmatory factor on scales derived from the World Value Survey.
analysis to reconfirm the factor analyses con- We developed new measures and collected
ducted in two prior pilot studies. The factor original data for our independent variables rather
structure of the Phase 2 leadership attribute and than relying on measures developed at other times
behavior questionnaire confirmed findings of in other places from other samples. Because the
the two prior pilot studies. Second-order factor organizational culture, societal culture, and leader-
analysis of the resulting 112 leader behavior ship measures used in Phase 2 were completed by
and attribute items produced four factors: different respondents, we were able to eliminate
(1) Team-Oriented Charismatic/Value-Based the frequently encountered problem of common
leadership (2) Humane-Oriented leadership source bias. The psychometric properties of the
(3), Participative-Self-Protective leadership and GLOBE scales and tests of their validity exceed
(4) Autonomous leadership. Assisted by prevail- normal empirical research standards and are
ing theory, we divided Factor 1 into Charismatic/ described in Chapter 8 by Hanges and Dickson.
Value-Based leadership and Team-Oriented
leadership to create two dimensions. We also
divided Factor 4 into two dimensions: Self- CONTRIBUTIONS
Protective leadership and Participative leadership
(the scores of the nonparticipative subscales The results of the GLOBE research program
were reversed to reflect Participative leadership). fill a substantial knowledge gap concerning
02-House.qxd 3/19/2004 4:07 PM Page 25
the cross-cultural forces relevant to effective conduct negotiations with commercial and
leadership, effective organizational practices, political organizations in other cultures.
national prosperity, and the physical and psy- Knowledge of the culturally endorsed
chological welfare of members of cultures. The implicit theories of leadership in each culture,
research findings will be useful not only in and the reported most- and least-effective leader
resolving several important theoretical social attributes and behaviors will be useful to those
science issues but also in answering a wide working with members of the cultures studied.
range of practical questions. Activities such as selecting, counseling, and
training these individuals will be facilitated by
the results of this study. Potential managers can
Practical Relevance
benefit from information regarding the kinds of
The quantitative findings presented in this behaviors and organizational practices that are
book will provide substantial enlighten- acceptable and effective and unacceptable and
ment concerning the processes by which culture ineffective in the cultures studied. Recall that
influences leadership and organizational prac- our sample of cultures includes at least three
tices. Chapter 1 represents an illustration of how societies from each major region of the world.
knowledge of the GLOBE core dimensions can Further, recall that we collected data from three
be useful for practical purposes. In subsequent quite diverse industries; within each industry,
chapters, the reader will find a specification organizations face substantially different prob-
of universal and culturally contingent leader lems, forms of competition, and demands from
attributes and behaviors, as well as commonly external sources. We found a high degree of
enacted and most-favored leader behavior correlation among the findings across the
patterns and organizational practices found in three industries. Also recall that our findings are
the cultures studied. The reader will also be supported by independently collected data. For
able to develop an understanding of the role of these reasons, we believe that our findings are
cultural influences on the reported effectiveness quite generalizable.
of leader behaviors and organizational practices. Information regarding the constraints imposed
In addition, leader behaviors that are culturally on leaders by cultural norms will be useful to
offensive are identified and described (see decision makers who need to anticipate and
Chapters 12 through 21). respond to the actions of leaders of other cultures.
Finally, we related our independent variables Knowledge about cultural and organizational
to objective dependent variables taken from norms and practices can be used to inform
independently collected archival data. The depen- the formulation of meaningful prescriptions for
dent variables are country prosperity (GNP per strategy and policy formulation, organizational
capita), scores on the Human Development improvement, human resource management
Index, and indicators of the psychological and practices, and the design of organization struc-
physical well-being of the members of the tures and incentive and control systems.
cultures studied. This information will be use-
ful as content for leadership training and career
Beneficial Social
development programs intended to prepare
and Economic Applications
individuals who will manage and lead others in
unfamiliar cultures. The research program is expected to have
The country scores on our independent several additional beneficial social and eco-
variables will be useful for the adjustment and nomic applications. Countries that share similar
effective interaction of individuals who work regional resources and backgrounds can make
with others from the cultures studied. More comparisons to determine similarities and dif-
specifically, this information will be useful to ferences among themselves and share ways to
expatriates, managers of diverse cultural and improve intercountry relationships, economic
ethnic groups both domestic and abroad, indi- productivity, and quality of life for their citi-
viduals involved in the management of public zens. The empirical findings have also led to
and private international affairs, and those who increased intercultural communication among
02-House.qxd 3/19/2004 4:07 PM Page 26
26– • –INTRODUCTION
educators who normally would not have contact expectancy rates, hygiene practices, preventative
with each other. The GLOBE CCIs have been or remedial medical practices, stress levels, sui-
extremely active in practicing cross-cultural cide rates, frequency of ethnic and border con-
communications. GLOBE-related research has flicts, indicators of social unrest, and violations
been presented in more than 250 conference of human rights. To date, cultural influences on
papers and more than 100 working papers, such variables have largely been ignored.
chapters of books, or journal articles. Chapter 7 by Javidan and Hauser reports
Many of the CCIs serve as university faculty relationships between GLOBE societal level
members, social scientists, and consultants. variables and economic prosperity of nations,
Being indigenous to their cultures, many CCIs the Human Development Index, and measures
are influence and change agents within those of the psychological and physical welfare of the
cultures. We expect that the GLOBE findings members of cultures. The worldwide data has
will be useful to individuals functioning as change been used to determine relationships between
agents within their societies. societal level GLOBE dimensions and many
indices of social and physical well-being.
Ancillary Social Science Contributions The following chapter presents a nontechni-
cal summary of the major research findings
The resulting data can be used for multiple resulting from Phases 1 and 2 of the GLOBE
purposes beyond the hypotheses of the study. research program.
For example, the worldwide Phase 2 data can be
used to compare countries with their trading
partners or their major competitors with respect APPENDIX: STRUCTURAL
to cultural, organizational, or leadership prac- CONTINGENCY THEORY
tices. These comparisons can lead to strategies
to improve trade or to facilitate harmonious Although a report of research findings relative
and productive relationships across borders. We to this theory is not included in this book
have already witnessed more than 50 research because the results of Phase 3 are not yet avail-
projects and papers presented at scholarly con- able, we describe this theory for the sake of
ferences in which cultural and managerial prac- completeness. The central proposition of this
tices have been compared among subsets of the theory is that there is a set of demands that are
GLOBE participating nations. imposed on organizations that must be met for
Relationships between the independent vari- them to ensure survival and guarantee effective-
ables of this study and economic prosperity have ness. These demands are referred to as organi-
also been subjected to analysis. The societal- zational contingencies. It is asserted that these
level data can be used in subsequent analyses of contingencies influence organizational form
important issues such as the construction of and practice and that congruence between the
econometric or sociological models. The data demands of the contingencies and organiza-
collected for our independent variables can be tional form and practice is associated with orga-
related to firm-level practices such as forms nizational effectiveness. Although this is a
of production systems and organization, transfer popular theory, its empirical verification is lim-
of technology, pricing, risk taking when ited to small sample studies of organizations in
entering new markets, investment in research industrialized countries (Child, 1981). Hickson,
and development, and foreign investment prac- Hinings, McMillan, and Schwitter (1974) have
tices. The measures of culture can also be related asserted that the propositions of structural
to national levels of saving, distribution of contingency theory are universal and cultur-
wealth and social privileges, consumption levels ally transcendent. This assertion rests on the
and patterns, issues of economic growth and assumption that organizational contingencies
development, regulatory practices, and national impose demands on organizations that are so
productivity and efficiency. Further, the GLOBE powerful and universal that it is imperative
societal culture dimensions can be also be for all organizations to respond in essentially
related to such outcomes as mortality rates, life the same way to function effectively and thus
02-House.qxd 3/19/2004 4:07 PM Page 27
survive in competitive environments. We refer to Hanges, P. J., Lord, R. G., Day, D. V., Sipe, W. P.,
this assertion as the task environment imperative. Smith, W. C., & Brown, D. J. (1997, April).
Child (1981) has presented a serious challenge Leadership and gender bias: Dynamic measures
to this assertion. and nonlinear modeling. In R. G. Lord (Chair),
Dynamic systems, leadership perceptions, and
gender effects. Symposium presented at the 12th
ENDNOTE Annual Conference of the Society for Industrial
and Organizational Psychology, St. Louis, MO.
1. We excluded the Czech Republic from the Hickson, D. J., Hinings, C. R., McMillan, J., &
analysis because of pervasive response bias. Schwitter, J. P. (1974). The culture-free context of
organization structure: A tri-national comparison.
Sociology 8, 59–80.
REFERENCES Hofstede, G. (1980). Culture’s consequences: Inter-
national differences in work-related values.
Ajzen, L., & Fishbein, M. (1970). Understanding London: Sage.
attitudes and predicting social behavior. Hofstede, G. (2001). Culture’s consequences:
Englewood Cliffs, NJ: Prentice Hall. International differences in work related values
Bass, B. M. (1985). Leadership and performance (2nd ed.). Thousand Oaks, CA: Sage.
beyond expectations. New York: Free Press. Hofstede, G., & Bond, M. H. (1988). The Confucius
Bass, B. M. (1990). Bass & Stogdill’s handbook of connection: From cultural roots to economic
leadership: Theory, research, and managerial growth. Organizational Dynamics, 16, 4–21.
applications (3rd ed.). New York: Free Press. House, R. J., & Aditya, R. N. (1997). The social
Berry, J. W, & Dasen, P. R. (1974). Introduction. In scientific study of leadership: Quo vadis? Journal
J. W. Berry & P. R Dasen (Eds.), Culture and of Management, 23(3), 409–473.
cognition (pp. 1–20). London: Methuen. House, R. J., Wright, N. S., & Aditya, R. N. (1997).
Burns, T., & Stalker, G. M. (1961). The management Cross-cultural research on organizational lead-
of innovation. London: Tavistock. ership: A critical analysis and a proposed theory.
Child, J. (1981). Culture, contingency, and capitalism In P. C. Earley & M. Erez (Eds.), New perspec-
in the cross-national study of organization. In tives in international industrial organizational
L. L. Cummings (Ed.), Research in organizational psychology (pp. 535–625). San Francisco: New
behavior (pp. 303–356). Greenwich, CT: JAI. Lexington.
Chokkar, J., Brodbeck, F., & House, R. J. (Eds.). (in Inglehart, R., Basanez, M., & Moreno, A. (1998).
press). Cultures of the world: A GLOBE anthol- Human values and beliefs: A cross-cultural
ogy. Thousand Oaks: Sage. sourcebook—Political, religious, sexual, and eco-
Cyert, R., & March, J. (1963). A behavioral theory of nomic norms in 43 societies. Ann Arbor: University
the firm. Englewood Cliffs, NJ: Prentice Hall. of Michigan Press.
Donaldson, L. (1993). Anti-management theories of Kluckhohn, F. R., & Strodtbeck, F. L. (1961). Variations
organization: A critique of paradigm proliferation. in value orientations. New York: HarperCollins.
Cambridge, UK: Cambridge University Press. Kopelman, R. E., Brief, A. P., & Guzzo, R. A. (1990).
Dorfman, P. W. (2004). International and cross-cultural The role of climate and culture in produc-
leadership research. In B. J. Punnett & O. Shenkar tivity. In B. Schneider (Ed.), Organizational cli-
(Eds.), Handbook for international management mate and culture (pp. 282–318). San Francisco:
research (2nd ed., pp. 265–355). Ann Arbor, MI: Jossey-Bass.
University of Michigan. Lawrence, P. R., & Lorsch, J. W. (1967). Organi-
Gerstner, C. R., & Day, D. V. (1994). Cross-cultural zation and environment. Cambridge, MA: Harvard
comparison of leadership prototypes. Leadership University Press.
Quarterly, 5(2), 121–134. Lombardo, M. M. (1983). I felt it as soon as I walked
Hanges, P. J., Braverman, E. P., & Rentsch, J. R. in. Issues and Observations, 3(4), 7–8.
(1991). Changes in raters’ impressions of subor- Lord, R., & Maher, K. J. (1991). Leadership and
dinates: A catastrophe model. Journal of Applied information processing: Linking perceptions
Psychology, 76, 878–888. and performance. Boston: Unwin-Everyman.
02-House.qxd 3/19/2004 4:07 PM Page 28
28– • –INTRODUCTION
3
A NONTECHNICAL SUMMARY
OF GLOBE FINDINGS
MANSOUR JAVIDAN
ROBERT J. HOUSE
PETER W. DORFMAN
his chapter provides a brief summary of the GLOBE research program, which is
29
03-House.qxd 3/19/2004 4:07 PM Page 30
30– • –INTRODUCTION
Power Distance: The degree to which members of Followers are (should be) expected to obey their
a collective expect power to be distributed equally. leaders without question.
Uncertainty Avoidance: The extent to which a Most people lead (should lead) highly structured
society, organization, or group relies on social lives with few unexpected events.
norms, rules, and procedures to alleviate
unpredictability of future events.
Humane Orientation: The degree to which a People are generally (should be generally) very
collective encourages and rewards individuals tolerant of mistakes.
for being fair, altruistic, generous, caring, and
kind to others.
Collectivism II (In-Group Collectivism): The Employees feel (should feel) great loyalty toward
degree to which individuals express pride, loyalty, this organization.
and cohesiveness in their organizations
or families.
Assertiveness: The degree to which individuals are People are (should be) generally dominant in
assertive, confrontational, and aggressive in their their relationships with each other.
relationships with others.
Gender Egalitarianism: The degree to which a Boys are encouraged (should be encouraged)
collective minimizes gender inequality. more than girls to attain a higher education.
(Scored inversely.)
Future Orientation: The extent to which More people live (should live) for the present
individuals engage in future-oriented behaviors rather than for the future. (Scored inversely.)
such as delaying gratification, planning, and
investing in the future.
Performance Orientation: The degree to which a Students are encouraged (should be encouraged)
collective encourages and rewards group members to strive for continuously improved performance.
for performance improvement and excellence.
item was measured using a scale ranging from and practices. The following are several of
1 to 7. GLOBE dimensions of societal culture the more interesting findings by the GLOBE
have been validated through the use of unobtru- research program.
sive measures and independently collected data
from the World Values Survey (Inglehart,
Basanez, & Moreno, 1998). Table 3.2 shows the SOCIETAL CULTURES
descriptive statistics for GLOBE cultural
dimensions. Appendix A provides the correla- • The means for the nine cultural practices scores
tion coefficients among all the cultural values range from 3.37 (Gender Egalitarianism) to
03-House.qxd 3/19/2004 4:07 PM Page 31
Table 3.2 Means and Standard Deviations for GLOBE Cultural Practices and Values Descriptive
Statistics
N = 61 societal cultures
5.17 (Power Distance). Most GLOBE societies relatively high levels of power distance and are
are reported to be somewhat male oriented and generally in-group oriented.
to experience high levels of power distance. • The mean score of Gender Egalitarianism
• The means for six of the cultural practices practices is the lowest among all practice
scores, namely, Uncertainty Avoidance, dimensions (3.37) indicating that GLOBE
Future Orientation, Institutional Collectivism, societies are reported to be male oriented.
Humane Orientation, Performance Orientation, • The means for the nine cultural values scores
and Assertiveness are around 4, which is the range from 2.75 (Power Distance) to 5.94
mid-point of the scale. (Performance Orientation). Most GLOBE socie-
• The means for two cultural practices scores, ties prefer lower levels of power differentiation
namely, Power Distance and In-Group Collec- and higher levels of performance orientation.
tivism, are above 5 on the 7-point scale. GLOBE • The largest range of societal scores for
societies are generally reported to experience cultural practices among GLOBE societies is
03-House.qxd 3/19/2004 4:07 PM Page 32
32– • –INTRODUCTION
related to In-Group Collectivism. The highest but very close on their In-Group Collectivism
reported country score on this dimension was values.
6.36 (Philippines) and the lowest reported • The mean scores for values are lower than
country score was 3.53 (Denmark) for a those for practices in the case of two dimen-
difference of 2.83. sions: Power Distance (− 2.42) and Asser-
• The smallest range of societal scores for cul- tiveness (−0.32). For all the other dimensions,
tural practices among GLOBE societies is the mean values scores are higher than the
related to Assertiveness. The highest reported mean practices scores.
country score on this dimension was 4.89 • Only for one dimension, Gender Egalitarianism,
(Albania) and the lowest reported country the correlation between the reported values
score was 3.38 (Sweden) for a difference and practices scores is significant and positive.
of 1.51. For seven other dimensions, it is significant
• Interestingly, the opposite is true for the and negative. It is insignificant for In-Group
GLOBE range of scores for cultural values. Collectivism.
The lowest range relates to In-Group • Ten cultural clusters were identified: Latin
Collectivism (1.58) and the largest range America, Anglo, Latin Europe (Italy, Portugal,
relates to Assertiveness (2.90). In other words, Spain, France, Switzerland [French-speaking],
GLOBE societies are reported to be rather Israel), Nordic Europe, Germanic Europe,
close in their Assertiveness practices, but very Confucian Asia, Sub-Saharan Africa, Middle
different in terms of Assertiveness values. In East, Southern Asia, and Eastern Europe. The
contrast, they are reported to be very different cultural profiles of these clusters are reported in
in terms of In-Group Collectivism practices Figures 3.1 to 3.10.
(Text continues on page 37)
Anglo
Performance Orientation
7.000
6.000
Uncertainty Avoidance Assertiveness
5.000
4.000
3.000
In-Group Collectivism 2.000 Future Orientation
1.000
Practices
Values
Figure 3.1 Cultural Dimension Scores for the Anglo Societal Cluster
03-House.qxd 3/19/2004 4:07 PM Page 33
Germanic Europe
Performance Orientation
7.000
6.000
Uncertainty Avoidance Assertiveness
5.000
4.000
3.000
1.000
Practices
Values
Figure 3.2 Cultural Dimension Scores for the Germanic Europe Societal Cluster
Confucian Asia
Performance Orientation
7.000
4.000
3.000
In-Group Collectivism 2.000 Future Orientation
1.000 Practices
Values
Figure 3.3 Cultural Dimension Scores for the Confucian Asia Societal Cluster
03-House.qxd 3/19/2004 4:07 PM Page 34
34– • –INTRODUCTION
Sub-Saharan Africa
Performance Orientation
7.000
1.000
Practices
Values
Figure 3.4 Cultural Dimension Scores for the Sub-Saharan Africa Societal Cluster
Middle East
Performance Orientation
7.000
6.000
Uncertainty Avoidance Assertiveness
5.000
4.000
3.000
1.000
Practices
Values
Figure 3.5 Cultural Dimension Scores for the Middle East Societal Cluster
03-House.qxd 3/19/2004 4:07 PM Page 35
Southern Asia
Performance Orientation
7.000
4.000
3.000
In-Group Collectivism 2.000 Future Orientation
1.000
Practices
Values
Figure 3.6 Cultural Dimension Scores for the Southern Asia Societal Cluster
Eastern Europe
Performance Orientation
7.000
4.000
3.000
In-Group Collectivism 2.000 Future Orientation
1.000
Practices
Values
Figure 3.7 Cultural Dimension Scores for the Eastern Europe Societal Cluster
03-House.qxd 3/19/2004 4:07 PM Page 36
36– • –INTRODUCTION
Latin Europe
Performance Orientation
7.000
4.000
3.000
1.000
Practices
Values
Figure 3.8 Cultural Dimension Scores for the Latin Europe Societal Cluster
Nordic Europe
Performance Orientation
7.000
5.000
4.000
3.000
In-Group Collectivism 2.000 Future Orientation
1.000
Practices
Values
Figure 3.9 Cultural Dimension Scores for the Nordic Europe Societal Cluster
03-House.qxd 3/19/2004 4:07 PM Page 37
Latin America
Performance Orientation
7.000
6.000
Uncertainty Avoidance Assertiveness
5.000
4.000
3.000
In-Group Collectivism 2.000 Future Orientation
1.000
Practices
Values
Figure 3.10 Cultural Dimension Scores for the Latin America Societal Cluster
38– • –INTRODUCTION
Gender
Egalitarianism
Assertiveness
Humane
Orientation
prosperity, government support for prosperity, Orientation, and Uncertainty Avoidance. They
and competitiveness ranking (Table 3.3). are both negatively and significantly related to
• The Societal Power Distance practice score Power Distance practice scores (Table 3.3).
is negatively and significantly related to • No cultural practice or value dimension is
all measures of economic health except related to economic productivity (Tables 3.3
productivity (Table 3.3). and 3.4).
• Economic prosperity and the World Compe- • Economic prosperity and competitive rankings
titiveness Index are positively and significantly are both negatively correlated with Perfor-
related to four cultural practices dimensions: mance Orientation, Future Orientation, Institu-
Performance Orientation, Future Orientation, tional Collectivism, and Uncertainty Avoidance
Institutional Collectivism, and Uncertainty value scores (Table 3.4).
Avoidance. They are both negatively and sig- • Societal health and general satisfaction are
nificantly related to In-Group Collectivism and positively and significantly correlated with
Power Distance practice scores (Table 3.3). practice scores of Performance Orientation,
• Government and societal support for prosperity Future Orientation, and Uncertainty Avoi-
and competitiveness are both positively and dance. They are negatively and significantly
significantly related to three cultural practice related with In-Group Collectivism and Power
dimensions: Performance Orientation, Future Distance practices (Table 3.5).
03-House.qxd 3/19/2004 4:07 PM Page 39
Gender .36**
Egalitarianism n = 57
Assertiveness
In-Group −.30*
Collectivism n = 57
Humane
Orientation
• The Human Development Index (HDI) and Each item is defined and measured on a scale
life expectancy are negatively and signifi- ranging from 1 to 7. A score of 1 means that
cantly related to practice scores on In-Group the attribute greatly inhibits outstanding leader-
Collectivism, Power Distance, and Humane ship and a score of 7 means that the attribute
Orientation. HDI is positively correlated with contributes greatly to outstanding leadership.
Gender Egalitarianism (Table 3.5).
• Societal health, general satisfaction, and HDI • GLOBE identified 22 leadership attributes that
are all negatively correlated with Future are universally desirable. Ninety-five percent
Orientation and Uncertainty Avoidance values of the societal average scores for these attrib-
scores (Table 3.6). utes were larger than 5 on a 7-point scale, and
• No cultural practice or value dimension is rela- the worldwide grand mean score exceeded 6
ted to psychological health (Tables 3.5 and 3.6). on a 7-point scale. “Decisiveness” and “fore-
sight” are examples of such attributes.
• Eight leadership attributes were identified as
CULTURE AND LEADERSHIP universally undesirable. Ninety-five percent of
the societal average scores for these attributes
Leadership attributes are measured through a were less than 3 on a 7-point scale, and the
questionnaire containing 112 leadership items. worldwide grand mean score was lower than 3
03-House.qxd 3/19/2004 4:07 PM Page 40
40– • –INTRODUCTION
Human
Cultural Societal Human Life General Psychological Development
Practices Health Health Expectancy Satisfaction Health Index (HDI)
Gender .29*
Egalitarianism n = 56
Assertiveness
Institutional
Collectivism
on a 7-point scale. “Irritable” and “ruthless” highly suspicious of individuals who are in
are examples of such attributes. positions of authority for fear that they will
• Many leadership attributes are culturally acquire and abuse power. In these cultures,
contingent. They are desirable in some cul- substantial constraints are placed on what
tures and undesirable in others. “Ambitious” individuals in positions of authority can and
is an example with a societal score ranging cannot do, and such individuals are given no
from 2.85 to 6.73. “Elitist” is another example special treatment, status, or privileges.
with a societal score range of 1.61 to 5.00. In • We were able to empirically establish that
some cultures the concept of leadership is there are culturally based shared conceptions of
romanticized and leaders are given excep- leadership, referred to as culturally endorsed
tional privileges and status and are held in implicit theories of leadership (CLT). That is,
great esteem. members of cultures share common observa-
• During focus interviews, it was reported that tions and values concerning what constitutes
in some cultures like the Netherlands or effective and ineffective leadership. Leadership
Switzerland, the concept of leadership is attributes were statistically grouped into 21
denigrated and members of the cultures are “first-order” primary factors (henceforth called
03-House.qxd 3/19/2004 4:07 PM Page 41
Human
Cultural Societal Human Life General Psychological Development
Values Health Health Expectancy Satisfaction Health Index (HDI)
Performance −.30*
Orientation n = 56
Assertiveness
Institutional −.33*
Collectivism n = 40
In-Group −.39*
Collectivism n = 40
Power
Distance
Humane
Orientation
primary leadership dimensions) that were then • Team Oriented leadership is generally
consolidated into six “second-order” global reported to contribute to outstanding leader-
leadership dimensions. Combined, they repre- ship. The range of mean societal scores
sent the culturally endorsed leadership theory among GLOBE countries is 4.7 to 6.2 on a
dimensions (CLTs). 7-point scale.
• The six global CLT leadership dimensions • Participative leadership is generally reported
are: (a) Charismatic/Value-Based leadership, to contribute to outstanding leadership,
(b) Team Oriented leadership, (c) Participa- although there are meaningful differences
tive leadership, (d) Autonomous leadership, among countries and clusters. The range of
(e) Humane-Oriented leadership, and (f) Self- mean societal scores among GLOBE countries
Protective leadership. is 4.5 to 6.1 on a 7-point scale.
• Charismatic/Value-Based leadership is gener- • Humane-Oriented leadership is reported to be
ally reported to contribute to outstanding neutral in some societies and to moderately
leadership. The range of mean societal scores contribute to outstanding leadership in others.
among GLOBE countries is 4.5 to 6.5 on a The range of mean societal scores is 3.8 to 5.6
7-point scale. on a 7-point scale.
03-House.qxd 3/19/2004 4:07 PM Page 42
42– • –INTRODUCTION
• Autonomous leadership is reported to range in the Anglo cluster (6.05) and the lowest score
from impeding outstanding leadership to in the Middle East cluster (5.35).
slightly facilitating outstanding leadership. • Team Oriented leadership receives the highest
The range of mean societal scores is 2.3 to 4.7 reported score in the Latin American cluster
on a 7-point scale. (5.96) and the lowest score in the Middle East
• Self-Protective leadership is generally cluster (5.47).
reported to impede outstanding leadership. • Participative leadership receives the highest
The range of mean societal scores among reported score in Germanic Europe cluster
GLOBE countries is 2.5 to 4.6 on a 7-point (5.86) and the lowest score in the Middle East
scale. cluster (4.97).
• Humane-Oriented leadership receives the
highest reported score in Southern Asia (5.38)
LEADERSHIP PROFILES OF and the lowest score in Nordic Europe (4.42).
CULTURAL CLUSTERS • Autonomous leadership receives the highest
reported score in Eastern Europe (4.20) and
• Cluster scores on the culturally endorsed lead- the lowest score in Latin America (3.51).
ership theory (CLT) dimensions are provided in • Self-Protective leadership receives the highest
Figures 3.11 to 3.16. Charismatic/Value-Based reported score in Southern Asia (3.83) and the
leadership receives the highest reported score lowest score in Nordic Europe (2.72).
Eastern Europe
7.00
5.74
Middle East Latin America
5.00 5.99
5.35
5.79 5.93
Sub-Sahara Africa Nordic Europe
6.05
Anglo
Maximum potential score of 7.
Figure 3.11 Charismatic/Value-Based Leadership Dimension Scores for Each Culture Cluster
03-House.qxd 3/19/2004 4:07 PM Page 43
Eastern Europe
7.00
5.88
Middle East Latin America
5.00
5.96
5.47
3.00
Germanic Europe Latin Europe
5.73
5.62
1.00
5.70 5.77
Sub-Sahara Africa Nordic Europe
5.74
Figure 3.12 Team Oriented Leadership Dimension Scores for Each Culture Cluster
Eastern Europe
7.00
3.00
Germanic Europe Latin Europe
5.86 5.37
1.00
5.31 5.75
Sub-Sahara Africa Nordic Europe
5.73
Anglo
Maximum potential score of 7.
Figure 3.13 Participative Leadership Dimension Scores for Each Culture Cluster
03-House.qxd 3/19/2004 4:07 PM Page 44
44– • –INTRODUCTION
Eastern Europe
7.00
4.80 4.85
3.00
Germanic Europe Latin Europe
4.71 4.45
1.00
5.38 5.04
Southern Asia Confucian Asia
4.42
5.16
5.08
Sub-Sahara Africa Nordic Europe
Anglo
Maximum potential score of 7.
Figure 3.14 Humane-Oriented Leadership Dimension Scores for Each Culture Cluster
Eastern Europe
7.00
4.16 3.66
1.00
3.99 4.04
Southern Asia Confucian Asia
3.63 3.94
3.82
Anglo
Maximum potential score of 7.
Figure 3.15 Autonomous Leadership Dimension Scores for Each Culture Cluster
03-House.qxd 3/19/2004 4:07 PM Page 45
Eastern Europe
7.00
3.83 3.72
Southern Asia 2.72 Confucian Asia
3.55 3.08
Anglo
Maximum potential score of 7.
Figure 3.16 Self-Protective Leadership Dimension Scores for Each Culture Cluster
46– • –INTRODUCTION
Charismatic/Value-Based Leadership
Performance Orientation
Figure 3.17 Cultural Value Drivers of the Charismatic Value-Based CLT Leadership Dimension
NOTES: This figure summarizes information from Chapters 12–19 and 21–22. As such, it combines results from both the
single and multiple HLM statistical tests linking cultural values to leadership dimensions.
* The most important cultural dimensions for this leadership CLT are bolded.
Team-Oriented Leadership
Humane Orientation +
• Collaborative team
Performance Orientation orientation
• Team integration
Future Orientation
• Diplomatic
• Malevolent (reverse-
[No dimensions were scored)
negatively related.] • Administratively
− competent
Figure 3.18 Cultural Value Drivers of the Team Oriented CLT Leadership Dimension
NOTES: This figure summarizes information from Chapters 12–19 and 21–22. As such, it combines results from both the
single and multiple HLM statistical tests linking cultural values to leadership dimensions.
* The most important cultural dimensions for this leadership CLT are bolded.
03-House.qxd 3/19/2004 4:07 PM Page 47
Participative Leadership
Humane Orientation
• Participative
Uncertainty Avoidance • Autocratic
(reverse-scored)
Power Distance
−
Assertiveness
Figure 3.19 Cultural Value Drivers of the Participative CLT Leadership Dimension
NOTES: This figure summarizes information from Chapters 12–19 and 21–22. As such, it combines results from both the
single and multiple HLM statistical tests linking cultural values to leadership dimensions.
* The most important cultural dimensions for this leadership CLT are bolded.
Figure 3.20 Cultural Value Drivers of the Humane Oriented CLT Leadership Dimension
NOTES: This figure summarizes information from Chapters 12–19 and 21–22. As such, it combines results from both the
single and multiple HLM statistical tests linking cultural values to leadership dimensions.
* The most important cultural dimensions for this leadership CLT are bolded.
03-House.qxd 3/19/2004 4:07 PM Page 48
48– • –INTRODUCTION
Autonomous Leadership
Autonomous
+ Leadership CLT
Performance Orientation
• Individualistic
• Independent
Humane Orientation • Autonomous
• Unique
Institutional Collectivism
−
Figure 3.21 Cultural Value Drivers of the Autonomous CLT Leadership Dimension
NOTES: This figure summarizes information from Chapters 12–19 and 21–22. As such, it combines results from both the
single and multiple HLM statistical tests linking cultural values to leadership dimensions.
* The most important cultural dimensions for this leadership CLT are bolded.
Self-Protective Leadership
Self-Protective
Power Distance + Leadership CLT
Uncertainty Avoidance
• Self-centered
• Status conscious
Gender Egalitarianism • Conflict inducer
In-Group Collectivism • Face saver
• Procedural
Performance Orientation −
Figure 3.22 Cultural Value Drivers of the Self-Protective CLT Leadership Dimension
NOTES: This figure summarizes information from Chapters 12–19 and 21–22. As such, it combines results from both the
single and multiple HLM statistical tests linking cultural values to leadership dimensions.
* The most important cultural dimensions for this leadership CLT are bolded.
REFERENCES
Inglehart, R., Basanez, M., & Moreno, A. (1998). Human World Economic Forum. (1998). The global com-
values and beliefs: A cross-cultural sourcebook. petiveness report. Geneva, Switzerland: Author.
Ann Arbor: University of Michigan Press. United Nations Development Program. (1998).
The World Bank. (2000). World development indica- Human development report. New York: Oxford
tors. Washington, DC: Author. University Press.
04-House.qxd 3/19/2004 4:08 PM Page 49
PART II
PRIOR LITERATURE
PETER W. DORFMAN
49
04-House.qxd 3/19/2004 4:08 PM Page 50
the mechanisms by which the influence is Thus, the information in Part II provides
enacted. For instance, the level and degree of ample evidence that cultural forces influence
governmental regulation, development of the many aspects of leadership, industry, and orga-
industry within a society, and national economic nizational phenomena. And clearly, we are
system are just a few of the many factors that much further along in our understanding of the
affect the ways in which a given industry is cultural influence on these processes than we
enacted in a given society. In recent years, orga- were 38 years ago when Haire, Ghiselli, and
nizational researchers have fortunately begun to Porter (1966) published their seminal study. Yet
identify potential explanatory mechanisms to much remains to be learned. These chapters set
understand these relationships more thoroughly. the stage for understanding the potential interac-
Theoretical models that are discussed in this tion between leadership and societal culture,
chapter include: Cultural immersion, social net- industry sectors, and organizational cultures.
work, resource dependency, and institutional
theory. The chapter concludes with the premise
that truly understanding when and how these REFERENCES
mechanisms affect an industry requires both cul-
turally generalizable and culturally specific per- Gordon, G. G. (1991). Industry determinants of organi-
spectives. The culturally generalizable approach zational culture. Academy of Management Review,
is necessary to identify which general principles 16, 396–415.
are commonly more active than others, whereas Haire, M., Ghiselli, E. E., & Porter, L. (1966).
the culturally specific perspective tells us about Managerial thinking: An international study.
the mechanisms active in specific situations. New York: John Wiley.
04-House.qxd 3/19/2004 4:08 PM Page 51
4
CULTURAL INFLUENCES ON
ORGANIZATIONAL LEADERSHIP
Literature Review, Theoretical
Rationale, and GLOBE Project Goals
PETER W. DORFMAN
ROBERT J. HOUSE
AUTHORS’ NOTE: Phases 1 and 2 of the GLOBE research program were funded by the Dwight D. Eisenhower
Leadership Education Program of the Department of Education of the United States. Funds to support data collec-
tion in three African countries were provided by a grant from the Reginald Jones Center for Strategic Leadership.
51
04-House.qxd 3/19/2004 4:08 PM Page 52
• The Malaysian leader is expected to behave in management practices should the Euromanager
a manner that is humble, modest, and dignified use, and for which subordinates, colleagues, and
(House et al., 1999). supervisors? The multicultural reality is not just
• In India, there is a preference for leadership that an abstract phenomenon to ponder; real people in
is proactive, morally principled, ideological, those companies are searching for more effective
bold, and assertive, in contrast to reactive, management practices. Titus Lokananta, for
pragmatic, instrumental, quiet, and nurturing example, is an Indonesian Cantonese holding a
(Chhokar, in press). German passport, managing a Mexican multi-
• The combination of family and tribal norms and national corporation producing Gummy Bears
bureaucratic organizational structures fosters in the Czech Republic (Millman, 2000).
authoritarian management practices that may Should we, however, conclude that cultural
be characterized as a “sheikocracy” leadership differences are so vast that finding common
style in Arabian countries (Al-Kubaisy, 1985). management practices among countries are
the exception rather than the rule and will ever
These examples speak to the multifaceted remain so? Not necessarily. Companies are
question “to what extent is leadership culturally forced to share information, resources, and train-
contingent”? For many years the academic ing in a global world economy. The best busi-
literature barely touched on this issue, yet expa- ness schools educate managers from all over
triate managers working in multinational com- the world in the latest management techniques.
panies hardly need to be reminded of the wide Using academic jargon, the issue of common
variety of management practices found around versus unique business and management prac-
the world. Laurent (1983), and more recently tices is framed using contrasting perspectives
Trompenaars (1993), document the astonishing embodied in the terms cultural universals versus
diversity of organizational practices worldwide, cultural specifics. The former are thought to be
of which many are acceptable and considered found from the process of cultural convergence
effective in one country but ineffective in a whereas the latter from maintaining cultural
neighboring country. For instance, supervisors divergence. Perhaps not surprisingly, empirical
are expected to have precise answers to sub- research supports both views. For example, in
ordinates’ questions in Japan, but less so in the their event management leadership research
United States. As another example, the effec- program Smith and Peterson (1988) found both
tiveness of working alone or in a group is per- commonalities and differences across cultures
ceived very differently around the world and in the manner by which managers handled rela-
this will certainly influence the quality and tively routine events in their work. All managers
aptitude of virtual teams found in multina- preferred to rely on their own experience and
tional organizations (Davis & Bryant, 2003). training if appointing a new subordinate, relative
An inescapable conclusion is that acceptable to other influences such as consultation with
management practices found in one country are others or using formal rules and procedures.
hardly guaranteed to work in a different country. However, there were major differences in
Although fascinating in its own right, does countries in the degree to which managers used
the diversity of management practices world- formal company rules and procedures in contrast
wide constitute a management problem? We to more informal networks, and these differ-
believe it does, given the interrelatedness of ences covary with national cultural values
the world economy. Consider the management (Smith, 2003; Smith, Dugan, & Trompenaars,
dilemma posed within the European Union. To 1996). As another example, Hazucha and col-
the extent that management practices in Europe leagues (Hazucha, Hezlett, Bontems-Wackens,
are said to stop at national boundaries, the & Ronnqvist, 1999) found a good deal of simi-
emergence of a generic “Euromanager” in the larity among European countries regarding the
European Union becomes problematic (Brodbeck importance of core management competencies
et al., 2000). Training a knowledgeable, glob- for a Euromanager. Yet there were significant
ally competent manager is an obvious problem differences among countries in the perceived
for multinational European companies: What attainment of these skills.
04-House.qxd 3/19/2004 4:08 PM Page 53
Should we also expect that leadership cultural variations as parameters of the theory.
processes, like management practices, be simi- Cross-cultural research may also help uncover
larly influenced by culture? The apparent new theoretical relationships by forcing the
answer is yes—substantial empirical evidence researcher to consider a much broader range
(see House et al., 1997) indicates that leader of noncultural variables (Chemers, 1983). For
attributes, behavior, status, and influence vary instance, models promoting participative lead-
considerably as a result of culturally unique ership may be valid for relatively sophisticated
forces in the countries or regions in which the employees in developed countries, but less valid
leaders function. But, as the colloquial saying for employees in less-developed countries
goes “the devil is in the details,” and current where egalitarianism may not be highly valued.
cross-cultural theory is inadequate to clarify and Cultural variations may therefore highlight
expand on the diverse cultural universals and relationships between theoretical constructs
cultural specifics that have been elucidated in and specify important theoretical boundary
cross-cultural research. Some researchers cur- conditions. The GLOBE research program was
rently subscribe to the philosophy that the pri- designed to further theory development by
mary impact of cultures depends on the level of devising an empirically based theory of cross-
analysis used in the research program. That is, cultural leadership in addition to helping man-
some view the basic functions of leadership agerial leaders by specifying the how, why,
as having universal importance and applica- what, and where of cultural impacts on leader-
bility, but the specific ways in which leadership ship processes.
functions are enacted are strongly affected by Before embarking on a discussion of our cur-
cultural variation (Chemers, 1997; Smith & rent state of knowledge regarding the influence
Peterson, 1988). The GLOBE project will test of culture on leadership, it may be worthwhile
this fundamental assumption. From a practical to consider the degree to which cultural influ-
viewpoint, because the meaning of specific ence is expected to play a significant role in
leadership actions and behaviors are influenced the future given today’s global world. We ask
by the cultural context (Smith, 1997), we need the question, are cultural influences a transient
more information to help leaders practice every- phenomenon in a fast changing global econ-
day leadership, such as information on partici- omy? Academics often frame this question as
pation, direction, coaching, and motivation. a discussion of the “culture convergence
There is also a scientific rationale, in addition hypothesis.”
to practical applications, for conducting cross-
cultural research. From a scientific and theoret-
ical perspective, compelling reasons exist for CULTURAL CONVERGENCE:
considering the role of culture in influencing HOW FAR WILL IT GO?
leadership processes. Conventional practice
labels theories that transcend cultures as “tran- Are cultural differences here to stay, or will
scultural,” “culture free,” “cultural universal,” forces of globalization blur differences among
or “etic processes.” However, we should realize nations and organizations? It is often suggested
the inherent limitations in transferring theories that there is a substantial amount of cross-
across cultures (Chemers, 1997; Poortinga & national convergence of management practices,
Malpass, 1986): what works in one culture may values, and beliefs (Boyacigiller, Kleinberg,
not necessarily work in another. Research that Phillips, & Sackmann, 1996). This occurs as a
supports cultural diversity is generally referred result of interactions among organizations
to as “cultural specific,” “culturally unique,” or engaged in cross-border trading, and wide-
“emic processes.” Through cross-cultural spread proliferation of management education
research, we may determine which aspects of programs that reflect Western assumptions, val-
a leadership theory are culturally universal ues, and practices. Some scholars subscribe to
and which are culturally unique. As Triandis the most extreme form of the cultural conver-
(1993) suggests, leadership researchers will gence thesis: convergence of management prin-
be able to fine-tune theories by investigating ciples and practices throughout the industrial
04-House.qxd 3/19/2004 4:08 PM Page 54
Nations can exhibit a profound backlash various cultures and a need for an empirically
against leaders who represent countercultural grounded theory to explain differential leader
values. Statues of high-level communist leaders behavior and effectiveness across cultures”
in Moscow and Budapest were retired to “ceme- (House, 1995, pp. 443–444; see also Bass, 1990;
teries of statues,” preserving them for posterity Boyacigiller & Adler, 1991; Dorfman, 2004;
but keeping them out of public view. The Irish Dorfman & Ronen, 1991).
Republican Army destroyed Lord Nelson’s
statue in Dublin and the Irish government never
replaced it because of its pro-British symbolism. THE GLOBE DEFINITION OF LEADERSHIP
Institutions, streets, and cities were converted to
their pre-Communist names in almost all former As previously discussed in the overview
Soviet Union nations: former Leningrad is once (Chapter 2), in August 1994 the first GLOBE
again St. Petersburg and Stalingrad is now research conference was held at the University
Volgograd. The former Karl Marx Institute of of Calgary in Canada. Fifty-four researchers
Economics in Budapest is now the Hungarian from 38 countries gathered to develop a collec-
Institute of Economics. tive understanding of the project and to initiate
its implementation. At this meeting, GLOBE
researchers generated a working definition
WESTERN-DOMINATED of leadership reflecting diverse viewpoints. A
LEADERSHIP RESEARCH consensus emerged for a universal definition
of organizational leadership: the ability of an
Most of the leadership research during the past individual to influence, motivate, and enable
half-century was conducted in the United States, others to contribute toward the effectiveness
Canada, and Western Europe (Yukl, 2002). The and success of the organizations of which they
same description applies to the researchers are members. Note that this is a definition of
who conducted these studies because they have organizational leadership, not leadership in
been primarily trained in the West. Perhaps as general.
a result, almost all prevailing theories of leader- The GLOBE project uses the concept of
ship and most empirical evidence is North implicit leadership theory (ILT) as a critical
American in character, that is individualistic explanatory mechanism by which culture
rather than collectivistic; emphasizing U.S. influences leadership processes. Leadership
assumptions of rationality rather than ascetics, prototypes within ILT are profiles of presumed
religion, or superstition; stated in terms of indi- effective or preferred leader attributes or behav-
vidual rather than group incentives; stressing iors (Lord & Maher, 1991). Using prototypes
follower responsibilities rather than rights; in this manner is a natural development of the
assuming hedonistic rather than altruistic moti- social information processing literature (Croker,
vation; and assuming centrality of work and Fiske, & Taylor, 1984; Rosch, 1975). Models
democratic value orientation (House, 1995, developed from this literature can illustrate
pp. 443–444). Hofstede (1993) makes a similar how cultures shape the basic ways people col-
point: “In a global perspective, US management lect, store, organize, and process information
theories contain several idiosyncrasies not about leaders. Shaw (1990) argues that much
necessarily shared by management elsewhere” of the cross-national literature indicating
(p. 81). Hofstede mentions three such idiosyn- differences in managerial beliefs, values, and
crasies: A stress on market processes, a stress on styles can be interpreted as showing culturally
the individual, and a focus on managers rather influenced differences in leader prototypes.
than workers. Cross-cultural psychological, Prototypes may include leader behaviors,
sociological, and anthropological research values, attitudes, and personality traits. At this
shows that many cultures do not share these point we are not certain whether there are
assumptions. As a result, many scholars have universally endorsed prototypes of effective
noted the obvious “need for a better understand- leaders, or even universal characteristics of
ing of the way in which leadership is enacted in effective leaders. Project GLOBE intends to
04-House.qxd 3/19/2004 4:08 PM Page 57
answer this question. The cross-cultural with the experiential definition of culture. The
research literature examining prototypes will be method of analysis for almost all of the studies in
explored in more detail later in this chapter and the Bass review was the group mean comparison
in Chapter 21. based on aggregation of individual responses.
Two major trends in the cross-cultural lead-
ership literature were evident from the Bass
VARIED PERSPECTIVES ON CULTURE review. First, substantial research has been con-
ducted to examine the applicability of Western
As with leadership, there is no consensually leadership theory in multiple national settings.
agreed-on definition among social scientists for Second, a great deal of effort has been made to
the term culture, a term used by social scientists compare the leadership styles and requirements
to refer to a set of parameters of collectives that of small groups of nations. Usually, the com-
differentiate the collectives from each other in parisons are made among the U.S., Western
meaningful ways. The focus is on the “shared- European, Latin American, and Asian nations.
ness” of cultural indicators among members Consequently, more is known about leadership
of the collective. Culture is variously defined in in these regions than is known about leadership
terms of several commonly shared processes: in Southern Asian, African, Arab, and Eastern
shared ways of thinking, feeling, and reacting; European countries.
shared meanings and identities; shared socially The Bass review also reveals a number of
constructed environments; common ways in which shortcomings in the cross-cultural leadership
technologies are used; and commonly experi- research literature. First, there is a lack of theo-
enced events including the history, language, retical cohesiveness among the studies cited.
and religion of their members. As discussed Whereas some investigators draw from well-
in the overview chapter, Project GLOBE defines established theories of leadership, many merely
culture as “shared motives, values, beliefs, describe national differences and draw on rather
identities, and interpretations or meanings of atheoretical and unsystematic intellectual
significant events that result from common frameworks. Second, Bass found that there was
experiences of members of collectives and a dearth of studies based on more than three or
are transmitted across age generations.” This four countries. Third, many studies make use
definition can apply for both societies and of existing standardized U.S. instruments that
organizations, and may be investigated at the may not fully capture non-Western or non-U.S.
societal and organizational levels of analysis. conceptualizations of leadership.
The good news is that since the Bass 1990
review, cross-cultural leadership theory and
PRIOR EMPIRICAL RESEARCH: research have improved immeasurably. More
CULTURE AND LEADERSHIP recent studies frequently are grounded in
theory, based on more than a comparison of
We now turn to a review of the empirical cross- two or three countries, use sophisticated quanti-
cultural leadership literature and specifically tative analysis, provide in-depth qualitative
concentrate on general topics of interest and, for descriptions, and often use perspectives from
the most part, studies conducted after the 1990 researchers in non-Western countries. Subsequent
review of cross-cultural leadership research to the review by Bass (1990), extensive reviews
that appeared in Bass and Stogdill’s Handbook of the cross-cultural leadership literature can be
of Leadership Research (Bass, 1990). Certain found in Chemers (1997); Dorfman (1996,
conclusions from the Bass review, however, are 2004); House, Wright, and Aditya (1997); and
worth noting. Specifically, many studies were Peterson and Hunt (1997).
concerned with the effects of differences in cul- Conducting cross-cultural research is, by its
tural or subcultural units on managerial behavi- very nature, difficult to conceptualize and con-
ors, attitudes, preferences, and motivations. duct. As a result, far more questions than answers
National boundaries were used to specify cultural still exist regarding the culturally contingent
units in almost all of these studies, consistent aspects of leadership; teams of cross-national
04-House.qxd 3/19/2004 4:08 PM Page 58
researchers find that cultural differences are real Location of residence, in addition to national
and affect everyday decisions and interactions origin, often has a powerful effect on individuals
among team members (Dorfman, Hanges, & (Huo & Randall, 1991).
Dickson, 1998). The following review must, by
necessity, be selective and focused on research Modernization
that paints a picture of what we currently know
about the influence of culture on leadership Differential expectations of, and preferences
processes and, perhaps more importantly, what for, specific leader behaviors will vary accord-
we do not know about these processes that the ing to the degree of societal modernization.
GLOBE project was designed to study. Merit is more often expected to be the basis
of performance evaluation and compensation
in modernized nations. Leaders in countries
LEADERSHIP TOPICS STUDIED where cultural norms are more tradition-bound
often take factors other than merit into account
WITHIN A CROSS-CULTURAL
if deciding on salary increases (Bass et al.,
LEADERSHIP FRAMEWORK 1979; Ryterbrand & Thiagarajan, 1968; Ulin,
1976). They are also more likely to find
We first summarize major topics that emerged bribery to be acceptable, and are more likely
from our literature review. Part of our review to relegate women to lower-status positions
considers non-Western conceptualizations and (Davis, Ming, & Brosnan, 1986). On the other
studies of leadership processes that often have hand, leaders from modern societies tend to
escaped notice. We then describe the multiple focus more on issues of merit, orderliness,
goals and practical applications of the GLOBE punctuality, intended rationality, and progress
Project and finish with a list of future projects. (Bass et al., 1979; Inkeles, 1966).
In the remainder of this section, we describe
findings relevant to the most frequent topics
addressed in the cross-cultural leadership Unique Role Demands of Leaders
literature. For heuristic purposes, the review is
arranged roughly in an order that mirrors the The differential role demands placed on
variables presented within the GLOBE inte- leaders may vary according to demographic
grated leadership model presented in the composition of organizations, national or regional
overview chapter (see Figure 1 in Chapter 2). political systems, or the strategic requirements of
the leader’s organizations (Anzizu & Nuenos,
1984; Bass, 1990; Heller, 1958). Organizational
Origin of Leaders management practices in China, India, and Hong
Kong are often based on kinship relationships—
Ascriptions of achievement provide normal that is, hiring relatives is often the norm rather
routes for an individual’s rise to leadership than the exception—a system used in many large-
and vary in a predictable manner according to scale enterprises in these countries as well. Large
one’s culture. Other influential factors such as Indian firms practice many of these same behav-
differences in education, class, occupation, own- iors, such as obedience to elders based on defer-
ership, and technical expertise are also deter- ence to the wisdom of experience (Chowdry &
minants to filling leadership roles in different Tarneja, 1961). This is a tendency that persists
countries (Boyd, 1974; Harbron, 1965; Lee & today. Five of the largest business organizations
Schwendiman, 1982; McClelland, 1961). A in India—Reliance (managed by the Ambani
leader’s origin also influences his or her behav- family), Birla, Goenka, Kirloskar, and Tata—
ior if working as an expatriate. For example, remain family-managed. In Mexico, the large
managers of Pakistani origin working in family owned and operated business structures
England exhibited a substantial effect of origin called grupos are examples in point. The impor-
not shared by British managers (Shackleton & tance of strong family ties and paternalistic
Ali, 1990). However, the effect of origin did not management practices are emphasized, and
overwhelm the context of expatriate managers. these businesses retain their management
04-House.qxd 3/19/2004 4:08 PM Page 59
characteristics even after expansion to larger storyteller, no one wearing eyeglasses, and
organizational entities (Dorfman & Howell, one who has run a marathon (Ritchie, 1999).
1997). Samsung and Hyundai Motor Company, Several relatively small-scale studies have
Korean chaebols, also fit this same model of demonstrated that leadership prototypes vary
family-centered conglomerates in which leader- across cultures.
ship succession is family dominated (Steers, An initial study by O’Connell, Lord, and
Shinn, & Ungson, 1989). O’Connell (1990) supports the argument that
societal culture plays a strong role in the
content of organizational leadership pro-
Antecedents to Preferred Leader Behavior
totypes. Japanese and American leadership
Several antecedents to preferred leader prototypes differed in significant respects.
behavior are addressed in the cross-cultural Gerstner and Day (1994) provide additional
leadership literature. Religious beliefs and evidence that leadership prototypes vary
the cultural norms expressed by the dominant across cultures. As expected, attributes that
elite are often predecessors to leadership posi- were seen as most characteristic of business
tions. Also, preferences for certain leader leaders varied across countries. The Gerstner
behaviors have been shown to be associated and Day (1994) study and the O’Connell and
with the dominant norms of cultural entities colleagues (1990) study are useful exemplars
(Redding & Casey, 1975; Stening & Wong, of research examining leadership prototypes,
1983), gender egalitarianism (Paris, 2003), and but findings are limited by small samples and
religious or ideological values such as Confuci- other methodological issues. More recently,
anism (Hofstede & Bond, 1988), Catholicism Den Hartog and colleagues (1999), using a
(Pelletier, 1966), and the Protestant work ethic subset of the GLOBE data presented in this
of saving, sacrifice, hard work, and investment book, tested the hypothesis that specific
(Weber, 1930). McClelland (1961) and Boyd aspects of transformational/charismatic lead-
(1974) noted the role that landed gentry played ership are strongly and universally endorsed
in the development of leadership traits in across cultures. Brodbeck and colleagues
Confucian China and Victorian England. Speci- (2000), also using a subset of the GLOBE
fically, leaders in both countries were expected data, tested the assumption that leadership
to possess good manners, good physical con- concepts are culturally endorsed and leader-
ditioning, and classical training. Interviews ship prototype dimensions are highly corre-
conducted in France by the second author of lated with cultural dimensions found within
this chapter indicate that French leaders are Europe. In-depth discussions of prior research
generally well educated and expected to be on ILT, rationale for using ILT as a main
“cultivated”—that is, classically educated. It is construct for GLOBE, and research results
likely that the dominant elite serve as role mod- linking culture to leadership prototypes are
els for aspiring leaders, especially if the elite are presented in Chapter 21.
respected and trusted.
Preferences for Leadership Styles
Leader Prototypes
Followers differ by nation in their prefer-
As introduced earlier in this chapter, leader- ences and acceptance for intensity and kind
ship prototypes are profiles of presumed typical of communication with leaders (Earley, 1984),
or preferred leader attributes or behaviors (Lord task versus person orientation (Blake &
& Maher, 1991). Prototypes may include leader Mouton, 1970; Misumi, 1974; Scandura, Von
behaviors, values, attitudes, and personality Glinow, & Lowe, 1999), and responses to orga-
traits. To cite one amusing example, executive nizational development efforts (Deyo, 1978).
recruiters reported that the following attributes They also differ regarding preferences for
were among the more unusual characteristics close versus general supervision, democratic
they were asked to look for: Over 6 feet tall, versus autocratic leader behavior, importance
10 handicap or lower in golf, fiddle player and of morality (Hui & Tan, 1999), and emphasis
04-House.qxd 3/19/2004 4:08 PM Page 60
placed on interaction facilitation (Gibson, example, Dorfman and colleagues (1997) found
1995). A prevailing difference, which will be evidence for the universal effectiveness of
supported by the GLOBE project, is subordi- three popular contingency leadership behaviors
nates’ preference of leadership styles that vary (leader supportiveness, contingent reward, and
along the differing degrees of participatory charismatic leadership) and also found culturally
practices (Bass, 1990; Bass et al., 1979; Osland, unique influences for three other leadership
Monteze, & Hunter, 1998; Torres, 2000). behaviors (participativeness, directiveness, and
contingent punishment). Although their study
was limited to five countries in the Pacific Rim,
Leadership Behavior Patterns
including the United States and Mexico, it
Modal leader behavior patterns differ widely reinforces the often stated belief that the “indi-
across countries in their emphasis on individ- vidualistic” United States is different from more
ualistic versus team orientation; particularism “collectivistic” countries. Differences were most
versus universalism (Dorfman, 1998; Dorfman & notably apparent by the unique effects of par-
Howell, 1988; Smith et al., 1996); performance ticipation and contingent punishment having
versus maintenance orientation (Shenkar, Ronen, positive impacts only in the United States.
Shefy, & Chow, 1998; Smith, Misumi, Tayeb, What follows is a summary of cross-cultural
Peterson, & Bond, 1989; Smith, Peterson, leadership studies with respect to commonly
Misumi, & Bond, 1992); authoritarian versus studied behavioral aspects of leadership. Task-
democratic orientation (Al-Hajjeh, 1984; Stening oriented and relationship-oriented leadership
& Wong, 1983); paternalism (Redding & Casey, dimensions, which were originally conceptual-
1975); reliance on personal abilities, subordi- ized in the United States, appear to be important
nates, or rules (Smith & Peterson, 1995); leader leadership behaviors in many cultures (Bond &
influence processes (Rahim, Kim, & Kim, 1994; Hwang, 1986; Chemers, 1997; Misumi, 1985;
Schmidt & Yeh, 1992); and consensual decision Sinha, 1980). In general, cross-cultural studies
making and service orientation (Bass et al., support the importance of considerate leader-
1979). ship in increasing subordinates’ satisfaction.
The universality of leader supportiveness
should not be surprising because supportive
Behavioral Impact of Leadership
leaders show concern for followers and are con-
Finally, there have been many studies investi- siderate and available to listen to follower’s
gating the effects of culture on the generally problems. Exceptions to this general finding,
accepted dependent variables in the discipline however, can be found and are particularly
of organizational behavior. Scholars have exam- informative in pointing out cultural differences.
ined the effect of cultural influences on employee As one example, leadership effectiveness in the
attitudes such as job satisfaction, motivation, job Middle East, given the strong role of Islam
performance, and other criteria such as general in the Arabic world, led to leadership effective-
welfare. Most studies conducted by leadership ness not being related to the supervisor’s level
researchers in different parts of the world reflect of consideration, but rather to the supervisor’s
an awareness and understanding of leadership strong directive behavior (Scandura, Von Glinow,
research conducted in America and other & Lowe, 1999).
Western countries (Smith & Peterson, 1988). Results concerning task-oriented leadership
Keeping with the behavioral tradition within the are complex and defy simple explanation (Bass,
leadership literature, much of the empirical cross- 1990). Korman’s (1966) review of studies in the
cultural research in the past 30 years has been U.S. concludes that there is a lack of consistent
ethnocentric (research designed and tested in results for task-oriented leadership. Not surpris-
one culture and replicated in another culture) ingly, cross-cultural studies examining the
rather than truly comparative (designed to test impact of directive behaviors on employees also
similarities and differences across two or more show conflicting and not easily interpretable
cultures) (Adler, 1984). Consistencies and differ- results (Anderson, 1983; Kakar, 1971; Kennis,
ences have been found among cultures. As one 1977; Yukl, 2002).
04-House.qxd 3/19/2004 4:08 PM Page 61
Participatory and charismatic leadership 2002), charismatic theories fare quite well under
styles have also been studied in cross-cultural empirical scrutiny. Numerous studies using a
contexts, with the former showing major differ- variety of methods (including field studies, case
ences among cultures and the latter having more histories, management games, interviews, and
universally positive effects. In the Dorfman and laboratory experiments) and samples (including
colleagues (1997) study, participatory leadership middle- and lower-level managers, top-level
had positive correlations with satisfaction with corporate leaders, educational, and national
one’s supervisor for all nations, but when con- leaders) attest to the significant and robust impact
trolling for simultaneous effects of several dif- of charismatic and transformational leader-
ferent leadership behaviors, the United States ship (Bass, 1997). Furthermore, Bass (1991)
was the single culture in which participative notes confirming cross-cultural evidence for
leadership had a positive influence on employee the proposition that there is a hierarchy of lead-
performance. The actual level of participation ership effectiveness among various leadership
for the United States was also the highest in all styles—transformational leaders are more effec-
sampled countries. We might also note that the tive than those practicing transactional leader-
Vroom and Jago (1988) model of participatory ship, who, in turn, are more effective than
decision making should be a useful theoretic laissez-faire leaders (who are not effective).
approach to examine leadership differences This hierarchy of relationships was found for
across cultures. In one exploratory study, Jago field grade officers in Germany and Canada;
and colleagues (1993) found that German, New Zealand professionals and administrators;
Austrian, and Swiss managers were the most senior managers in Italy and Sweden; and middle-
participative; Polish and Czech managers the level managers in Japan, Belgium, Canada, Spain,
most autocratic; and the U.S. and French man- Saudi Arabia, and India (Bass, 1991; Bass &
agers between the extremes. Two other findings Avolio, 1993; Bass, Waldman, Avolio, & Bebb,
were particularly interesting. First, unlike man- 1987; Bass & Yokochi, 1991; Howell & Avolio,
agers of other nations, Polish managers were 1993). These studies generally support a near
more likely to be participative only on trivial universalistic position regarding the potential
matters and not on important issues. Second, impact of charismatic and transformational
there were differences across countries in leadership across cultures.
scenarios in which subordinate conflict was On a less supportive note, a methodologi-
likely—for example, the U.S. and Polish man- cally sophisticated study examining the effects
agers were likely to become autocratic in of charismatic leadership on individuals at
conflict-producing situations. The researchers multiple organizational levels found little sup-
developed predictions of participation based on port for the Shamir, House, and Arthur’s self-
Hofstede’s power distance scores and results concept-based charismatic theory (Shamir,
were generally supportive for the prediction that Zakay, Breinin, & Popper, 1998). Clearly, much
participation scores would be higher for lower remains to be learned about the process by
power distance cultures (Jago et al., 1993). We which charismatic leadership affects followers.
suspect that further understanding about partici- Certainly, if considered from a cross-cultural
patory leadership will be even more important perspective, one can entertain a hypothesis that
for multinationals as societies move toward more the enactment of charismatic leadership and
egalitarian practices and because power sharing transformational leadership will likely be cul-
is critical within virtual teams that span time and ture specific. In addition, it would be wise not to
geographic boundaries (Davis & Bryant, 2003). forget that individuals in societies previously
Not surprisingly, strong empirical evidence dominated by charismatic dictators generally
supports the importance of charismatic leader- view autocratic charismatic leadership as unde-
ship in the West, and growing evidence suggests sirable. Interviews and focus groups conducted
that it is also important in non-Western societies. as an early part of the GLOBE research program
Despite substantive and legitimate criticism, revealed strong reservations, suspicions, and
including definition and measurement problems distaste for authoritarian charismatic leadership
(Hunt, 1991; Smith & Peterson, 1988; Yukl, among German, Mexican, Portuguese, and
04-House.qxd 3/19/2004 4:08 PM Page 62
adequately characterize Chinese leadership (Ling cultures that value hierarchical and personalized
& Fang, 2003; Wang, 1994). relationships. It is likely not a coincidence that
Similar to Misumi’s research in Japan, early Sinha’s NT leader resembles Misumi’s PM
efforts to study leadership in India were influ- leader in Japan, the “benevolent/paternalistic”
enced by conceptual links to Western social leader in Iran (Ayman & Chemers, 1983), the
scientists. However, research results were often paternalistic patron in Mexico (Bass, 1990),
inconsistent and at odds with prevailing beliefs or the family-dominated leaders of Korean
about the nature of effective leadership in Indian chaebols (Steers, Shin, & Ungson, 1989). The
organizations. Research data often supported the importance of both the Misumi and Sinha
interpretation that democratic, participative, and research programs is that they emphasize the
considerate leaders were most effective, whereas importance of context as determining which
managers and workers often voiced a preference processes of leadership are considered effective
for paternalistic and nurturing leaders who are and which are not (Smith, 1997).
also authoritarian and assertive (Kakar, 1971; As a final consideration of non-Western
Sinha, 1994). To reconcile these inconsistent conceptualizations of leadership, it is interesting
beliefs about effective leadership, Sinha (1980, to consider leadership styles in countries very
1984) developed a Nurturant-task oriented different from the West such as in the Arabic
model (NT) that incorporated a combination of world of the Middle East. The pervasive influence
leadership styles. The model suggests that an of the Islamic religion is a key to understanding
ideal leader in India is both nurturant the Arab world and, presumably, leadership in
and task oriented. According to the theory, a the Arab world (Hagan, 1995). In Arabic, the
nurturant-task-oriented (NT) leader needs to word for leadership is al kiyada, which refers to
show affection, care for subordinates, and com- officers in the military or high-ranking members
mit toward their growth. However, the leader’s of the government. Historically, a leader is a great
nurturance is contingent on the subordinate’s hero who leads warriors into battle, and therefore
task accomplishment—the leader becomes a not unexpectedly, the concept of leadership is
benevolent source of support provided that the rooted in traditional military concepts of leader-
subordinate respects and obeys the supervisor, ship (Scandura et al., 1999). Modern Arab
works hard, and is highly productive. In short, the management practices have been influenced by
leader is paternalistic and authoritative. At this Islamic religion, tribal and family traditions, the
point in the relationship, however, the leader can legacy of colonial bureaucracies, and contact with
encourage the subordinate to be more indepen- Western nations (Ali, 1990). Tribal traditions
dent and to actively participate with the leader in influence all aspects of life and, as a consequence,
decision making. Still, this relationship is very managers are expected to act as fathers—viewing
much like the Japanese management-familyism their role in a highly personalized manner charac-
system in which the supervisor benevolently terized by providing and caring for employees
guides the subordinate who, in turn, must recip- and favoring individuals within the family and
rocate with complete obedience and absolute tribe over outsiders. The legacy of a highly struc-
loyalty (Whitehill & Takezawa, 1968). tured bureaucracy left by the ruling Ottoman
Sinha (1980, 1984) presents research evi- Empire and European nations is superimposed on
dence that the effectiveness of NT is affected by these Islamic family–tribal traditions.
a number of contingency variables such as the The combination of family and tribal
subordinate’s desire for a dependency relation- norms in addition to bureaucratic organiza-
ship and acceptance of a hierarchical relation- tional structures fosters authoritarian manage-
ship. Triandis (1994) speculates that even in ment practices that may be characterized as a
cultures not typically appreciative of democratic “sheikocracy” leadership style (Al-Kubaisy,
relationships, participatory leadership can be 1985). This style is characterized by a patriar-
effective given worker training to accept, expect, chal approach to managing that includes strong
and appreciate such styles. We might also spec- hierarchical authority, subordination of effi-
ulate that a form of the NT model would hold ciency to human relations and personal con-
for other high power distance and collectivist nections, and sporadic conformity to rules and
04-House.qxd 3/19/2004 4:08 PM Page 64
regulations contingent on the personality and Nevertheless, the contrasts have significant
power of those who make them. theoretical and practical significance.
Leadership studies in the Middle East are
almost nonexistent due to the inherent difficulty The Cultural Congruence Proposition
of conducting organizational research there.
This proposition asserts that cultural forces
However, a few research studies shed some light
affect the kind of leader behavior that is
on leadership this area of the world. Scandura
usually accepted, enacted, and effective within a
and colleagues (1999) found striking differences
collective. Accordingly, behavior that is consis-
in effective leadership styles between a U.S.
tent with collective values will be more accept-
managerial sample and a sample of managers
able and effective than behavior that represents
from the Middle East (Jordan and Saudi Arabia).
conflicting values. This hypothesis is taken as an
Although they used the fairly old concept of lead-
article of faith among culture theorists. The avail-
ership as being composed of two leadership styles
able empirical evidence supports this hypothesis
(initiating structure and consideration), the results
(House, Wright, & Aditya, 1997); however, a rig-
were dramatic. Whereas the people-oriented style
orous test of this proposition is important for both
(i.e., consideration) was related to job satisfaction
and leader effectiveness for the U.S. sample, the practical and theoretical reasons.
task-oriented style (i.e., initiating-structure) was Following are several examples illustrating
not. Exactly opposite results were found for the the cultural congruence proposition. Nations
Middle East sample. They conclude that strong in Hofstede’s sample that have high power dis-
and decisive leadership is expected from an Arab tance and collectivism scores experience a ten-
person, hence the effectiveness of setting high dency toward behaviors that are consistent with
goals and standards of performance characteristic high scores on these dimensions. For instance,
of initiating structure, whereas a considerate the heavy emphasis placed by Asian managers
leader might be perceived as being weak and on paternalism (Dorfman & Howell, 1988;
indecisive. Scandura and Dorfman, in a series Farmer & Richman, 1965) and group mainte-
nance activities (Bass et al., 1979; Bolon &
of Leadership Quarterly letters (Scandura &
Crain, 1985; Ivancevich et al., 1986) is consis-
Dorfman, in press), discuss culture and charis-
tent with countries that are highly collective.
matic leadership, particularly as they relate to the
Hofstede’s uncertainty avoidance scores are
Middle East. Of course given the current world
associated with less risky entry into foreign
situation, we need to know a great deal more than
markets and more full disclosure of accounting
we do about leadership in this part of the world.
information. Achievement motivation reflected
in grammar school books was found to be pre-
dictive of entrepreneurial behavior 25 years later
THREE COMPETING PROPOSITIONS: (McClelland, 1961). In individualistic societies,
CULTURAL CONGRUENCE, CULTURAL people prefer individual rather than group-based
DIFFERENCE, AND NEAR UNIVERSALITY compensation practices and exhibit greater will-
OF LEADER BEHAVIORS ingness to take risks (Erez, 1997). These empir-
ical findings support the congruence proposition.
The discussion thus far suggests three contrasting A corollary of the cultural congruence pro-
propositions testable within Project GLOBE. The position is that violation of cultural norms by
propositions differ in the extent to which effec- leaders or managers will result in dissatis-
tive leadership is characterized by the congru- faction, conflict, and resistance on the part of
ence, or lack of congruence, between endemic followers or subordinates and, at times, lower
cultural forces and specific styles, behaviors, and performance of leaders, their work units, and
images associated with leaders. The three propo- their subordinates. Anecdotal examples from
sitions are labeled: cultural congruence, cultural the literature on expatriate adjustment illustrate
difference, and near universality. Each of these lower productivity and satisfaction if collective
propositions enjoys some empirical support, but norms and expatriate values conflict (Lindsay &
they have not been explicitly tested to date. Dempsey, 1985; Weiss & Bloom, 1990).
04-House.qxd 3/19/2004 4:08 PM Page 65
The Cultural Difference Proposition Thus, it appears that leaders can initiate
change by being different with respect to some
Juxtaposed to the cultural congruence
leader behaviors. No research has been con-
proposition is the cultural difference proposi-
ducted to determine the precise leader behaviors
tion. According to this proposition, increased
associated with such improvement. We hypoth-
task performance of followers, organizations,
esize that the leader behaviors associated with
and institutions in societies will be induced by
breakthrough improvements in organizations
the introduction of selected values, techniques,
and societies are often those associated with the
and behavior patterns that are different from
introduction of constructive change, such as
those commonly valued in the society. The
articulating a vision of a different way of life
rationale for this hypothesis is that by being
and communicating high performance expecta-
different with respect to some behaviors,
tions and confidence in followers.
leaders introduce more changes of the kind
required for innovation and performance
improvement. The Near Universality
Several examples of minority groups of Leader Behaviors Proposition
leading the way to industrialization, business
development, and entrepreneurship illustrate This proposition asserts that there are some
this hypothesis. Some examples include leader behaviors that are universally, or near
the Methodists in England, the Protestant universally, accepted and effective. Despite
Huguenots in France prior to the Edict of wide-ranging differences in cultural norms
Nantes, and Cubans in the United States who across countries studied, there is some support
fled from Cuban communism. These groups for this proposition. Bass and colleagues (1979)
and their leaders acted and behaved in ways found that managers from 12 culturally diverse
that were largely outside the cultural norms countries indicated a desire to get work done
of the larger societies but were effective in while using less authority. Similarly, Smith and
business development. The incredible success Peterson (1995) found that managers in 30
of the mainland Chinese entrepreneurs in countries reported satisfaction with events for
Southeast Asia presents an equally fascinating which they were delegated substantial discre-
case study of entrepreneurs who are different tion. Transformational leadership has been
from the culture in which they live (Redding, found to be more acceptable and effective in
1990). However, as argued by Smith (1997), Canada (Howell & Avolio, 1993; Howell &
the success of such leaders is due in part to Frost, 1989), India (Pereira, 1987), Japan (Bass,
the congruence between the kinds of ethni- 1997), the Netherlands (Koene, Pennings, &
cally homogeneous businesses created and Schreuder, 1991), and Singapore (Koh,
requisite networks of region-of-origin and Terborg, & Steers, 1991), as well as the United
family connections. In contrast, the proposition States (Bass & Avolio, 1993).
of successful leadership through induction of Bass (1997) argues that the three compo-
practices that are different from the model nents of transformational leadership are near
cultural practices is still viable if we consider universal: charisma, intellectual stimulation of
leaders as shapers of change rather than simply followers, and individualized consideration
embodiments of the status quo. For instance, toward followers. Bass (personal communica-
a central feature of charismatic leadership is tion, June 15, 1996) reports that
the envisioning of change and an innovative
vision to which followers should aspire. Sinha Although some fine tuning may be required, on
(1995) argues that a successful Indian leader all continents people’s ideal leader is transfor-
may have to buck the normally relationship- mational, not transactional. . . . Transformational
oriented culture by placing greater emphasis leadership is more effective than contingent
on task performance, which will eventually reward which in turn is more effective than man-
change the leadership style from being more aging by exception. Laissez faire leadership is
directive to more participative. contraindicated.
04-House.qxd 3/19/2004 4:08 PM Page 66
Bass also argues that leaders can learn to degree to which leadership roles are filled by
become more transformational in formal training ascription or achievement; modal leader behavior
programs. This assertion remains to be empiri- patterns; preferences for and expectations of
cally demonstrated. It should be noted that the leaders; and follower and subordinate reactions
Bass measures operationalize leader behaviors to different kinds of leader behavior. In addition,
(and attributions) as rather general items such as culturally defined antecedents to preferred leader
“inspires loyalty to him or her,” or “is a model behaviors have been revealed through research,
for me to follow.” It is likely that these behaviors including dominant norms, dominant elites, ideo-
(or attribution process) will be enacted differen- logical and religious values, modernization,
tially in a culture-specific manner. For example, unique role demands, and historical experiences
as noted by House and colleagues (1997), trans- with leaders.
formational leadership may be enacted in a Although we are much further along in our
highly assertive manner (as in the cases of John understanding of the cultural influence on
F. Kennedy, Martin Luther King, Jr., Theodore leadership processes than we were almost 40
Roosevelt, or Winston Churchill) or in a quiet, years ago when Haire and colleagues (1966)
non-aggressive manner (as in the cases of conducted their seminal study, much remains
Mahatma Gandhi, Nelson Mandela, or Mother to be learned. It seems likely that the meaning
Teresa). In all of these cases, however, charis- and importance of leadership vary across cul-
matic qualities are evident. These individuals tures, and actions and behaviors are context
articulated an ideological message, set personal specific, but a more precise understanding of
examples of the values inherent in their message, the subtle nuances and precise mechanisms by
conveyed a sense of strong confidence in them- which culture works await further research—
selves and in their followers, and were in turn hence Project GLOBE. Hopefully, the first
highly respected and trusted by their followers. few chapters of the book have made it clear
Several specific hypotheses relating culture that the GLOBE project should be seen as a
to leadership processes will be advanced in journey. The ultimate destination is to validate
chapters throughout this book. One such a cross-level integrated theory specifying the
hypothesis, given the previous discussion of relationship among culture and societal, orga-
universality, might take the position that charis- nizational, and leadership effectiveness. This
matic/value-based leadership and integrity chapter has concentrated on the relationship
attributes will be universally endorsed as con- among culture, leadership, and organizational
tributors to outstanding leadership. Although effectiveness. Although the GLOBE project
not specifically predicted as a hypothesis, we is also concerned with societal effectiveness,
expect that most respondents will view some the present chapter has not reviewed this
attributes such as “malevolent” and “dictatorial” literature.
as universal impediments to outstanding leader- We have learned a great deal during this
ship. Most interesting from a cross-cultural current phase of Project GLOBE with regard
viewpoint are the attributes that in some to the relationships between culture and leader-
societies may be considered to enhance out- ship. The remaining chapters of the book will
standing leadership and in other societies con- provide information with respect to
sidered to impede outstanding leadership.
• Identifying leader characteristics that are
universally accepted and considered effective
CONCLUSION: CULTURAL INFLUENCES across cultures.
ON LEADERSHIP AND GLOBE GOALS • Specifying leadership attributes considered
effective in some cultures, but not others.
The above review of empirical research clearly • Testing the hypothesis that organizations mirror
shows that cultural forces influence many aspects the societies in which they are embedded.
of the leadership phenomena. These include: pro- • Extending the conceptualization of implicit
totypical requisites for leadership positions; priv- leadership theory (ILT) to culturally endorsed
ileges, power, and influence granted to leaders; implicit leadership theory (CLT).
04-House.qxd 3/19/2004 4:08 PM Page 67
• Grouping leadership attributes into meaningful Let the journey continue through the enigmatic
combinations to form primary and global CLT landscape of leadership.
leadership dimensions.
• Determining which dimensions of culture are
related to culturally endorsed CLT leadership SUMMARY: ANTICIPATED
dimensions. THEORETICAL AND PRACTICAL
• Generating profiles of CLT leadership dimen- CONTRIBUTIONS OF PROJECT GLOBE
sions considered effective for specific cultures
and culture clusters. The GLOBE research program is directed
• Determining which cultural dimensions con- toward filling a substantial knowledge gap
sidered together have a maximal influence on concerning the cross-cultural forces relevant
CLT leadership dimensions. to effective leadership and organizational
• Discussing the importance of our findings for practices.
leaders in multicultural environments. It is designed to contribute to the develop-
ment of empirically based cross-cultural leader-
The Project GLOBE journey is far from over. ship and organizational theory by investigating
On our way to validating the integrated theory, the roles of societal and organizational values
additional GLOBE projects, some proposed and and institutionalized practices, organizational
some underway, are intended to contingency variables, and implicit leadership
theories as antecedents to cross-cultural vari-
• Validate relationships as specified in multiple ance in leader behavior, leader influence, leader
propositions within the integrated model effectiveness, and organizational practices and
presented in Chapter 2. performance. GLOBE research is also designed
• Expand the model as suggested by research to contribute to organizational theory and prac-
findings. This has already occurred during our tice by exploring relationships among societal
research and will be an ongoing endeavor. and organizational cultural variables and orga-
• Provide a rich ethnographic description of nizational effectiveness and by conducting
managerial leadership and organizational pro- cross-cultural tests of structural contingency
cesses for countries participating in GLOBE. theory. The research findings will be useful for
Culture specific chapters will be published as resolving several important theoretical social
a companion volume to this book. science issues and for a wide variety of practical
• Examine the effect of violating cultural norms purposes.
relevant to leadership and organizational GLOBE results should also lead to improve-
practices. ments in the practice of leadership. Each leader-
• Determine the extent to which leadership uni- ship profile from geographic areas around the
versals are enacted in a culturally specific or world contains a description of the attributes
universal manner. perceived as facilitating or impeding outstand-
• Identify leader behaviors and organizational ing leadership. We also include information
practices that are universal and those that are specifying culturally endorsed values and com-
culturally specific. mon practices in the societies and organizations
• Determine if the effectiveness of specific lead- studied. Both kinds of information should have
ership behaviors are invariant to culturally substantial practical value for leaders who prac-
endorsed (CLT) leadership prototypes (i.e., can tice management in the cultures studied or who
one lead democratically in a high power dis- interact with individuals from these cultures.
tance culture with strong authoritarian norms?). This information can also supply effective and
valuable case content for leadership training and
We hope that the information presented in career development programs, and for the
this book will challenge readers to develop design of management and leadership education
new theories, hypothesize relationships that we programs to prepare individuals who will man-
have not considered, and conduct research age and lead personnel in cultures other than
related to the influence of culture on leadership. their own.
04-House.qxd 3/19/2004 4:08 PM Page 68
In P. B. Paulus (Ed.), Basic group processes comparative management (Vol. 12, pp. 53–65).
(pp. 9–39). New York: Springer-Verlag. Greenwich, CT: JAI.
Chemers, M. M. (1997). An integrative theory of Dorfman, P. W. (2004). International and cross-
leadership. London: Lawrence Erlbaum. cultural leadership research. In B. J. Punnett &
Chhokar, J. S. (in press). Leadership and culture in O. Shenkar (Eds.), Handbook for international
India: The GLOBE research project. In J. Chhokar, management research (2nd ed., pp. 267–349).
F. Brodbeck, & R. J. House (Eds.), Managerial Ann Arbor: University of Michigan Press.
cultures of the world: GLOBE in-depth studies of Dorfman, P. W., Hanges, P. J., & Dickson, M. W.
the cultures of 25 countries (Vol. 2). Thousand (1998). Challenges in cross-cultural research:
Oaks, CA: Sage. The GLOBE Project—A mini-United Nations
Chowdry, K., & Tarneja, R. (1961). Developing better or the champion Chicago Bulls? Paper presented
managers: An eight-nation study. New York: at the International Conference of Applied
National Industrial Conference Board. Psychology, San Francisco.
Craig, C. S., Douglas, S. P., & Grein, A. (1992). Dorfman, P. W., & Howell, J. P. (1988). Dimensions
Patterns of convergence and divergence among of national culture and effective leadership
industrialized nations: 1960–1988. Journal of patterns. In R. N. Farmer & E. G. McGoun (Eds.),
International Business Studies, 4, 773–787. Advances in international comparative manage-
Croker, J., Fiske, S. T., & Taylor, S. E. (1984). ment (Vol. 3, pp. 127–150). London: JAI Press.
Schematic bases of belief change. In J. R. Eisen Dorfman, P. W., & Howell, J. P. (1997). Managerial
(Ed.), Attitudinal judgment (pp. 197–226). leadership in the United States and Mexico:
New York: Springer-Verlag. Distant neighbors or close cousins? In C. S.
Davis, D. D., & Bryant, J. L. (2003). Influence at Granrose & S. Oskamp (Eds.), Cross cultural
a distance: Leadership in global virtual teams. work groups. Thousand Oaks, CA: Sage.
In W. H. Mobley & P. W. Dorfman (Eds.), Dorfman, P. W., Howell, J. P., Hibino, S., Lee, J. K.,
Advances in global leadership (Vol. 3, pp. 303– Tate, U., & Bautista, A. (1997). Leadership in
340). Oxford: JAI. Western and Asian countries: Commonalities
Davis, H. J., Ming, L. W., & Brosnan, T. F. (1986, and differences in effective leadership processes
August). The Farmer-Richman model: A biblio- across cultures. Leadership Quarterly, 8(3),
graphic essay emphasizing applicability to 233–274.
Singapore and Indonesia. Paper presented at the Dorfman, P. W., & Ronen, S. (1991, August). The
Academy of Management, Chicago. universality of leadership theories: Challenges
Den Hartog, D., House, R. J., Hanges, P. J., Ruiz- and paradoxes. Paper presented at the National
Quintanilla, S. A., & Dorfman, P. W., & GLOBE Academy of Management annual meeting, Miami.
Associates. (1999). Culture specific and cross Earley, P. C. (1984). Social interaction: The frequency,
culturally generalizable implicit leadership use of, and valuation in the United States,
theories: Are attributes of charismatic/trans- England, and Ghana. Journal of Cross-Cultural
formational leadership universally endorsed? Psychology, 15, 477–485.
Leadership Quarterly, 10(2), 219–256. Erez, M. (1997). A culture-based model of work
Deyo, F. C. (1978). The cultural patterning of organi- motivation. In P. C. Earley & M. Erez (Eds.),
zational development: A comparative case study New perspectives on international industrial/
of Thailand and Chinese industrial enterprises. organizational psychology (pp. 193–242). San
Human Organization, 37, 68–72. Francisco: New Lexington Press.
Dorfman, P. W. (1996). International and cross- Farmer, R. N., & Richman, B. M. (1965). Comparative
cultural leadership research. In B. J. Punnett & management and economic progress. Homewood,
O. Shenkar (Eds.), Handbook for international IL: Irwin.
management research (pp. 267–349). Oxford: Gerstner, C. R., & Day, D. V. (1994). Cross-cultural
Blackwell. comparison of leadership prototypes. Leadership
Dorfman, P. W. (1998). Implications of vertical and Quarterly, 5(2), 121–134.
horizontal individualism and collectivism for Gibson, C. (1995). An investigation of gender differ-
leadership effectiveness. In L. C. Change & ences in leadership across four countries. Journal
R. B. Peterson (Eds.), Advances in international of International Business Studies, 26(2), 255–279.
04-House.qxd 3/19/2004 4:08 PM Page 70
Graumann, C. F., & Moscovici, S. (1986). Changing In P. C. Earley & M. Erez (Eds.), New perspec-
conceptions of leadership. New York: Springer- tives in international industrial/organizational
Verlag. psychology (pp. 535–625). San Francisco: New
Hagan, C. M. (1995). Comparative management: Africa, Lexington Press.
the Middle East, and India. (Working Paper). Boca Howell, J. P., & Avolio, B. J. (1993). Transformational
Raton, FL: Florida Atlantic University. leadership, transactional leadership, locus of
Haire, M., Ghiselli, E. E., & Porter, L. (1966). control and support for innovation. Journal of
Managerial thinking: An international study. Applied Psychology, 78, 891–902.
New York: John Wiley. Howell, J. M., & Frost, P. J. (1989). A laboratory
Harbron, J. D. (1965). The dilemma of an elite group: study of charismatic leadership. Organizational
The industrialist in Latin America. Inter-American Behavior and Human Decision Processes, 43(2),
Economic Affairs, 19, 43–62. 243–269.
Hazucha, J. F., Hezlett, S. A., Bontems-Wackens, S., & Hui, C. H., & Tan, G. C. (1999). The moral component
Ronnqvist. (1999). In search of the Euro-manager: of effective leadership: The Chinese case. In
Management competencies in France, Germany, W. Mobley, M. J. Gessner, & V. Arnold (Eds.),
Italy, and the United States (Vol. 1, pp. 267–290). Advances in global leadership (Vol. 1, pp. 249–
Stamford, CT: JAI. 266). Stamford, CT: JAI Press.
Heller, T. (1958). Changing authority patterns: A Hunt, J. G. (1991). Leadership: A new synthesis.
cultural perspective. Academy of Management Newbury Park, CA: Sage.
Review, 10, 488–495. Huo, P. Y., & Randall, D. M. (1991). Exploring
Hofstede, G. (1980). Culture’s consequences: subcultural differences in Hofstede’s survey:
International differences in work-related values. The case of the Chinese. Asia Pacific Journal of
Beverly Hills, CA: Sage. Management, 8, 159–173.
Hofstede, G. (1993). Cultural constraints in manage- Inkeles, A. (1966). The modernization of man. In
ment theories. Academy of Management Executive, M. Weiner (Ed.), Modernization: The dynamics
7(1), 81–94. of growth (pp. 138–150). New York: Basic
Hofstede, G. (2001). Culture’s consequences: Com- Books.
paring values, behaviors, institutions, and orga- Ivancevich, J. M., Schweiger, D. M., & Ragan, J. W.
nizations across nations. Thousand Oaks, CA: (1986). Employee stress, health, and attitudes: A
Sage. comparison of American, Indian and Japanese
Hofstede, G., & Bond, M. H. (1988). The Confucius managers. Paper presented at the Academy of
connection: From cultural roots to economic Management, Chicago.
growth. Organizational Dynamics, 16, 4–21. Jago, A. G., Reber, G., Bohnisch, W., Maczynski, J.,
Hoppe, M. (1993). The effects of national culture on Zavrel, J., & Dudorkin, J. (1993, November).
the theory and practice of managing R&D Culture’s consequences? A seven nation study of
professionals abroad. R&D Management, 23(4), participation. Paper presented at the Proceed-
313–325. ings of the 24th annual meeting of the Decision
House, R. J. (1995). Leadership in the 21st century: Sciences Institute, Washington, DC.
A speculative inquiry. In A. Howard (Ed.), The Jones, A. P., Rozelle, R. M., & Chang, W. C. (1990).
changing nature of work. San Francisco: Jossey Perceived punishment and reward values of
Bass. supervisor actions in a Chinese sample. Psycho-
House, R. J., Hanges, P. J., Ruiz-Quintanilla, S. A., logical Studies, 35, 1–10.
Dorfman, P. W., Javidan, M., Dickson, M., et al. Kakar, S. (1971). Authority patterns and subordinate
(1999). Cultural influences on leadership and behavior in Indian organizations. Administrative
organizations: Project GLOBE. In W. F. Mobley, Science Quarterly, 16, 298–307.
M. J. Gessner, & V. Arnold (Eds.), Advances in Kennis, I. (1977). A cross-cultural study of personal-
global leadership (Vol. 1, pp. 171–233). ity and leadership. Group and Organizational
Greenwich, CT: JAI. Studies, 2(1), 49–60.
House, R. J., Wright, N. S., & Aditya, R. N. (1997). Koene, H., Pennings, H., & Schreuder, M. (1991).
Cross-cultural research on organizational lead- Leadership, culture, and organizational effec-
ership: A critical analysis and a proposed theory. tiveness. In K. E. Clark, M. E. Clark, &
04-House.qxd 3/19/2004 4:08 PM Page 71
D. P. Campbell (Eds.), The impact of leadership. Misumi, J. (1974, July). Action research on the
Greensboro, NC: Center for Creative Leadership. development of leadership, decision-making
Koh, W. L., Terborg, J. R., & Steers, R. M. (1991). processes and organizational performance in a
The impact of transformational leaders on orga- Japanese shipyard. Paper presented at the
nizational commitment, organizational citizen- International Congress of Applied Psychology,
ship behavior, teacher satisfaction and student Liege, Belgium.
performance in Singapore. Paper presented at Misumi, J. (1985). The behavioral science of leadership:
the Academy of Management, Miami, FL. An interdisciplinary Japanese research program.
Korman, A. K. (1966). Consideration, initiating Ann Arbor: University of Michigan Press.
structure and organizational criteria. Personnel Mobley, W. H., & Dorfman, P. W. (Eds.). (2003).
Psychology, 18, 349–360. Advances in global leadership (Vol. 3). Oxford:
Laurent, A. (1983). The cultural diversity of Western JAI.
conceptions of management. International O’Connell, M. S., Lord, R. G., & O’Connell, M. K.
Studies of Management and Organization, (1990). Differences in Japanese and American
13(2), 75–96. leadership prototypes: Implications for cross-
Lee, J. K., Roehl, T., & Choe, S. (2000). What cultural training. Paper presented at the Meeting of
makes management style similar and distinct the Academy of Management, San Francisco, CA.
across borders? Growth, experience and Osland, J. S., Monteze, M. S., & Hunter, L. (1998). A
culture in Korean and Japanese firms. Journal comparative study of managerial styles among
of International Business Studies, 31(4), female executives in Nicaragua and Costa Rica.
631–652. International Studies of Management and
Lee, S. M., & Schwendiman, G. (1982). Japanese Organization, 28(2), 54–73.
management: Cultural and environmental Ouchi, W. (1981). Theory Z: How American business
considerations. New York: Praeger. can meet the Japanese challenge. Reading, MA:
Lindsay, C. P., & Dempsey, B. L. (1985). Experiences Addison-Wesley.
in training Chinese business people to use U.S. Paris, L. (2003). The effect of gender and culture on
management techniques. Journal of Applied implicit leadership theories: A cross-cultural
Behavioral Science, 21, 65–78. study. Unpublished doctoral dissertation, New
Ling, W. Q., & Fang, L. (2003). The Chinese leader- Mexico State University, Las Cruces.
ship theory. In W. H. Mobley & P. W. Dorfman Pascale, R. T., & Athos, A. G. (1981). The art of
(Eds.), Advances in global leadership (Vol. 3, Japanese management. New York: Simon &
pp. 183–204). Oxford: JAI. Schuster.
Lord, R. G., & Maher, K. J. (1991). Leadership and Pelletier, G. (1966). Business management in French
information processing: Linking perceptions and Canada. Business Quarterly—Canada Manage-
performance (Vol. 1). Cambridge, MA: Unwin ment Journal, 31(3), 56–62.
Hyman. Pereira, D. (1987). Factors associated with trans-
Martinez, S. M. (2000). An ethnographic study of formational leadership in an Indian engineering
the Mexican entrepreneur: A configuration of firm. Paper presented at the Administrative
themes and roles impacting managerial leader- Science Association of Canada, Vancouver.
ship in an emerging economy Unpublished Peterson, M. F., Brannen, M. Y., & Smith, P. B.
doctoral dissertation, New Mexico State (1994). Japanese and U.S. leadership: Issues in
University, Las Cruces. current research. In S. B. Prasad (Ed.), Advances
McClelland, D. C. (1961). The achieving society. in international comparative management: A
Princeton, NJ: Van Nostrand. research annual (Vol. 9, pp. 57–82). Greenwich,
Meindl, J. R. (1990). On leadership: An alternative CT: JAI.
to the conventional wisdom. In B. A. Staw (Ed.), Peterson, M. F., & Hunt, J. G. (1997). Overview: Inter-
Research in organizational behavior (Vol. 12, national and cross-cultural leadership research
pp. 159–203). New York: JAI. (Part II). Leadership Quarterly, 8(4), 339–342.
Millman, J. (May 9, 2000). Trade wins: The world’s Peterson, M. F., Smith, P. B., & Peng, T. K. (1995).
new tiger on the export scene isn’t Asian; it’s Japanese and American supervisors of a U.S.
Mexico. Wall Street Journal, p. A1. workforce: An intercultural analysis of behavior
04-House.qxd 3/19/2004 4:08 PM Page 72
meanings. In S. El-Badry, H. Lopez-Cepero, & Scandura, T. A., Von Glinow, M. A., & Lowe, K. B.
T. Hoppe (Eds.), Navigating the Japanese mar- (1999). When East meets West: Leadership
ket: Business and socio-economic perspectives “best practices” in the United States and the
(pp. 229–249). Austin, TX: IC2 Institute. Middle East. In W. Mobley, M. J. Gessner, &
Peterson, M. F., Smith, P. B., & Tayeb, M. H. (1993). V. Arnold (Eds.), Advances in global leadership
Development and use of English versions of (Vol. 1, pp. 235–248). Greenwich, CT: JAI.
Japanese PM leadership measures in electronics Schein, E. H. (1992). Organizational culture and
plants. Journal of Organizational Behavior, leadership: A dynamic view (2nd ed.). San
14(3), 251–267. Francisco: Jossey-Bass.
Poortinga, Y. H., & Malpass, R. S. (1986). Making Schmidt, S. M., & Yeh, R. S. (1992). The structure of
inferences from cross-cultural data. In W. J. leader influence: A cross-national comparison.
Lonner & J. W. Berry (Eds.), Field methods in Journal of Cross-Cultural Psychology, 23(2),
cross-cultural research (pp. 12–46). Beverly 251–264.
Hills, CA: Sage. Shackleton, V. J., & Ali, A. H. (1990). Work-related
Rahim, A. M., Kim, N. H., & Kim, J. S. (1994). values of managers: A test of the Hofstede
Bases of leader power, subordinate compliance, model. Journal of Cross-Cultural Psychology,
and satisfaction with supervision: A cross- 21(1), 109–118.
cultural study of managers in the U.S. and Shamir, B., Zakay, E., Breinin, E., & Popper, M.
S. Korea. The International Journal of Organiza- (1998). Correlates of charismatic leader behav-
tional Analysis, 2(2), 136–154. ior in military units: Subordinates’ attitudes, unit
Ralston, D. A., Gustafson, D. J., Elsass, P. M., characteristics, and superiors’ appraisals of leader
Cheung, F. M., & Terpstra, R. H. (1992). performance. Academy of Management Journal,
Eastern values: A comparison of managers in 41(4), 387–409.
the United States, Hong Kong, and the People’s Shaw, J. B. (1990). A cognitive categorization model
Republic of China. Journal of Applied for the study of intercultural management.
Psychology, 77, 664–671. Academy of Management Review, 15(4),
Redding, S. G. (1990). The spirit of Chinese capital- 626–645.
ism. New York: deGruyter. Shenkar, O., Ronen, S., Shefy, E., & Chow, I. H.
Redding, S. G., & Casey, T. W. (1975, August). (1998). The role structure of Chinese managers.
Managerial beliefs among Asian managers. Human Relations, 51(1), 51–72.
Paper presented at the Annual Meeting of the Sinha, J. B. P. (1980). The nurturant task leader. New
Academy of Management, New Orleans. Delhi: Concept.
Ritchie, R. J. (1999). Reconciling I/O psychology and Sinha, J. B. P. (1984). A model of effective leader-
executive perspectives on global leadership ship styles in India. International Studies of
competencies. In W. Mobley, M. J. Gessner, & Management and Organization, 14(3), 86–98.
V. Arnold (Eds.), Advances in global leadership Sinha, J. B. P. (1994). Cultural imbeddedness and
(Vol. 1, pp. 115–125). Greenwich, CT: JAI. the developmental role of industrial organiza-
Rosch, E. (1975). Universals and cultural specifics tions in India. In H. C. Triandis, M. D. Dunette, &
in human categorization. In R. Brislin & L. M. Hough (Eds.), Handbook of industrial
S. Bochner & W. Lonner (Eds.), Cross-cultural and organizational psychology (2nd ed., Vol. 4,
perspectives in learning (pp. 177–206). Beverly pp. 727–764). Palo Alto, CA: Consulting
Hills, CA: Sage. Psychologists Press.
Ryterbrand, E. C., & Thiagarajan, K. M. (1968). Sinha, J. B. P. (1995). The cultural context of leader-
Managerial attitudes toward salaries as a ship and power. Thousand Oaks, CA: Sage.
function of social and economic development Smith, P. B. (1997). Cross-cultural leadership: A path
(Technical Report No. 24). Rochester, NY: to the goal? In P. C. Earley & M. Erez (Eds.),
University of Rochester, Management Research New perspectives on international industrial/
Center. organizational psychology (pp. 626–639). San
Scandura, T. A., & Dorfman, P. (in press). Francisco: New Lexington Press.
Leadership research in a post-GLOBE world. Smith, P. B. (2003). Leaders’ sources of guidance
Leadership Quarterly Letters. and the challenge of working across cultures. In
04-House.qxd 3/19/2004 4:08 PM Page 73
W. Mobley & P. Dorfman (Eds.), Advances in using Jackson’s return potential model. The
global leadership (Vol. 3, pp. 167–182). Sciences and Engineering, 60, 8–13.
Oxford: JAI. Triandis, H. C. (1993). The contingency model in
Smith, P. B., & Bond, M. H. (1993). Social psychol- cross-cultural perspective. San Diego: Academic.
ogy across cultures: Analysis and perspectives. Triandis, H. C. (1994). Cross cultural industrial
London: Harvester Wheatsheaf. psychology. In H. C. Triandis, M. D. Dunette, &
Smith, P. B., Dugan, S., & Trompenaars, F. (1996). L. M. Hough (Eds.), Handbook of industrial and
National culture and the values of organizational organizational psychology (2nd ed., Vol. 4,
employees: A dimensional analysis across 43 pp. 104–172). Palo Alto, CA: Consulting
nations. Journal of Cross-Cultural Psychology, Psychologists Press.
27(2), 231–264. Trompenaars, F. (1993). Riding the waves of culture:
Smith, P. B., Misumi, J., Tayeb, M. H., Peterson, Understanding cultural diversity in business.
M. F., & Bond, M. H. (1989). On the generality London: Breatley.
of leadership style across cultures. Journal of Ulin, R. D. (1976). African leadership: National goals
Occupational Psychology, 30, 526–537. and the values of Botswana University students.
Smith, P. B., & Peterson, M. F. (1988). Leadership, Comparative Education, 12, 145–155.
organizations and culture: An event manage- Vroom, V. H., & Jago, A. (1988). The new leader-
ment model. Newbury Park, CA: Sage. ship: Managing participation in organizations.
Smith, P. B., & Peterson, M. F. (1995, August). Englewood Cliffs, NJ: Prentice Hall.
Beyond value comparisons: Sources used to give Wade, R. (1990). Governing the market: The role of
meaning to management events in 30 countries. government in East Asian industrialization.
Paper presented at the Academy of Management Princeton, NJ: Princeton University Press.
Annual Meeting, Vancouver, Canada. Wang, Z. M. (1994). Culture, economic reform
Smith, P. B., Peterson, M. F., Misumi, J., & Bond, and the role of industrial and organizational psy-
M. H. (1992). A cross-cultural test of Japanese chology in China. In H. C. Triandis, M. D. Dunette,
PM leadership theory. Applied Psychology: An & L. M. Hough (Eds.), Handbook of industrial
International Review, 42(1), 5–19. and organizational psychology (2nd ed., Vol. 4,
Steers, R. M., Shinn, Y. K., & Ungson, G. R. (1989). pp. 689–725). Palo Alto, CA: Consulting Psycho-
The Chaebol: Korea’s new industrial might. logists Press.
New York: Harper. Weber, M. (1930). The Protestant ethic and the spirit
Stening, B. W., & Wong, P. S. (1983). Australian of capitalism. London: Allen & Unwin.
managers’ leadership beliefs. Psychological Weiss, J. W., & Bloom, S. (1990). Managing in
Reports, 53, 274–278. China: Expatriate experiences and training.
Stogdill, R. M. (1974). Handbook of leadership: Business Horizons, 33, 23–29.
A survey of the literature. New York: Free Whitehill, A. M., & Takezawa, S. (1968). The other
Press. worker. Honolulu: East-West Center Press.
Torres, C. (2000). Leadership style norms among Yukl, G. A. (2002). Leadership in organizations
Americans and Brazilians: Assessing differences (5th ed.). Upper Saddle River, NJ: Prentice Hall.
05-House.qxd 3/19/2004 4:08 PM Page 74
5
THE IMPACT OF SOCIETAL
CULTURE AND INDUSTRY ON
ORGANIZATIONAL CULTURE
Theoretical Explanations
MARCUS W. DICKSON
RENEE S. BESHEARS
VIPIN GUPTA
ecause national culture and industry researchers have begun to understand the need
provide brief descriptions of the theories and evidence suggesting that: (a) values and practices
present examples to better convey the proposed both serve to differentiate between societies and
mechanism in action. Finally, we propose pos- organizations; (b) the values and practices each
sible conditions under which each of the mech- account for unique variance; (c) the values and
anisms is more or less likely to be active than practices scales interact; and (d) the dimension
the others. First, however, we briefly outline of values and practices can be meaningfully
GLOBE’s assumptions about the measurement applied at both levels.
and dimensionality of culture as it is enacted at Unlike the field of personality, in which
the society and organization levels. many researchers have concluded that a specific
taxonomy of dimensions (the five-factor model)
is sufficient and the most parsimonious, we do
GLOBE’S CULTURAL not believe that cultural researchers have yet
DIMENSIONS: VALUES AND PRACTICES reached the point in this field of research where
we can say which taxonomy of cultural dimen-
Throughout this volume, members of the GLOBE sions is most parsimonious. We thus look for-
Project presume that it is useful to conceive of ward to further explorations of the dimensionality
cultural dimensions operating at both the society of culture, and the assessment of systematic
and organization levels of analysis. GLOBE variation in practices and values across levels of
questionnaire items are largely constructed culture, in the hope of reaching such a taxon-
around “quartets” in which the same concept is omy. Prior to that future date, however, we turn
addressed in terms of both practices and values now to our proposals for mechanisms by which
at both the societal and organizational levels societal culture can have its impact on the homol-
(Hanges & Dickson, Chapter 8, this volume). ogous dimensions of organizational culture.
Thus, a deliberate choice was made in the
GLOBE Project to treat cultural dimensions as
homologous across levels of analysis. LEVELS OF IMPACT ON
Other cross-cultural scholars have advocated ORGANIZATIONAL CULTURE
different approaches to this issue. Hofstede, for
example, argues that societies are differentiated There are, of course, myriad factors that affect
by values whereas organizations are differenti- the creation and evolution of organizational cul-
ated by practices, and that a different set of tures: The presence or absence of competitors;
dimensions of practice will be more useful for local, regional, national, and global economic
differentiating organizations than the set of conditions; the nature of the business (manufac-
dimensions of values that best differentiate turing or service or other types of ventures); the
societies (e.g., Hofstede, Neuijen, Ohayv, & nature of the available labor supply, and so
Sanders, 1990; Hofstede & Peterson, 2000). forth. Many of these factors are highly idiosyn-
Given the unique opportunity to collect data cratic and are likely to have unique effects on
from a large number of organizations from a specific kinds of organizations. However, some
large number of countries from every region of factors are likely to have similar effects on a
the world, GLOBE investigators chose to variety of organizations, and thus are likely to
address this question empirically. By collecting be more systematic in their impact. We are
data on both practices and values at both the especially interested in these types of effects,
society and organization levels of analysis and and we broadly categorize them three ways:
by using advanced modeling techniques to test effects in which the nature of the societal culture
the relationships among and between these has an impact on the nature of the organization;
dimensions and levels, we believed that we effects in which the nature of the global indus-
would be able to provide a more definitive try has an impact on the nature of the organiza-
answer to the question of whether both values tion; and effects in which the nature of the
and practices differ meaningfully at both the industry as it is manifested in a given society
society and organization levels. Indeed, Hanges has an impact on the nature of the organization.
and Dickson (Chapter 8, this volume) present We will hereafter in this chapter refer to these
05-House.qxd 3/19/2004 4:08 PM Page 76
types of effects as society effects, industry systems prevent the company from developing
effects, and society-by-industry interaction strategies, structures, or processes that would
effects, respectively. conflict with these assumptions and be ‘antago-
nistic’ to the culture” (p. 398). Gordon goes on
to argue that differences in organizational cul-
Society Effects
ture largely have their basis in differences in
As noted previously, a wealth of literature industries. In support of this argument, Reynolds
indicates that organizations reflect a variety of (1986) conducted a study using 14 cultural
aspects of the societies in which they exist. For dimensions derived from the dimensions most
example, Shane, Venkataraman, and MacMillan commonly used in the organizational culture
(1995) found that differences between societies literature. In surveying respondents from three
on uncertainty avoidance, power distance, and international advanced technology industrial
individualism were related to preferred champi- firms, 14 restaurants in an international fast-
oning behavior in organizations in those societies. food chain, and a computer software and custom
Hayes and Prakasam (1989) found that societal software firm, Reynolds found statistically
differences in power distance predicted the type significant differences for 12 of the 14 cultural
of consultant–client relationship preferred in dimensions. These results suggest a strong indus-
those societies; firms in countries characterized try impact on organizational culture across these
by high power distance preferred relationships three industries.
that were more directive and less collaborative. Analogous to societal effects, in looking for
Indeed, Tata and Prasad (1992) present a model industry effects, one should expect to see (a) sim-
describing the potential negative outcomes that ilarities among most organizations within a given
can accrue to an organization that strays from the industry across cultures and (b) clear evidence that
expectation to match society and industry norms. the origin for the organization-level similarities
Thus, it seems clear that societal charac- resides at the industry level of analysis.
teristics can and do influence the characteristics Unfortunately, there is less empirical evi-
of the organizations within the society. In look- dence for an effect of industry that is consistent
ing for such effects, therefore, one should across societies than there is for societal effects
expect to see (a) similarities among all (or cer- operating on organizational culture. Although
tainly most) organizations within the society, this does not mean that such an effect does not
and (b) a clear origin at the societal level for the exist, we can only speculate as to the size of such
organization-level similarities. effects based on the extant literature. As argued
in the next section of the paper, we expect the
effect of industry to primarily be carried through
Industry Effects
society–industry interactions.
The fact that different industries impose
different demands on organizations is obvious
Society-by-Industry Interaction Effects
and well supported in the organizational strat-
egy literature (Gordon, 1991). Dickson, Aditya, Anyone who has traveled internationally
and Chhokar (2000) argue that “The nature of will recognize that similar industries all over
the industry influences organizational culture the world are enacted differently in different
through the constraints it places on the behavior countries. Utilities are an obvious example, with
of all persons in the organization” (p. 454). some countries having government-run monop-
Further, specific organizational practices olies, others having nongovernmental monopo-
may be essential if an organization hopes to lies, others having state or province or regional
survive in the industry (Burns & Stalker, 1961; monopolies, and still others having varying
Lawrence & Lorsch, 1967). In addition, Gordon forms and levels of competition. Thus, the
(1991) cites Gagliardi (1986) in making similar telecommunications industry may exist in virtu-
points, when he notes that “industry-driven ally every society, but country-level differences
assumptions lead to value systems that are con- in how the industry functions are significant.
sistent with these assumptions, and these value Level and degree of governmental regulation,
05-House.qxd 3/19/2004 4:08 PM Page 77
development of the industry within a society, and in that culture and may in many ways forget
national economic system are just a few of the that other cultures perceive and experience the
other factors that can affect the ways in which a world differently. In the language of cognitive
given industry is enacted in a given society.2 psychology, they develop “shared schemas” or
An interesting example of this type of effect common patterns of thinking, responding, and
is provided by Soeters and Schreuder (1988) in interpreting stimuli they encounter.
their examination of Dutch and U.S. accounting One outcome of this process is that, if people
firms. They found that there were national dif- from a given society share schemas, then the
ferences in the firms’ practices, in that the U.S. organizations within that society are likely to
firms were all lower on uncertainty avoidance have structures and cultures that reflect those
than the Dutch firms, but all of the accounting schemas. We believe that this process is likely to
firms in both societies had uncertainty avoidance lead to both a direct effect of societal culture on
scores that were lower than even the lowest scor- organizational culture, and an effect mediated by
ing country in Hofstede’s (1980) study, suggest- organizational founder or leader characteristics.
ing that there was an industry effect that differed
in its magnitude between societies. Tata and
Direct Effect of
Prasad (1992) provide another, nonempirical,
Society on the Organization
example of this type of interaction. They pro-
pose a model of organizational design for effec- As noted above, the normative prescriptions
tiveness that incorporates a consideration of the and behavioral expectations that characterize a
effects of the market, the current technology, the given culture are likely to lead to patterns of
industrial environment, and the societal culture. thought and behavior that are highly routinized.
More importantly, they discuss the interactions Researchers approaching the study of culture from
of the market, the technology, and the industry a cognitive psychological perspective (e.g., Erez &
with the societal culture as having an impact on Earley, 1993; Lord & Maher, 1991; Smith &
the design of an effective organization. Peterson, 1988; Smith, Peterson, & Wang, 1996)
describe a process whereby sociocultural events
activate particular cultural meaning systems
EXPLANATORY MECHANISMS within a cognitive network, and within a given
culture certain ways of thinking and certain
We now turn to descriptions of the several mech- response patterns to stimuli (such as schemas)
anisms by which we believe industry and society are activated more often than others as a result
effects occur. Although there are myriad poten- of cultural expectations. As a consequence, over
tial explanations, some more macro and some time these schema become readily available and
more micro, we have identified several specific more easily activated than alternatives. In other
theoretical models that we believe help to explain words, “the more often a particular pattern is
the influence process among society, industry, activated in a connectionist network, the more
industry within society, and organization levels efficient its activation is in the future” (Hanges,
of analysis. Specifically, we will refer to: Lord, & Dickson, 2000, p. 143). Eventually,
people will no longer be consciously aware of the
• Cultural immersion theory fact that there are other ways of perceiving the
• Social network theory world or of responding to situations—the cultur-
• Resource dependency theory ally congruent schema have become chronic, and
• Institutional theory as Hanges, Lord, and Dickson (2000) note, it
• Other potential mechanisms, at micro, meso, is difficult to get people off of the familiar path
and macro levels. to consider less familiar (and thus less cogni-
tively available) alternative responses. In this
way, a direct societal effect is created, in that the
Cultural Immersion Theory
shared expectations of the society lead directly to
Most people live their entire lives within a patterns of behavior practices and values that
single societal culture—they become immersed characterize organizations within that society.
05-House.qxd 3/19/2004 4:08 PM Page 78
For example, Lee and Barnett (1997, p. 398), Harrison found that culture did influence this
following earlier theorizing such as that by relationship but that managerial personality did
Emery and Trist (1965), view organizations as not. Again, however, the fact that the societal-
“open systems influenced by the environment.” level sharing of values is treated as sufficient
They operationalized organizational culture in explanation for the pattern of characteristics
terms of perceived distances between pairs of at the organizational level implies a cultural
concepts such as happiness, seniority, success, immersion explanation.
the self, one’s job, supervisor, money and the
names of the countries. They observed signifi-
Effect Mediated by
cant differences among organizational cultures
Founders or Leaders
of a Taiwanese-owned, a Japanese-owned, and
an American-owned bank located in their When the founder of an organization decides
respective countries, but found little difference who to bring in as initial key players in the orga-
between the Taiwanese-owned bank and an nization and how to structure the organization,
American-owned bank located in Taiwan. he or she typically makes decisions on the basis
These findings highlight the influence of of what “makes sense” to them. Organizational
national culture more than that of leader’s leaders later in the organizational life cycle
values in determining organizational culture. In operate in a similar manner. In so doing, they
addition, Meschi and Roger (1994) report a hire people who share their own values, and they
strong linear relationship between perceived create organizational structures that reflect those
distance separating national cultures and that values (Giberson & Resick, 2001). Schneider’s
separating organizational cultures (r = .71). (1987; Schneider, Goldstein, & Smith, 1995)
Further, a great deal of the more atheoretical Attraction-Selection-Attrition (ASA) model states
literature on society effects seems to imply this this explicitly, and argues that an eventual out-
process, though it is rarely explicitly described. come of this process (along with organizational
For example, Schuler and Rogovsky (1998) socialization) is increasing homogeneity within
tested for society effects on the generally accepted organizations on a variety of dimensions, includ-
methods of managing human resources. They ing personality (Schneider, Smith, Taylor, &
found that Hofstede’s (1980) four dimensions of Fleenor, 1998) and values and cultural perceptions
culture were associated with specific compensa- (Haudek, 2001). The ASA model is addressed in
tion practices, though they did not focus on the more detail below.
mechanisms by which that might occur. We suggest that the values systems that the
Similarly, Jang (1997) found that the network founders and leaders hold and the schemas that
structure of shared public relations firms was have become chronic are likely to be reflective
influenced by the differences in national cultures of the values and shared schemas of the larger
(as well as the companies’ business types), such society (Hanges, Lord, & Dickson, 2000; Lord,
that American companies were more centralized Brown, & Freiberg, 1999), although of course
than Japanese firms. Again, however, the expla- there is variation among organizational founders
nation for why this might be is implied more than and leaders in the hierarchy of those values
stated, with the implication being that this is (Locke, 1991). Similarly, the degree to which
reflective of the larger culture: that is, that this common behavioral patterns and cultural schemas
mode of operations “makes sense” to the people have become chronic for specific founders may
in those cultures. As a last of a great many poten- be somewhat individualized, but are likely to
tial examples, Harrison (1993) tested the influ- also reflect the societal culture. Thus, the cul-
ence of societal culture and managerial personality tural immersion process has, in addition to the
on the relation between reliance on accounting direct effect described above, an effect on orga-
performance measures in the evaluative style nizational culture and structure that is mediated
of superiors. Using data from Singapore and by the values of founders and other organiza-
Australia (representing high power distance and tional leaders. In other words, the culture affects
low individualism, and low power distance the founder or leader directly, and the founder
and high individualism cultures, respectively), or leader then affects the organization directly.
05-House.qxd 3/19/2004 4:08 PM Page 79
Schein’s (1992) description of the creation cohesive network provide an effective channel
of organizations culture fits with the perspective for building and transmitting shared perceptions
that leaders have direct effects on organizational and cues. The organizations participating in
culture and structure. He argues that organiza- a cohesive network interact intensively on
tional leaders select people who reflect their learning and innovative activities, and help one
own values and assumptions, and put into place another select and develop shared best practices.
organizational structures and features that The influence mechanism in such a network
reflect their values and assumptions, and it is tends to be cooperative, and there is greater use
these structures and features that encourage the of direct, dyadic ties for interorganizational
development of certain types of culture. In addi- influence. Further, in a cohesive network, non-
tion, Hambrick and Mason (1984) identified business as well as business organizations can
several hypotheses about the demographics of transmit social culture to various member orga-
top management teams and how those demo- nizations. The cohesive relationships also define
graphics would relate to organizational structure the group boundaries, which tend to be stronger
and performance. Miller and Droege (1986) within an identifiable social geographic area
and Miller, Kets de Vries, and Toulouse (1982) than across geographical boundaries, thus sug-
found a relationship between specific leader gesting society effects. From this perspective,
personality components (such as need for bounded information and contact within the
achievement and locus of control) and gross group tend to generate shared values and prac-
measures of organization structure (such as tices among organizations within a society. In a
centralization and formalization). cohesive group, the members tend to demon-
To summarize the preceding discussion, strate strong behavioral conformity and coordi-
societal culture influences perceptions and nated patterns of behavior (Levine & Moreland,
values of societal members, and because leaders 1990). Cohesion allows the group to punish
are members of societal cultures they are thus deviation and reward behaviors consistent with
affected to varying degrees by their societal group norms (Guzzo & Dickson, 1996; Guzzo &
cultures. It is also well established that organi- Shea, 1992). Further, cohesion can facilitate rep-
zational founders and leaders affect their orga- utation enhancement, thus leading to enhanced
nizations’ structures and cultures. Thus, there perceptions of credibility and reliability. A nat-
is significant evidence for both the direct ural outcome of this process is the “bandwagon-
influence of societal culture on organizational effect” in which people and organizations
culture, and the influence of societal culture on “follow the crowd,” such as adopting popular
organizational culture mediated by the values organizational innovations or reading and rely-
of founders or leaders. ing on best-selling business press books whether
they are appropriate to the organizational context
or not. As such, cohesion can be a powerful
Social Network Theory
source of society effects on organizational culture.
Social network theory approaches the A structural equivalence perspective, on the
issues raised in this chapter from the assump- other hand, can often lead to similar predictions
tion that organizations are influenced by the as a cohesion perspective, but through different
social networks in which they are embedded explanatory mechanisms. From this perspective,
and that these social networks constrain the an actor (whether an individual or an organiza-
behaviors and choices of various actors within tion) may play a similar role in a social network
the networks. According to this theory, these as another actor. She or he may provide similar
constraints are either based on cohesion or struc- input to the network while maintaining similar
tural equivalence mechanisms (Burt, 1987), relationships with other network members. These
which are described below. two similar actors may be seen, to some extent, as
Cohesion research suggests that direct inter- interchangeable with each other as they fill the
action with others results in socially constructed same role within the network. In other words,
perceptions (Ibarra & Andrews, 1993). The they are structurally equivalent. The point is that
multiplicity and density of interaction within a if one actor adopts an innovation that could serve
05-House.qxd 3/19/2004 4:08 PM Page 80
to differentiate it from the other actor (making the supply conditions, and other shared experiences,
two nonequivalent), the nonadopter faces pres- all generating industry effects that partially
sures to quickly adopt the innovation to reduce share experiences and behaviors common across
the possibility of a negative comparison with the the society. Thus, a combination of cohesion
adopting actor (Burt, 1983). To illustrate, if two at the society level and structural equivalence at
graduate students work with the same professors the industry level can contribute to interactive
on similar research and one learns a new data society and industry effects.
analytic technique, the other is pressured to learn At the international level, the social network
the new technique as well. Similarly, if two theory predicts some degree of fragmentation
companies produce similar products for a target in the cohesion network (organizations forming
market and one introduces a new feature to their subgroups) because it is difficult for each orga-
products, the other is pressured to adopt that new nization to interact with every other member of
feature as quickly as possible. the network (Berelson & Steiner, 1964). Within
Structural equivalence research thus suggests each subgroup there would be intensive inter-
that the actors sharing similar patterns of role action among organizations with similar values
relationships with others face common role con- and practices. These strong sustained ties gener-
straints and opportunities, and therefore tend to ate cliques: densely connected subgroups of reci-
have similar behaviors associated with manage- procated ties within a social network (Doreian,
ment of those role constraints and opportunities. 1979). The clique-type behavior can be particu-
Structural equivalence generates competition larly strong among organizations belonging to
among organizations because their roles are traditional societies. In these cases, society as well
substitutable due to similarity in their relation- as industry effects can be quite weak because
ships and interaction patterns (Burt, 1987). The there may be different patterns of organizational
intense competition may limit the ability of the behaviors depending on the clique to which the
firms to adopt new social innovations, or alter- firms belong.
natively add to the willingness of the firms to In summation, firms within each industry
be more organic, flexible, and responsive, so tend to have considerable interaction with each
that they can differentiate from one another. In other, and so tend to have similar characteristics,
either case, a structural equivalence perspective because of social cohesion and social influence.
would lend itself to explanation and prediction Social cohesion effects tend to be weaker for
of what we refer to as industry effects. organizations across industries than for organiza-
In addition to competition effects, structural tions within industries. Further, for industries
equivalence may also generate other kinds of across societies, social cohesion is complemented
social influences. Structurally equivalent actors by structural equivalence, which can also generate
may influence each other via indirect ties to similar organizational cultures across different
similar others, such as customers and vendors, societies for firms facing similar network charac-
and thereby adopt similar values and practices. teristics. The degree of cohesion and structural
Because similar environmental conditions induce equivalence varies for each industry and society,
similarity among organizations, this process may presumably leading to society–industry interac-
be termed as ecological influence (Cartwright, tion effects. One of the interesting implications
1965). In addition, structural equivalence may of this process is that effects that appear to be
also operate through social information process- at the societal level may in reality be driven by
ing and social comparison channels, whereby the interorganizational interaction rather than by a
organizations compare with and benchmark cross-level effect of the society.
against firms with equivalent roles. Again, some
degree of similarity in values and practices can
Resource Dependency Theory
emerge as a result of such social comparison,
resulting in knowledge spillovers and organiza- Resource dependency theory is rooted in the
tional learning. Alternatively, the similar network need for organizations to control necessary
relationships may generate similar socialization resources (Pfeffer, 1981; Pfeffer & Salancik,
effects, support service conditions, demand and 1978). It focuses on how the organizations adopt
05-House.qxd 3/19/2004 4:08 PM Page 81
practices and values, and introduce them to their to enhance their trade relationships as a result of
network partners so as to gain access to and best the North American Free Trade Agreement, and
use those resources that will reduce their depen- we see similar shifts among the nations of the
dence on critical contingencies or will otherwise European Union. As new trade regulations pro-
improve their competitive position. Dependency moting trade across these national borders have
theorists view the organization as being embed- taken effect, and trade has indeed increased,
ded within larger interorganizational and societal organizations in these countries find that they
networks and cultural systems. This environ- have grown more dependent upon organizations
ment not only influences the organization’s input in other countries either because of supply or
and output markets, but its beliefs, norms, and sales relationships. Thus, dependency effects
historical traditions. To prosper in their environ- become evident.
ment, organizations must strive to achieve social The changes in the health care industry in the
legitimacy in addition to technological and United States provide a vivid example of depen-
operational efficiency (Abernathy & Chua, 1996). dency theory within a single industry and
Thus, resource dependency theory research country. Health care providers in the United
focuses on (a) the nature of environmental States are dependent on insurance companies
uncertainty and complexity and the impact of for payment, and insurance companies are for-
this uncertainty on the strategic imperatives the profit enterprises. When the American insur-
organization must meet and (b) the organiza- ance industry began to promote managed care
tion’s subsequent efforts to develop an appropri- and the use of health maintenance organizations
ate organizational structure. as a means of reducing costs, most health care
Bartol and Martin’s (1988, 1989, 1990) providers (both individuals and organizations)
dependency theory is largely a more narrowly were compelled to participate in the system or
focused, microlevel version of resource depen- lose access to patients and the accompanying
dency theory. Dependency theory builds on two revenue stream. Health care providers thus went
basic assumptions about managerial behavior: along with managed care reforms despite the
(a) managers are dependent, to varying degrees, fact that many concerns about compromises in
upon their subordinates; and (b) managers use quality of care were raised from the beginning
pay and other valued resources to manage their (cf. Patient Right to Know Act of 1996, 1996;
dependence upon their subordinates (Klein Views of Healthcare Providers, 1993).
et al., 2000). Dependency theory and resource Although there is a fair amount of research
dependency theory are largely homologous: from a dependency theory perspective, we see
They propose that similar dynamics occur the bulk of dependency theory’s arguments as
between managers and subordinates within being incorporated in institutional theory. We
organizations as well as among organizations thus now turn to institutional theory and will
and their suppliers, vendors, and consumers. focus in much more detail on that perspective.
Thus, individual managers make decisions
based on dependencies and threats to depen-
Institutional Theory
dency relationships in much the same way as
organizations and organizational subunits do. Institutional theory is in some ways similar
Some examples of dependency relationships to dependency theory, but focuses primarily on
and expected outcomes might facilitate under- the organization’s institutional environment
standing of the resource dependency model. rather than on tangible resources exclusively.
First, countries that have formalized relation- This environment consists of the elaboration of
ships as trading partners are likely to become rules, practices, symbols, beliefs, and normative
more similar to each other in terms of their requirements to which individual organizations
behaviors and structures, even if the deeply held must conform to receive support and legitimacy
values driving those behaviors do not coalesce (DiMaggio & Powell, 1983; Meyer & Rowan,
(see Chapter 20, this volume, by Brodbeck 1977). Institutional theory can best be explained
et al.). Thus, we see Canadian, Mexican, and as a theory of legitimacy-seeking. Organizations
American organizations making strategic shifts respond to pressures from society and other
05-House.qxd 3/19/2004 4:08 PM Page 82
U.S. health care providers, again showing the organizations may be more commonplace in
similarity of prediction between resource depen- societies in which bureaucratic governments
dency theory and institutional theory. exist, or in cultures in which power distance and
Coercive isomorphism also includes the uncertainty avoidance are high. What is per-
pressures put on an organization to conform to ceived as being cutting edge may also reflect
the cultural norms of the society in which they overall societal or national cultural values.
operate. This type of influence seems clearly Some cultures may be more apt to follow trends
to be operating at the society effects level. This and fads than others, and this would influence
type of influence is evident in different pay the degree to which organizations adopt these
structures, which may reflect cultural differ- practices.
ences in power distance, uncertainty avoidance, It follows from the description above that
and individuality. In addition, the work–family mimetic isomorphism would primarily be active
accommodation issues cited above (Goodstein, at the society and the society-by-industry inter-
1994; Klein, et al., 2000) would likely be exam- action effect levels. We would trace the basis
ples of this form of isomorphism, in that for this argument back to Katz and Kahn (1966).
generalized society-level cultural expectations According to Katz and Kahn, people are over-
probably drive these organizational practices. whelmed with information and of necessity
This form of isomorphism seems likely to us develop systems by which they can filter incom-
to contribute to society effects, industry effects, ing information. The probable outcome of such
and society-by-industry interaction effects on filtration is that events that occur and data
organizational values and practices. Interestingly, generated within the boundaries of their system
the more recent movements in resistance to become much more salient than events and
managed care can be thought of in this context: information generated elsewhere. Dickson,
Health care providers eventually grew weary of Aditya, and Chhokar (2000) argue that for most
being compelled to provide what they perceived people the most cognitively available compara-
as substandard care (i.e., not meeting the legiti- tors (i.e., the sources of information most likely
mate expectations of the broader society) and, to be within the filtration system) are organiza-
so as to restore perceptions of legitimacy, have tions within their own society and industry. This
in some cases left the managed care system and occurs because most people are significantly
in others have attempted to band together to less knowledgeable about organizations in
resist the pressures of managed care from the other countries or even in other industries
insurance industry. within their own country. Of course, some
Mimetic isomorphism generally involves a people have broad knowledge of other societies
transformation that takes place in the face of and industries than their own, but even for these
uncertainty, if an organization models itself few individuals, the level of perceived similarity
after another, more successful, organization. of organizations within their own national bor-
Organizations usually mimic other organiza- ders may lead them to more easily attend to and
tions that are either in their field and perceived recall, and thus be influenced by, organizations
as more legitimate, outside their field (and per- “closer to home” (Festinger, 1954).
ceived as legitimate) but similar in complexity, Finally, educational or professional pressures
or those organizations that are perceived as to conform to a set of rules and norms charac-
being “on the cutting edge” (DiMaggio & terize normative isomorphism. Educational or
Powell, 1983). If organizations mimic other training programs and professional societies or
organizations in the same field, an industrial associations influence professionals operating
culture tends to emerge in which organizations within organizations. This type of pressure can
within a given industry behave similarly in the be powerful if these professional entities push
manner that is perceived as most successful. If for state-sanctioned actions, such as licensing
organizations mimic other organizations similar or certification. On a less severe level, norma-
in complexity or organizations seen as being on tive isomorphism can dictate how certain orga-
the cutting edge the influence of national culture nization members will speak, behave or dress
may emerge. For example, highly bureaucratic (Ouchi, 1980). This type of isomorphism can
05-House.qxd 3/19/2004 4:08 PM Page 84
again operate at the different levels, either is likely to vary from organizational relationship
national or industrial. The degree to which to relationship. This inconsistency of behavior
normative influences will affect organizations patterns does not imply that the culture is unclear
will depend upon the number of professionals or unstable, but rather that the culture includes the
from the same field that exist within a particular idea that organizations engage differentially with
society or industry. For example, medical doc- other organizations on the basis of their relative
tors who comprise a large portion of the work- dominance.
force influence the health care industry. Their In societies high in uncertainty avoidance, on
high involvement in associations and societies the other hand, there may also be a commonal-
can pressure organizations to adopt certain ity of coercive isomorphism, but more so in
practices that the profession as a whole deems response to the values of the greater society.
most appropriate, such as continuing education Institutional theory would predict that organiza-
or licensing. This type of influence can be tem- tions that are viewed as legitimate by the greater
pered, however, by the broader societal culture. society are going to prosper, whereas organiza-
Although the professional associations exert pres- tions that do not adopt and reflect the societal
sures on organizations, the norms of the national culture’s values will not.
or societal culture will be operating within the Even more complex, there may be a three-
professional societies. way interaction between the cultural dimension,
the type of isomorphism, and the level of cul-
tural influence. The previous description of
Complexities of Isomorphic Pressures
organizations high in power distance provides
Although institutional theory and isomor- an example. Although the dependency portion
phism can operate at different cultural levels, as of coercive isomorphism may be operating if
we have noted above, it is most likely that the power distance is high, the cultural influence
influences at each level interact with one another may stem from the specific industry within
so that ultimately we see society-by-industry which the organization operates. Thus, although
interaction effects as a result of isomorphic pres- the cultural influence might be expected to be
sure. Although industries may operate in similar great due to high power distance, the specific
ways regardless of the country in which they cultural effect may be dependent upon what is
exist, it is impossible to completely override viewed as legitimate within the industry. For
the cultural norms that exist within a particular example, the power distance norm may gener-
society. Because culture essentially exists within ally be enacted in a given society through the
the minds of people and is manifested in behav- lack of superiors consulting subordinates when
ior, organizational members will adopt norms making decisions. This approach is likely to be
and standards that are most familiar to them. acceptable and workable in industries in which
Whereas different types of isomorphism may leadership arises through demonstrated exper-
operate at various levels of cultural influence tise at lower levels in the hierarchy (e.g., finan-
(national or societal versus industrial), specific cial services). However, it may not be workable
cultural dimensions may drive organizational in a dynamic and technology-driven industry
isomorphism differentially. For example, in a high such as telecommunications, in which upper-
power distance societal culture, the most preva- level managers may have little if any under-
lent form of institutional influence may be coer- standing of the technologies they supervise,
cive isomorphism, specifically the dependency yielding a situation in which it is imperative for
aspect of coercive isomorphism. Because societies the manager to seek opinions and suggestions
high in power distance perceive the wielding from lower-level employees. In such a case, we
of influence and domination as appropriate would expect other manifestations of the power
behavior, organizations will fall into either a distance norm to emerge in that industry. We
powerful or a dependent position relative to thus see that, although institutional theory can
other specific organizations. However, whether provide a framework to interpret and predict the
they are dependent or dominant will depend upon transmission of cultural values and establish-
the comparator, and the organizational behavior ment of common organizational behavioral
05-House.qxd 3/19/2004 4:08 PM Page 85
ISO 9000 certification because their primary culture, we conclude with the argument that to
customer instituted a policy that all suppliers truly understand when and how these mecha-
must be ISO 9000 compliant. nisms have their effects requires both culturally
generalizable and culturally specific perspec-
tives. The culturally generalizable approach is
Convergence of Expected Outcomes
necessary to identify whether one or more of the
Finally, one of the primary points that we general principles we have outlined here are in
want to make in this section is that many of the general more commonly active than are others,
theoretical perspectives presented lead to similar whereas the culturally specific perspective tells
expected outcomes, though perhaps through dif- us about the mechanisms active in specific situ-
fering processes. For example, as we noted above, ations. In addition, taking both the “outsider”
predictions building from dependency theory and “insider” senses leads us to conclude that
and institutional theory are inevitably going to be sometimes important aspects of a culture are not
similar, given that they each focus, at least in part, evident to insiders, but those aspects of culture
on relationships between organizations in which may not be interpretable by outsiders. For these
one organization is dependent upon, and thus reasons, in GLOBE we have taken both per-
influenced by, another. Cultural immersion theory, spectives. This book presents our attempts to
institutional theory, and social networks theory uncover general principles and cross-cultural
all lead to the conclusion that organizations with findings. The companion volume (edited by
whom other organizations have frequent contact Chhokar, Brodbeck, and House) allows GLOBE
and who are seen as highly effective by other country teams for each culture, made up of both
organizations will be emulated by those other cultural insiders and outsiders, to present their
organizations. The level of conscious awareness interpretations of their target culture and to
of the process by the individuals involved and the articulate how a wide variety of factors, includ-
focus on action at the individual or organizational ing those identified in this chapter, play out
level may differ, but the ultimate outcome in that specific context. We believe that the
expected is quite similar. This is perhaps one rea- interplay between outsider and insider makes
son why there has not been a wealth of research GLOBE unique in its approach and explanatory
on identifying which specific mechanisms are power for addressing these important issues.
actually in play—the outcomes are often similar,
regardless of the mechanism, and so there is little
to be gained by teasing apart the processes by
ENDNOTES
which the outcome occurs. Nonetheless, we
1. Following the GLOBE approach outlined
believe that it does matter whether organizations
throughout this volume, we consider culture at both
consciously choose to emulate or acquiesce to
the societal and organizational levels to include both
other organizations or do so based on decisions of
common practices and shared values.
individuals influenced at nonconscious levels, and
2. Of course, the influence that societal culture
that it does matter whether the decision is made
may exert on industry-level variables of a country
out of a desire for enhanced performance or a fear
only serves to make the situation more complicated
of losing business. Thus, as Brodbeck and col-
(e.g., McClelland, 1961).
leagues (Chapter 20, this volume) point out, even
though the same behaviors may ultimately occur
at the societal level, the reasons behind them REFERENCES
may differ. It is important to understand those
reasons if we are to truly understand organiza- Abernathy, M. A., & Chua, W. F. (1996). A field study
tional and individual behavior and values. We of control system “redesign”: The impact of insti-
cannot simply ignore the black box because the tutional processes on strategic choice. Contem-
outcomes predicted by the different theoretical porary Accounting Research, 13, 569–595.
approaches are likely to be similar. Bartol, K. M., & Martin, D. C. (1988). Influences
Having discussed these potential mechanisms on managerial pay allocations: A dependency
by which societal culture influences organizational perspective. Personnel Psychology, 41, 361–378.
05-House.qxd 3/19/2004 4:08 PM Page 88
Bartol, K. M., & Martin, D. C. (1989). Effects of Festinger, L. (1954). A theory of social comparison
dependence, dependency threats, and pay secrecy processes. Human Relations, 7, 117–140.
on managerial pay allocations. Journal of Gagliardi, P. (1986). The creation and change of
Applied Psychology, 74, 105–113. organizational cultures: A conceptual framework.
Bartol, K. M., & Martin, D. C. (1990). When politics Organization Studies, 7, 117–134.
pays: Factors influencing managerial compensa- Giberson, T. R., & Resick, C. (2001, April). Transferring
tion decisions. Personnel Psychology, 43, 599–614. leader values: Using ASA to understand organiza-
Bechtel, W., & Abrahamsen, A. (1996). Connec- tional culture creation. In M. W. Dickson (Chair),
tionism and the mind: An introduction to paral- The attraction-selection-attrition model: Current
lel processing in networks. Cambridge, MA: research and theory. Symposium conducted at the
Blackwell. meeting of the Society for Industrial/Organizational
Berelson, B., & Steiner, G. (1964). Human behavior: Psychology, San Diego, CA.
An inventory of scientific findings. New York: Goodstein, J. D. (1994). Institutional pressures and
Harcourt, Brace, and World. strategic responsiveness: Employer involvement
Burns, T., & Stalker, G. M. (1961). The management of in work-family issues. Academy of Management
innovation. London: Tavistock Publications Ltd. Journal, 37, 350–382.
Burt, R. S. (1983). Toward a structural theory of Gordon, G. G. (1991). Industry determinants of
action. New York: Academic. organizational culture. Academy of Management
Burt, R. S. (1987). Structural contagion and innova- Review, 16, 396–415.
tion: Cohesion vs. structural equivalence. American Guilhon, A., Martin, J., & Weill, M. (1998). Quality
Journal of Sociology, 92, 1287–1335. approaches in small or medium-sized enter-
Cartwright, D. (1965). Influence, leadership, control. prises: Methodology and survey results. Total
In J. G. March (Ed.), Handbook of organizations Quality Management, 9(8), 689–701.
(pp. 1–47). Chicago: Rand McNally. Guzzo, R. A., & Dickson, M. W. (1996). Teams in
Dickson, M. W., Aditya, R. N., & Chhokar, J. S. organizations: Recent research on performance
(2000). Definition and interpretation in cross- and effectiveness. Annual Review of Psychology,
cultural organizational culture research: Some 47, 307–338.
pointers from the GLOBE research program. In Guzzo, R. A., & Shea, G. P. (1992). Group perfor-
N. Ashkanasy, C. Wilderom, & M. Petersen (Eds.), mance and intergroup relations in organiza-
Handbook of organizational culture and climate, tions. In M. D. Dunnette & L. M. Hough (Eds.),
(pp. 447–464). Thousand Oaks, CA: Sage. Handbook of industrial and organizational psy-
DiMaggio, P. J., & Powell, W. W. (1983). The iron chology (Vol. 1, pp. 269–313). Palo Alto, CA:
cage revisited: Institutional isomorphism and Consulting Psychologists Press.
collective rationality in organizational fields. Hambrick, D. C., & Mason, P. A. (1984). Upper ech-
American Sociological Review, 48, 147–160. elons: The organization as a reflection of its top
Doreian, P. (1979). Mathematics and the study of social managers. Academy of Management Review, 9,
relations. London: Weidenfeld & Nicolson. 193–206.
Earley, P. C., & Erez, M. (1997). Introduction. In Hanges, P. J., Lord, R. G., & Dickson, M. W. (2000).
P. C. Earley & M. Erez (Eds.), New perspectives An information processing perspective on leader-
on international industrial/organizational psy- ship and culture: A case for connectionist archi-
chology (pp. 1–10). San Francisco: New Lexington tecture. Applied Psychology: An International
Press. Review, 49, 133–161.
Emery, F. E., & Trist, E. L. (1965). The causal tex- Harrison, G. L. (1993). Reliance on accounting per-
ture of organizational environments. Human formance measures in superior evaluative style:
Relations, 18, 21–32. The influence of national culture and personal-
Erez, M., & Earley, P. C. (1993). Culture, self-identity, ity. Accounting, Organizations and Society,
and work. New York: Oxford University Press. 18(4), 319–339.
Farh, J., Leong, F. T. L., & Law, K. S. (1998). Haudek, G. (2001, April). A culture variance test
Cross-cultural validity of Holland’s model in of the ASA homogeneity hypothesis. In M. W.
Hong Kong. Journal of Vocational Behavior, Dickson (Chair), The attraction-selection-
52, 425–440. attrition model: Current research and theory.
05-House.qxd 3/19/2004 4:08 PM Page 89
Symposium conducted at the meeting of the Leong, F. T. L., Austin, J. T., Sekaran, U., &
Society for Industrial/Organizational Psychology, Komarraju, M. (1998). An evaluation of the
San Diego, CA. cross-cultural validity of Holland’s theory:
Hayes, J., & Prakasam, R. (1989). Culture: The Career choices by workers in India. Journal of
efficacy of different modes of consultation. Vocational Behavior, 52, 441–455.
Leadership and Organization Development Levine, J. M., & Moreland, R. L. (1990). Progress
Journal, 10, 24–32. in small group research. In M. R. Rosenzweig &
Hofstede, G. (1980). Culture’s consequences: Inter- L. W. Porter (Eds.), Annual review of psychol-
national differences in work-related values. ogy (Vol. 41, pp. 585–634). Palo Alto: Annual
Beverly Hills, CA: Sage. Reviews.
Hofstede, G., Neuijen, B., Ohayv, D. D., & Sanders, G. Locke, E. A. (1991). The motivation sequence, the
(1990). Measuring organizational cultures: A qual- motivation hub, and the motivation core.
itative and quantitative study across twenty cases. Organizational Behavior and Human Decision
Administrative Science Quarterly, 35, 286–316. Processes, 50, 288–299.
Hofstede, G., & Peterson, M. F. (2000). Culture: Lord, R. G., Brown, D. J., & Freiberg, S. M. (1999).
National values and organizational practices. In Understanding the dynamics of leadership: The
N. Ashkanasy, C. Wilderom, & M. Petersen (Eds.), interaction of self-concepts in the leader/
Handbook of organizational culture and climate follower relationship. Organizational Behavior
(pp. 401–415). Thousand Oaks, CA: Sage. and Human Decision Processes, 78, 167–203.
Holland, J. L. (1996). Integrating career theory and Lord, R. G., & Maher, K. J. (1991). Leadership and
practice: The current situation and some poten- information processing: Linking perceptions
tial remedies. In M. L. Savickas & W. B. Walsh and performance. Boston: Unwin Hyman.
(Eds.), Handbook of career counseling theory McClelland, D. C. (1961). The achieving society.
and practice (p. 1–11). Palo Alto, CA: Davies- Princeton, NJ: Van Nostrand.
Black Publishing. Meschi, P., & Roger, A. (1994). Cultural context and
Ibarra, H., & Andrews, S. B. (1993). Power, social social effectiveness in international joint ven-
influence, and sense making: Effects of network tures. Management International Review, 34(3),
centrality and proximity on employee perceptions. 197–215.
Administrative Science Quarterly, 38, 277–303. Meyer, J. W., & Rowan, B. (1977). Institutional
Jang, H. (1997). Cultural differences in an interorga- organizations: Formal structure as myth and
nizational network: Shared public relations ceremony. American Journal of Sociology, 83,
firms among Japanese and American compa- 340–363.
nies. Public Relations Review, 23(4), 327–341. Meyer, J. W., & Rowan, B. (1991). Institutional orga-
Katz, D., & Kahn, R. L. (1966). The social psychology nizations: Formal structure as myth and cere-
of organizations. New York: John Wiley & Sons. mony. In W. W. Powell & P. J. DiMaggio (Eds.),
Klein, K. J., Berman, L. M., & Dickson, M. W. (2000). The new institutionalism in organizational
May I work part-time? An exploration of pre- analysis (pp. 41–62). Chicago: The University of
dicted employer responses to employee requests Chicago Press.
for part-time work. Journal of Vocational Miller, D., & Droege, C. (1986). Psychological and
Behavior, 57, 85–101. traditional determinants of structure. Adminis-
Lawler, E. E., III. (1986). High-involvement manage- trative Science Quarterly, 31, 539–560.
ment: Participative strategies for improving Miller, D., Kets de Vries, M. F., & Toulouse, J.
organizational performance. San Francisco: (1982). Top executive locus of control and its
Jossey-Bass. relationship to strategy-making, structure, and
Lawrence, P. J., & Lorsch, J. W. (1967). Organization environment. Academy of Management Journal,
and environment. Boston: Harvard Business 25, 237–253.
School, Division of Research. Oleski, D., & Subich, L. M. (1996). Congruence and
Lee, M., & Barnett, G. A. (1997). A symbols-and- career change in employed adults. Journal Of
meaning approach to the organizational cultures Vocational Behavior, 49, 221–229.
of banks in the United States, Japan, and Taiwan. Ouchi, W. G. (1980). Market, bureaucracies and clans.
Communication Research, 24(4), 394–412. Administrative Science Quarterly, 25, 124–141.
05-House.qxd 3/19/2004 4:08 PM Page 90
Patient Right to Know Act of 1996, 104th Cong., 2d Smith, E. R. (1996). What do connectionism and
Sess. (1996) (testimony of Mark E. Rust). social psychology offer each other? Journal of
Pfeffer, J. (1981). Power in organizations. Marshfield, Personality and Social Psychology, 70, 893–912.
MA: Pitman. Smith, P. B., & Peterson, M. F. (1988). Leadership,
Pfeffer, J., & Salancik, G. R. (1978). The external organizations, and culture: An event manage-
control of organizations: A resource dependence ment model. Newbury Park, CA: Sage.
perspective. New York: Harper & Row. Smith, P. B., Peterson, M. F., & Wang, Z. M. (1996). The
Reynolds, P. D. (1986). Organizational culture as manager as mediator of alternative meanings: A
related to industry, position, and performance: A pilot study from China, the USA, and U.K. Journal
preliminary report. Journal of Management of International Business Studies, 27(1), 115–137.
Studies, 23(3), 333–345. Soeters, J., & Schreuder, H. (1988). The interaction
Schein, E. H. (1992). Organizational culture and lead- between national and organizational cultures in
ership (2nd ed.). San Francisco: Jossey-Bass. accounting firms. Accounting, Organizations
Schneider, B. (1987). The people make the place. and Society, 13, 75–85.
Personnel Psychology, 40, 437–453. Strauss, C., & Quinn, N. (1997). A cognitive theory of
Schneider, B., Goldstein, H. W., & Smith, D. B. (1995). cultural meaning. New York: Cambridge
The ASA framework: An update. Personnel University Press.
Psychology, 48, 747–773. Tata, J., & Prasad, S. (1992). Optimum production
Schneider, B., Smith, D. B., Taylor, S., & Fleenor, J. process, national culture, and organization
(1998). Personality and organizations: A test of design. European Business Review, 92(1), VII.
the homogeneity of personality hypothesis. Tolbert, P. S., & Zucker, L. G. (1983). Institutional
Journal of Applied Psychology, 83, 462–470. sources of change in the formal structure of orga-
Schuler, R. S., & Rogovsky, N. (1998). Understanding nizations: The diffusion of civil service reform,
compensation practice variations across firms: 1880–1935. Administrative Science Quarterly,
The impact of national culture. Journal of 28, 22–39.
International Business Studies, 29(1), 159–177. Townsend, R. (1970). Up the organization: How to
Scott, W. R., & Meyer, J. W. (1991). The rise of stop the corporation from stifling people and
training programs in firms and agencies: An strangling profits. New York: Fawcett Books.
institutional perspective. In L. L. Cummings & Views of Healthcare Providers, 103rd Cong., 1st Sess.
B. M. Staw (Eds.), Research in organizational (1993) (testimony of James S. Todd).
behavior (Vol. 13, pp. 297–326). Greenwich, Wrigley, A. (1997, December 16). Although only
CT: JAI Press. about 12% of tier 1 suppliers to the big three
Shane, S., Venkataraman, S., & MacMillan, I. (1995). automakers were QS 9000 certified at the begin-
Cultural differences in innovation champion- ning of 1997, the vast majority are QS 9000
ing strategies. Journal of Management, 21(5), certified as of the end of 1997. American Metal
931–952. Market Metal Technology Quarterly, 2A.
06-House.qxd 3/19/2004 4:08 PM Page 91
PART III
RESEARCH METHODOLOGY
PAUL J. HANGES
he chapters in this section focus on the Following the GLOBE conceptual model, we
of scales (i.e., societal, organizational, and with possible analysis at four different levels:
leadership). individual, organizational, industry, and soci-
Psychometric properties of the GLOBE etal. That is, individuals work within organiza-
scales were assessed and documented in tions, which in turn are found within three
Chapter 7. We found that our scales, on average, industries (financial, food processing, and
exceeded professional standards with regard telecommunication). Industries in turn are found
to properties such as dimensionality, interitem within societal cultures. The GLOBE project,
reliability, interrater agreement, and aggregata- like the Hofstede (1980) study that preceded it,
bility. This chapter also discusses and computes uses societal culture rather than individuals as
measures to deal with potential cultural response units of analysis. We also use organizational
bias contamination. It also summarizes the initial culture as units of analysis. Thus, unlike the
work that provides evidence for the construct Hofstede study, the research design used in
validity of the GLOBE scales. Construct validity GLOBE allowed exploration of our hypotheses
provides important information about the nature at multiple levels of analysis.
of the constructs measured by the GLOBE scales It should be noted that we make no assump-
as well as possible boundary conditions of these tions that relationships among variables at the
scales. ecological level (organizations and societies)
To further demonstrate construct validity, also are valid at the individual level. Never-
evidence presented in Chapter 8 supports the theless, we are fortunate to be able to determine
generalizability of the GLOBE constructs. Some the extent to which relationships hold at multi-
of this research required obtaining archival data ple levels—Leung and Bond (1989) refer to
to develop alternative nonquestionnaire-based these as “strong etic relationships.” Although
measures of the GLOBE societal cultural con- we do not dwell on the intracultural analyses,
structs. This work is fully described by Gupta, these results provide additional evidence to the
Sully de Luque, and House in Chapter 9. These validity of our overall findings described at the
authors provide strong evidence for convergent societal level.
and discriminant validity. Further, Javidan and Taken together, the information provided
Hauser in Chapter 7 also used archival data to in Chapters 7, 8, 9, and Appendix C provide
assess the implications of the GLOBE scale empirical support for the assertion that the
scores. constructs measured by the GLOBE scales
In Chapter 10, Gupta and Hanges describe generalize beyond the sample from which the
the process of grouping the GLOBE societies data were obtained (i.e., middle manager from
into a smaller cluster of societies we label a one of three industries), the method used to
culture cluster. These culture clusters are based collect these data (i.e., self-report), and the
on an assessment of commonly used indicators “sets of operations” applied on these data (i.e.,
such as language, history, religion, and geogra- aggregation). The findings reflect the broader
phy. Gupta and Hanges empirically demon- societal and organizational cultures under
strate the validity of these clusters using the study.
GLOBE societal culture dimensions. These
authors also report on a second type of group-
ing in which the GLOBE societies were sorted REFERENCES
in clusters on the basis of similarity in physical
climate experienced by people in those Boring, E. G. (1945). The use of operational definitions
societies. Both societal clustering procedures in science. Psychological Review, 52, 243–245.
are used and discussed in the chapters appear- Bridgman, P. W. (1927). The logic of modern physics.
ing in Part IV of this book. New York: Macmillan.
Last, Hanges, Dickson, and Sipe in Chapter Hofstede, G. (1980). Culture’s consequences: Inter-
11 explain the rationale underlying the statisti- national differences in work-related values.
cal techniques we used to test the GLOBE London: Sage.
hypotheses. It is important to reiterate that Hofstede, G. (2001). Culture’s consequences:
the GLOBE data is multilevel in structure Comparing values, behaviors, institutions and
06-House.qxd 3/19/2004 4:08 PM Page 93
organizations across nations (2nd ed.). Thousand Shadish, W. R., Cook, D. T., & Campbell, T. D. (2002).
Oaks, CA: Sage. Experimental and quasi-experimental designs
Leung, K., & Bond, M. H. (1989). On the empirical for generalized causal inference. Boston, MA:
identification of dimensions for cross-cultural Houghton Mifflin.
comparisons. Journal of Cross-Cultural Psycho- United Nations Development Program. (1998).
logy, 20, 133–151. Human development report. New York: Oxford
Messick, S. (1981). Constructs and their vicissitudes University Press.
in educational and psychological measurement. van de Vijver, F., & Leung, K. (1997). Methods and
Psychological Bulletin, 89, 575–588. data analysis for cross-cultural research.
Robinson, W. S. (1950). Ecological correlations and Thousand Oaks, CA: Sage.
the behavior of individuals. American Sociological World Economic Forum. (1998). The global compet-
Review, 15, 351–357. itiveness report. Geneva, Switzerland: Author.
06-House.qxd 3/19/2004 4:08 PM Page 94
06-House.qxd 3/19/2004 4:08 PM Page 95
6
RESEARCH DESIGN
ROBERT J. HOUSE
PAUL J. HANGES
95
06-House.qxd 3/19/2004 4:08 PM Page 96
datasets used to test the GLOBE propositions. the samples included demographic information,
Finally, we discuss the implications of the some did not. The reasons for this varied.
GLOBE research design for handling a number Several of our country co-investigators (CCIs)
of common problems found in the cross-cultural did not include this demographic information in
literature. their version of the questionnaire. The decision
was a conscious omission on the part of the
CCIs in that they believed that asking demo-
SAMPLING DESIGN graphic information would substantially lessen
the response rate, something akin to asking
We used a stratified sampling strategy in which about personal income in questionnaires in the
four different strata (i.e., individuals, organiza- United States. In other cases, respondents did
tions, industries, and societies) were included. not complete the demographic section of the
In particular, the GLOBE Phase 2 sampling questionnaire. A complete set of demographic
strategy required that data from each society data was collected from more than 8,000
met the following criteria: (a) respondents had respondents. We can use this subsample to
to be middle managers, (b) multiple respon- estimate some of the characteristics of the
dents had to be obtained from organizations, GLOBE respondents. In particular, approxi-
(c) two or more organizations had to be mately 74.8% of our respondents were men.
obtained from two of three types of industries Further, all respondents indicated that they had
(financial, food processing, and telecommuni- an average full-time work experience of 19.2
cation), and (d) at least two industries had to be years, of which an average of 10.5 years were
obtained for each society. Half of the respon- spent as managers. They indicated that they had
dents from a given organization completed one worked for their current organizations an aver-
version of the GLOBE culture and leadership age of 12.2 years. Finally, 51.4% indicated that
questionnaire (labeled version Alpha). The they had worked for a multinational corpora-
other half completed a second version of the tion, 44.0% indicated that they belonged to pro-
GLOBE culture and leadership questionnaire fessional associations, and 43.8% indicated that
(labeled version Beta). By administering these they participate in industrial or trade associa-
questionnaires to separate samples of middle tion activities.
managers from the same organization or
society, we minimized or even eliminated com-
Sampling of Organizations and Industries
mon source response bias concerning societal
and organizational phenomena. The Phase 2 data were obtained from middle
managers employed in 951 separate organiza-
tions. Only corporations headquartered in
Sampling of Individuals
the host cultures were included in our sample.
A total of 17,370 middle managers from 951 Therefore, we deliberately excluded from our
organizations in three industries completed the sample foreign multinational corporations.
culture and leadership questionnaires in both Multinational corporations were excluded
Phases 1 and 2 of GLOBE. Specifically, 1,943 because their members would be from multiple
respondents participated in the Phase 1 pilot cultures and their responses would not be
studies and 15,427 respondents participated in indicative of the societal culture in which these
Phase 2. As reported in Chapter 8 by Hanges organizations functioned. We specified this
and Dickson, the number of respondents by sample requirement because respondents from
country ranged from 27 to 1,790 with an aver- these organizations would most likely, with
age per country of 251 respondents. More than very few exceptions, be almost completely from
90% of our societies had sample sizes of 75 the societal culture in which these organizations
respondents or greater.1 functioned.
In addition to culture and leadership items, Finally, only organizations from one of
the Phase 2 questionnaires also contained sev- three industries (food processing, financial
eral demographic questions. Although most of services, and telecommunications services) were
06-House.qxd 3/19/2004 4:08 PM Page 97
included in the GLOBE sample. We limited our were finally included in the GLOBE sample.
organizations to these three industries because, These societies come from North and South
after polling the CCIs involved in the initial Africa, Asia, Europe (Eastern, Central, and
phases of the project, we determined that these Northern), Latin America, North America, the
industries were present in all countries of the Middle East, and the Pacific Rim. The final
world. Further, we chose these industries sample size of 62 societies included at least 3
because they systematically differed from one countries from each major region of the world.
another and these differences have important See Chapter 10 by Gupta and Hanges for a
implications for organizational culture. In partic- detailed description of these regions.
ular, the food-processing industry is a relatively It should be noted that the reader may want
stable industry whereas the telecommunications to substitute the term country or nation rather
and financial industries may be stable or unsta- than the label society or societal culture that we
ble depending on country and economic condi- use in Project GLOBE. We use the term societal
tions. As discussed by Brodbeck, Hanges, culture to recognize the complexity of the cul-
Dickson, Gupta, and Dorfman in Chapter 20, a ture concept and because we occasionally sam-
review of both the qualitative and quantitative pled two subcultures from a single nation. It was
literature concerning these three industries veri- recognized that national borders may not be an
fied that these industries systematically differed adequate way to demarcate cultural boundaries
in meaningful ways. Thus, organizations within because many countries have large subcultures.
these industries were expected to use different It is impossible to obtain representative samples
organizational practices to successfully adapt to within each nation of such multicultural nations
their different types of environments. such as China, India, or the United States.
Nonetheless, the samples drawn from such
countries need to be comparable with respect to
Sampling of Societal Cultures
the dominant forces that shape cultures, such as
A total of 62 societal cultures were included ecological factors, history, language, and reli-
in the GLOBE sample. It should be noted that gion. The country samples also need to be rela-
the original sampling strategy did not include tively homogeneous within cultures with regard
all of the final participating societies. Indeed, to their reports of the construct being measured.
the original sampling strategy only specified For multicultural countries, whenever possible
that data from 20 countries would be collected. we sampled the subculture in which there is
These original 20 countries represented a con- the greatest amount of commercial activity.
venience sample in that Robert House knew We sampled two subcultures in South Africa,
social scientists living in these countries. Once Switzerland, and Germany. These subcultures
CCIs in these original countries were identified were indigenous and Caucasian subcultures in
and after the first pilot study was started, the South Africa, French and German subcultures
sampling strategy was modified to include an in Switzerland, and East and West subcultures
additional 20 countries. The additional recruit- in Germany.
ment of countries went very quickly and by the
end of the first pilot study, CCIs from a total
Summary
of 40 countries were actively participating in
GLOBE. The addition of the second 20 countries By following our sampling strategy, we cre-
fortuitously allowed us to replicate Pilot Study ated a “nested structure” in the Phase 2 database.
1 with an independent sample of countries. In other words, multiple middle managers were
Given the ease of recruiting these 40 sampled from organizations (i.e., “individuals
countries, we modified the sampling strategy were nested within organizations”). Further,
again. Specifically, this modification focused on only domestic organizations were sampled, thus
collecting data from additional countries to ensure organizations were nested within societies. The
that Project GLOBE had data from at least three consequence of having a nested structure in a
countries in each of the major geographic database is that it produces covariation among
regions of the world. A total of 62 societies the responses that need to be accounted for in
06-House.qxd 3/19/2004 4:08 PM Page 98
the data analytic procedures. As discussed in scale that ranged from a low of “This behavior or
Chapter 11 by Hanges, Dickson, and Sipe, we characteristic greatly inhibits a person from being
used a number of the statistical techniques, such an outstanding leader” (response score = 1) to a
as hierarchical linear modeling, to test the high of “This behavior or characteristic con-
GLOBE Phase 2 propositions. tributes greatly to a person being an outstanding
leader” (response score = 7).
Organizational and societal culture items
were written for the nine core GLOBE dimen-
GLOBE MEASURES sions, described above, at both the societal and
the organizational levels. We also wrote the
To test the GLOBE propositions, we collected items to reflect two culture manifestations: soci-
original data in a field study using question- etal practices reported “As Is” and values
naires as well as accessing archival data to reported in terms of what “Should Be.” These
develop additional measures. The archival data two culture manifestations correspond to Schein’s
served two purposes. First, we used the data as (1992) concepts of artifacts and espoused val-
a mechanism for construct validation of the cul- ues as levels of culture. Specifically, Schein
ture dimension scales. Second, we were able to argues that both artifacts and espoused values
determine the relationships between the culture are important elements of culture. Artifacts are
dimensions and important economic and human the “visible products of the group” (p. 17) and
condition variables. the organizational practices and processes that
The following sections will describe the make these visible products and behavior rou-
GLOBE culture and leadership questionnaire tine. In contrast, espoused values are “some-
that was used to collect original information one’s sense of what ought to be as distinct from
from the middle managers. After this, we will what is” (p. 19). The GLOBE culture items were
also briefly describe the archival data sources. written as “quartets” having isomorphic struc-
tures across the two levels of analysis (societal
and organizational) and across these two culture
Culture and Leadership Questionnaire
manifestations (practices and values).
As described in Chapter 8 by Hanges and The basic structure of the items comprising
Dickson, the Phase 2 GLOBE culture and lead- quartets is identical, but the frame of reference
ership questionnaire contained items that mea- varied according to the particular cultural mani-
sured organizational culture, societal culture, festation and levels of analysis being assessed.
and culturally endorsed leadership attributes. Items were derived from a review of relevant
Item generation for the culture scales was literature and interviews and focus groups held
accomplished by Robert House with substantial in several countries, as well as from extant orga-
help from Paul Koopman, Henk Thierry, and nizational and culture theory. Psychometric
Celeste Wilderom of the Netherlands and analyses indicated justification for grouping the
Phillip Podsakoff of the United States. The ini- items into scales relevant to nine core GLOBE
tial item pool contained 753 items, of which 382 dimensions of societies and organizations.
were leadership items and 371 were societal and
organizational culture items.
Versions Alpha and Beta
In generating the leadership items, our focus
was on developing a comprehensive list of leader Two versions of the GLOBE questionnaire
attributes and behaviors rather than on develop- were developed. Independent samples of middle
ing a priori leadership scales. The initial pool of managers completed one of two questionnaires.
leadership items was based on leader behaviors Half of the respondents in each organization
and attributes described in several extant leader- completed the organizational culture ques-
ship theories. The theories are described in tionnaire (version Alpha), and the other half
House and Aditya (1997). These leadership completed the societal culture questionnaire
items consisted of behavioral and attribute (version Beta). On version Alpha of the ques-
descriptors. Items were rated on a seven-point tionnaire, there were 75 questions asking about
06-House.qxd 3/19/2004 4:08 PM Page 99
organizational culture. On version Beta of the measures of a society’s economic health; the
questionnaire, there were 78 questions asking society’s success in science and technology;
about societal culture. Both versions of the the overall quality of the living in a society;
questionnaire contained the same 112 leader- and societal values surrounding service to
ship attributes questions. These two different community, tolerance of diversity, and the
versions of the questionnaire enabled us to col- nature of work. Javidan and Hauser develop
lect independent assessments of organizational several hypotheses concerning relationships
and societal culture. This process eliminates between these societal level constructs and the
common source method variance when deter- GLOBE scales.
mining relationships between organizational
and societal variables. Further, for the hierarchi-
Unit of Analysis
cal linear modeling (HLM) analyses comparing
societal culture with leadership, we compared The units of analysis for Project GLOBE
the societal culture scores from the middle consisted of cultural-level aggregated responses
manager sample that completed version Beta of middle managers. There are several issues
with the CLT dimension scores from the middle that need to be recognized with regard to this
manager sample that completed version Alpha. unit of analysis as well as the nested structure in
We found that the culture and leadership scales, the GLOBE data. First, when developing the
on average, exceeded professional standards with research strategy for the GLOBE project, we
regard to properties such as unidimensional- were quite aware of the general “levels of analy-
ity, reliability, and aggregatability. A complete sis” problem that bedevils many cross-cultural
description of the questionnaire development research projects. Inappropriately assuming that
and scale validation process is discussed in cultural-level characterizations and relation-
Chapter 8 by Hanges and Dickson. ships apply to individuals within these cultures
is commonly labeled the “ecological fallacy”
error (Robinson, 1950). The “reverse ecological
Archival Databases
fallacy,” described by Hofstede (2001), is said
To further demonstrate construct validity to occur if one compares cultures on measures
and generalizability of the GLOBE scales, created for use at the individual level. The point
archival data was used to develop alternative, he is making is that cultures are not simply
non-questionnaire-based measures of the GLOBE “king-sized individuals” and cannot be under-
societal cultural constructs. Some of this work is stood simply on the basis of the internal logic of
described by Gupta, Sully de Luque, and House individuals. The GLOBE team was well aware
in Chapter 9. These authors report strong rela- of both potential problems of ecological fallacy
tionships between scales assessing the same and reverse ecological fallacy when designing
construct but measured using different methods our study. Indeed, as described in Chapter 8, we
(i.e., convergent validity) versus the relation- carefully constructed our scales so that they
ships found if comparing different traits mea- measured constructs at the targeted level of
sured with either the same method or across analysis (organizational or societal, not individ-
different methods (i.e., discriminant validity). ual). Also we empirically assessed the extent to
In addition, Javidan and Hauser in Chapter 7 which we were successful in developing scales
also used archival data to assess the implica- assessing constructs at the desired level.
tions of the GLOBE scale scores. However, Further, van de Vijver and Leung (1997) dis-
rather than create alternative measures of the cuss how hierarchical linear modeling (HLM)
same cultural constructs, Javidan and Hauser is an ideal statistical procedure for testing hypo-
used this additional archival data to assess addi- theses with data containing a multilevel struc-
tional societal level constructs. More speci- ture. As Hanges, Dickson, and Sipe indicate in
fically, using data resources such as the United Chapter 11, HLM can be thought of as analo-
Nations’ Report on Human Development (United gous to regression analysis except that it inher-
Nations Development Program, 1998) and other ently takes levels of analysis issues into account.
economic reports, these authors obtained Therefore, HLM can “address the separate and
06-House.qxd 3/19/2004 4:08 PM Page 100
joint effects of variables at different levels on managers from one of three industries studied)
the dependent measure” (p. 126, van de Vijver is an excellent one. This question clearly requires
& Leung, 1997). Hofstede (2001) also suggests some attention.
that critical insights can be obtained by studying As described in Chapter 8 by Hanges
the same database simultaneously at different and Dickson, our societal culture scales
levels of analysis. To our knowledge, project meaningfully correlate with other researchers’
GLOBE is the first large-scale cross-cultural self-report measures in which the data were
and cross-organizational project to take advan- collected from populations different from
tage of multilevel models and to use HLM in middle managers. Further, as discussed by
hypothesis testing. Gupta, Sully de Luque, and House in Chapter 9,
As a second important issue, we should the GLOBE societal culture scales showed
discuss the potential boundary conditions for significant positive relationships with alterna-
the GLOBE scales. Consider that the GLOBE tive culture measures that were developed using
societal and organizational cultural practices archival sources. This comparison revealed that
and values scale scores originate from self- there are stronger relationships between scales
report responses of individual middle managers. assessing the same construct but measured
One possible boundary condition is whether using different methods (convergent validity)
the GLOBE scale constructs generalize beyond versus the relationships found if comparing
the constraints of the methods (i.e., self-report different traits measured with either the same
methodology) and sample (i.e., middle man- method or across different methods (discrimi-
agers that come from one of three industries) that nant validity). Further, the results comparing the
were used to collect the original GLOBE data. measures of economic and the human condition
There are researchers and methodologists variables developed by Javidan and Hauser and
that hold a measurement philosophy in which the GLOBE societal culture scales, discussed in
constructs are believed to be completely bounded Chapters 12–19, also provides evidence that the
by the methods by which they are measured. generalizability of the GLOBE constructs
This measurement philosophy, called opera- extend beyond the boundaries of the methodol-
tionalism, was extremely influential during ogy used to collect them.
the 1940s and 1950s. Operationalism was first Finally, we have evidence of strong corre-
proposed by Bridgman (1927), a Nobel prize- spondence between the questionnaire and the
winning physicist, but made famous in the social qualitative data (interviews, focus groups,
sciences by B. F. Skinner and others. According media analyses) concerning the CLTs. Although
to Bridgman, a construct is “nothing more not reported in this book, these results are dis-
than a set of operations.” In other words, con- cussed in the forthcoming Chokkar, Brodbeck,
cepts such as intelligence, motivation, and even and House (in press) GLOBE book.
culture are synonymous with the way they are In addition to this evidence, there is more
measured. For example, Boring’s (1929) defini- direct evidence that our aggregated societal cul-
tion of intelligence (i.e., “intelligence is what ture scale extends beyond just aggregated indi-
tests test”) is a classic illustration of the belief vidual reports of cultural practices and values.
that constructs are bounded by the way they are This direct evidence comes from the HLM
measured. statistical analyses. As reported in Appendix C
In contrast to this measurement philosophy, by Hanges, Sipe, and Godfrey, these additional
it is now believed that constructs can extend analyses found that the aggregated culture scales
beyond their measurement methodology (Cook & added significant variance in the prediction of
Campbell, 1979; Messick, 1981; Shadish, Cook, & the CLTs over and beyond that accounted for
Campbell, 2002). However, this belief is not by individual level variation. The fact that the
taken on faith but must be empirically demon- aggregated HLM coefficients at the ecological
strated. Thus, the question of whether the GLOBE level of analysis provide information beyond
constructs extend beyond a “set of operations” individual level results is interpreted as direct
(e.g., aggregation of individual responses) or evidence that there is a context effect (Bliese,
the characteristics of the sample (e.g., middle 2000) in our data (i.e., the results obtained at the
06-House.qxd 3/19/2004 4:08 PM Page 101
society level are not simply a reflection of the relatively underdeveloped and unstudied societies
average individual level results). (e.g., Namibia, Zimbabwe).
Taken together, this information provides
empirical support for the assertion that the con-
structs measured by the GLOBE scales general- REFERENCES
ize beyond the sample from which the data were
obtained (middle manager from one of three Bliese, P. D. (2000). Within-group agreement, non-
industries), the method used to collect these data independence, and reliability: Implications for
(self-report), and the sets of operations applied data aggregation and analysis. In K. J. Klein &
on these data (aggregation). The findings reflect S. W. J. Kozlowski (Eds.), Multilevel theory,
the broader societal and organizational cultures research, and methods in organizations:
under study. Foundations, extensions, and new directions
(pp. 349–381). San Francisco, CA: Jossey-Bass.
Boring, E. G. (1929). A history of experimental
UNIQUE STRENGTHS psychology. New York: Century.
Bridgman, P. (1927). The logic of modern physics.
OF THE GLOBE RESEARCH DESIGN
New York: MacMillan.
Chokkar, J., Brodbeck, F., & House, R. J. (Eds.).
Project GLOBE differs from previous cross-
(in press). Cultures of the world: A GLOBE
cultural research in several ways. We use multi-
anthology. Thousand Oaks: Sage.
ple measurement methods to empirically validate
Cook, T. D., & Campbell, D. T. (1979). Quasi-
the hypothesized relationships in our integrated
experimentation: Design and analysis issues
model. This is most evident in the development
for field settings. Chicago: Rand-McNally.
of our questionnaire measures assessing culture:
Hofstede, G. (2001). Culture’s consequences: Inter-
(a) based on self-reports of shared values of
national differences in work-related values
organizational or society members, (b) based on
(2nd ed.). Beverly Hills, CA: Sage.
self-reports of current organizational and societal
House, R. J., & Aditya, R. (1997). The social scien-
practices, and (c) unobtrusive measures as well as
tific study of leadership: Quo vadis? Journal of
the Work Values Survey (Inglehart, Basanez, &
Management, 23(3), 409–473.
Moreno, 1998). In addition, we developed mea-
Inglehart, R., Basanez, M., & Moreno, A. (1998).
sures of leader attributes that differentiate cultures
Human values and beliefs: A cross-cultural
in terms of perceived effectiveness.
sourcebook. Ann Arbor: University of Michigan.
We developed new measures and collected
Messick, S. (1981). Constructs and their vicissitudes
original data to test our hypotheses and answer
in educational and psychological measurement.
research questions, rather than relying on
Psychological Bulletin, 89, 575–588.
measures developed at other times in other
Robinson, W. S. (1950). Ecological correlations and
places from other samples. Because different
the behavior of individuals. American Sociolo-
people completed the organizational culture,
gical Review, 15, 351–357.
societal culture, and leadership measures used
Schein, E. (1992). Organizational culture and leader-
in Phase 2, we were able to eliminate the fre-
ship (2nd ed.). San Francisco: Jossey-Bass.
quently encountered problem of common
Shadish, W. R., Cook, T. D., & Campbell, D. T.
source bias. By use of multiple indicators of
(2002). Experimental and quasi-experimental
societal culture, we were able to minimize, if
designs for generalized causal inference.
not eliminate, common method variance.
Boston, MA: Houghton-Mifflin.
United Nations Development Program. (1998).
ENDNOTE Human development report. New York: Oxford
University Press.
1. We made the difficult decision to keep data van de Vijer, F., & Leung, K. (1997). Methods and
from societies with less than 75 respondents. We did data analysis for cross-cultural research.
so particularly because these samples came from London: Sage.
07-House.qxd 3/19/2004 4:08 PM Page 102
7
THE LINKAGE BETWEEN
GLOBE FINDINGS AND OTHER
CROSS-CULTURAL INFORMATION
MANSOUR JAVIDAN
MARKUS HAUSER
in its economic performance and prosperity as 6. Societies that are lower on power distance
well as the extent of societal and human health (Chapter 17) tend to
and quality of life. A. Be more economically prosperous and com-
As suggested earlier, culture evolves over petitively successful
time as the community learns the ways to deal B. Enjoy higher levels of societal health
with its external adaptation and internal integra- C. Enjoy higher levels of human development.
tion challenges. We would thus expect a direct
7. Societies that are more humane oriented
connection between cultural dimensions and the
(Chapter 18) tend to
various elements and outcomes of the society’s
A. Have better human conditions
attempts to deal with its challenges. To guide the
B. Be more economically prosperous
search for the relevant archival data, a series of
C. Enjoy higher levels of satisfaction.
hypotheses were generated as to the type of infor-
mation that theoretically should be correlated 8. Societies that are high on uncertainty avoid-
with each GLOBE dimension. Each dimension ance (Chapter 19) tend to
chapter will present the relevant literature and the A. Enjoy healthier state of mind
exact nature of the relationship between a partic- B. Enjoy stronger scientific progress
ular type of data and each cultural dimension. C. Have governments that support economic
The following are the hypotheses that were gen- activities.
erated to link cultural practices and external data.
As for the sources of data, the United
1. Societies that are highly performance oriented Nations and the World Bank are highly reputable
(Chapter 12) tend to: providers of cross-national data. They produce
A. Be more economically prosperous and extensive information on economic performance
competitively successful and demographic features of different countries.
B. Have a stronger social support for The World Economic Forum (WEF) and Inter-
competitiveness national Institute for Management Develop-
C. Enjoy higher levels of human development. ment (IMD) in Switzerland both produce annual
reports on competitiveness of different countries.
2. Societies that are highly future oriented Because the Global Competitiveness report by
(Chapter 13) tend to WEF shares a larger number of countries with
A. Be more economically productive GLOBE, we decided to use their findings.
B. Enjoy higher levels of societal health A comprehensive source of information on
C. Have higher rates of national savings. societal values is the World Values Survey, which
continues to expand its database of values in
3. Societies that are high on gender egalitarianism
different countries. The reports produced by
(Chapter 14) tend to
the above sources consist of a very large number
A. Be economically more prosperous
of items on each country. A method had to be
B. Enjoy higher levels of human development
adopted for selecting a parsimonious number of
C. Enjoy higher levels of psychological health.
items that were most relevant to GLOBE cultural
4. Societies that are high on assertiveness dimensions. The solution was to follow the tra-
(Chapter 15) tend to ditional process of data reduction: The different
A. Perform better in terms of global competi- types of data from these sources were first clus-
tiveness tered into conceptual groupings. For example, it
B. Be more economically prosperous was hypothesized that the number of telephone
C. Exhibit lower levels of psychological health. lines in use per 1,000 inhabitants, electricity con-
sumption per capita, and paper consumption per
5. Societies that are more collectively oriented 1,000 people are all related to the construct of
(Chapter 16) tend to economic prosperity.
A. Enjoy higher levels of societal health In the second step, factor analysis was con-
B. Enjoy higher levels of human health ducted to verify the unidimensionality of each
C. Enjoy higher levels of human development. conceptual grouping. For example, all the items
07-House.qxd 3/19/2004 4:08 PM Page 105
Table 7.1 Sources and Types of Archival Data Used in the Upcoming Chapters
HDI is a simple average of the life product per capita. The 1998 Human
expectancy index, educational attainment index Development Report includes data from 174
(a combination of adult literacy and elementary countries. Table 7.2 shows the GLOBE
school enrollment), and real gross domestic countries’ HDI values and rankings.
Hungary 0.857 47
Mexico 0.855 49
Poland 0.851 52
Colombia 0.850 53
Kuwait 0.848 54
Qatar 0.840 57
Thailand 0.838 59
Malaysia 0.834 60
Brazil 0.809 62
Turkey 0.782 69
Ecuador 0.767 73
Kazakhstan 0.695 93
Indonesia 0.679 96
Philippines 0.677 98
NOTE: HDI = Human Development Index. From Human Development Report, United Nations Development Program, 1998.
07-House.qxd 3/19/2004 4:08 PM Page 108
topics. The 1999 version has 237 items. A brief CREATING THE RELEVANT
listing of the topics is provided below: ARCHIVAL DATA FACTORS
• Ecology: Values toward the environment, As explained earlier in this chapter, a two-stage
nuclear energy, and animal rights. data reduction process was used to create unidi-
• Economy: Values toward social welfare, third mensional factors consisting of the data reported
world development, management and owner- by the various sources. The process for data
ship of business organizations, income equality, reduction was as follows: First, in collaboration
technological development, the state of the with Robert House, we identified the items that
country’s economy, and individual freedom. in their judgment were relevant to GLOBE
• Education: Values toward scientific advances, dimensions. We then grouped these items into
educational and cultural organizations, and conceptually related clusters. The result was four
confidence in the educational system. conceptual factors: Economic Health, Success in
• Emotions: Feelings and emotions, and values Science and Technology, Human Condition, and
toward voluntary work. Societal Attitudes.
• Family: Importance of family, relationships Then, we conducted factor analyses of the
with partners and parents, values toward items in each of the four conceptual factors.
marriage and children, and trust in family. This step produced several nonorthogonal
• Gender and sexuality: Views toward rotated subfactors for each one of the four gen-
women’s groups, sexual freedom, women, and eral factors. With two exceptions, Cronbach’s
homosexuality. alphas are over 0.70. A brief description of the
• Government and politics: Views toward various factors and subfactors will follow.
politics and political parties, political attitudes
toward the society and country’s goals, and
confidence in the various organs of the society. Economic Health
• Health: Views toward the state of health and
health voluntary organizations.
The GLOBE measure of a country’s eco-
• Individual: Views toward free choice, individual
nomic health is a composite measure using a
development, responsibility, and self-confidence.
variety of economic indicators. It includes data
• Leisure and friends: Importance of friends,
related to the country’s prosperity, productivity,
views toward leisure and sports.
and competitiveness. It also reflects the extent
• Morality: Views toward abortion, cheating,
of government support for economic prosperity
drugs, and lying.
and the general societal support for economic
• Religion: Importance of religion, meaning of
competitiveness of the country. Table 7.4 shows
life, views toward the church, belief in good and
the specific items from the various sources that
evil, and views toward God.
generated each subfactor.
• Society and nation: Views toward youth
work organizations, tolerance of people who are
different, confidence in social security systems,
Success in Science and Technology
trust in different types of groups and individuals. A country’s accomplishments in regard to
• Work: Importance of work, types of voluntary science and technology are a strong indicator of
work, important aspects of a job, and reasons its present and future success and prosperity.
for working. We identified two overall measures to assess
the participating countries: 1. Their success in
The results of this survey are provided in a applied science, which refers to the country’s
variety of formats. A valuable and comprehen- performance in terms of commercializing and
sive source is Human Values and Beliefs: A Cross- exploiting technology. 2. A country’s ability to
Cultural Sourcebook (Inglehart, Basanez, & invest in basic science. These two measures are
Moreno, 1998). The program’s Web site related but quite different. Basic science tends
address is http://wvs.isr.umich.edu/ to require greater investment and a longer time
(Text continues on page 112)
07-House.qxd 3/19/2004 4:08 PM Page 111
Economic Prosperity
(A higher score on this scale means a more prosperous country) (Alpha = .949)
• Telephone lines (Number of main lines in use per 1,000 inhabitants) (WCY)
• Cellular mobile telephone subscribers (Number of subscribers per 1,000 inhabitants) (WCY)
• Electricity consumption per capita, 1995 (HDR)
• Personal computers per 1,000 people (HDR)
• Fax machines per 1,000 people, 1995 (HDR)
• Paper consumption per 1,000 people (HDR)
Economic Productivity
(A higher score on this item means a more productive country)
• Public service (1 = The public service is exposed to political interference, 10 = The public service
is not exposed to political Interference) (WCY)
• Bribing and corruption (1 = Bribing and corruption exist in the public sphere, 10 = Bribing and
corruption does not exist in the public sphere) (WCY)
• Financial resources (1 = Lack of sufficient financial resources constrains technological development,
10 = Lack of sufficient financial resources does not constrain technological development) (WCY)
• Development and application of technology (1 = Development and application of technology is
constrained by the legal environment, 10 = Development and application of technology is not
constrained by the legal environment) (WCY)
• Competition laws (1 = Competition laws do not prevent unfair competition in your country,
10 = Competition laws prevent unfair competition in your country) (WCY)
• Economic literacy (1 = Economic literacy is generally low among the population, 10 = Economic
literacy is generally high among the population) (WCY)
• Industrial relations (1 = Labor relations are generally hostile, 10 = Labor relations are generally
productive) (WCY)
• Restructuring of the economy (1 = Restructuring the domestic economy is not adapted for
long-term competitiveness, 10 = Restructuring the domestic economy is adapted for long-term
competitiveness) (WCY)
• Political system (1 = The political system is not well adapted to today’s economic challenges,
10 = The political system is well adapted to today’s economic challenges) (WCY)
• Labor regulations (1 = Labor regulations [hiring and firing practices, minimum wages, . . .] are too
restrictive, 10 = Labor regulations are flexible enough) (WCY)
(Continued)
07-House.qxd 3/19/2004 4:08 PM Page 112
These rankings are produced based on the competitiveness index calculated for each country.
NOTE: GCR = Global Competitiveness Rankings by World Economic Forum (1998); HDR = Human Development Report
(United Nations Development Program, 1998); WVS = World Values Survey, 1994, WCY = World Competitiveness Yearbook
(IMD, 1999). WDI = World Development Indicators (World Bank, 2000).
frame. It also needs a strong university-based dependent on the corporate sector. Table 7.5
scientific capability. Applied science, in contrast, shows the specific factors and items used to
tends to be shorter-term oriented, and is heavily measure each category.
NOTE: GCR = Global Competitiveness Report (World Economic Forum, 1998); HDR = Human Development Report (United
Nations Development Program, 1998); WVS = World Values Survey, 1994; WCY = World Competitiveness Yearbook (IMD,
1999).
07-House.qxd 3/19/2004 4:08 PM Page 113
Societal Health
(A higher score means a healthier society) (Alpha = .863)
• Justice (1 = Justice is not fairly administered in society, 10 = Justice is fairly administered in
society) (WCY)
• Personal security and private property (1 = Personal security and private property are not adequately
protected, 10 = Personal security and private property are adequately protected) (WCY)
• Quality of life (1 = Quality of life in your country is not high, 10 = Quality of life in your country
is high) (WCY)
• Alcohol and drug abuse (1 = Alcohol and drugs abuse pose a serious problem at the work place,
10 = Alcohol and drugs abuse do not pose a serious problem at the work place) (WCY)
Human Health
(A higher score reflects healthier individuals) (Alpha = .724)
• Doctors per capita
• Nurses per capita
• Infant survival per 1,000
Life Expectancy
(A higher score on this item means longer life expectancy)
This statistic is provided in the Human Development Report (United Nations Development
Program, 1998).
General Satisfaction
(A higher score on this scale means greater satisfaction) (WVS) (Alpha = .880)
V10. Taking all things together, would you say you are . . . 1) Very happy, 2) Quite Happy, 3) Not
Very happy, 4) Not at all happy, 5) Don’t know (percentage who are very happy and quite
happy)
(Continued)
07-House.qxd 3/19/2004 4:08 PM Page 114
V11. All in all, how would you describe your state of health these days? Would you say it is very
good, good, fair, poor, or very poor? (percentage of “very good” or “good”)
V65. All things considered, how satisfied are you with your life as a whole these days? (Ten-point
scale: 1 = Dissatisfied and 10 = Satisfied); (percentage of “satisfied”—codes 7 to 10.)
V82. During the past few weeks, did you ever feel “Pleased about having accomplished something”?
(percentage of “yes”)
V165. How satisfied are you with the way the people now in national office are handling the country’s
affairs? (percentage of “fairly satisfied” and “very satisfied”).
V213. How widespread do you think bribe taking and corruption is in this country? (percentage of
“a few public officials are engaged in it” and “almost no public officials are engaged in it”).
Psychological Health*
(A lower score means a higher level of psychological health) (WVS) (Alpha = .850)
V85. During the past few weeks, did you ever feel “So restless you couldn’t sit long in a chair”?
(percentage of “yes”)
V87. During the past few weeks, did you ever feel . . . “Very lonely or remote from other people”?
(percentage of “yes”)
V89. During the past few weeks, did you ever feel . . . “Bored”? (percentage of “yes”)
V91. During the past few weeks, did you ever feel . . . “Depressed or very unhappy”? (percentage
of “yes”)
This statistic is provided in the Human Development Report (United Nations Development
Program, 1998).
NOTE: GCR = Global Competitiveness Report (World Economic Forum, 1998); HDR = Human Development Report (United
Nations Development Program, 1998); WVS = World Values Survey, 1994; WCY = World Competitiveness Yearbook (IMD,
1999)
* The items in this scale are from the 1993 version of WVS. They were not included in the later version of
the survey.
ACHIEVEMENT (WVS)
Achieving Results
(A higher score means greater importance) (Alpha =.76)
V66. How much freedom of choice and control do you feel you have over the way your life turns
out? (1 = not at all, 10 = a great deal) (percentage of selecting “7” to “10”)
V67. I always continue to work on a task until I am satisfied with the results (percentage of “agree”
and “strongly agree”)
V73. Regardless of whether you are actually looking for a job, which would you, personally, place
first if you were looking for a job? (percentage of “doing an important job which gives you a
feeling of accomplishment”)
Initiative
(A higher score means greater importance) (Alpha = .90)
V80. Which aspects of a job are important to you? (percentage of “an opportunity to use initiative”)
V83. Which aspects of a job are important to you? (percentage of “a responsible job”)
(Continued)
07-House.qxd 3/19/2004 4:08 PM Page 116
Passiveness
(A higher score means greater passiveness) (Alpha = .80)
V118. Have you ever, or would you ever, sign a petition? (percentage of “would never do”)
V119. Have you ever, or would you ever, join in a boycott? (percentage of “would never do”)
V120. Have you ever, or would you ever, attend lawful demonstrations? (percentage of “would never do”)
V121. Have you ever, or would you ever, join unofficial strikes? (percentage of “would never do”)
Lack of Voice
(A higher score means less voice) (Alpha = .85)
V104. What should be the most important goal of the country? (reverse of percentage of “seeing that
people have a say in how things are done in their jobs and communities”)
V106. What should be the most important goal of the country? (reverse of percentage of “giving
people more say in important government decisions” and “protecting freedom of speech”)
Role of Government
(A higher score means a more active role for the government) (Alpha = .72)
V126. 1 = Private ownership of business and industry should be increased, 10 = Government owner-
ship of business and industry should be increased (percentage of selecting “7” to “10”)
V127. 1 = Government should take more responsibility to ensure that everyone is provided for, 10 = People
should take more responsibility to take care of themselves (percentage of selecting “1” to “4”).
Stability
(A higher score means more emphasis on stability) (Alpha = .77)
V106. Which one is most important? (percentage of “maintaining order in the nation”)
V108. Which one is most important? (percentage of “a stable economy”)
RELIGION (WVS)
Religious Devotion
(A higher score means stronger devotion) (Alpha = .93)
V9. How important is religion in your life? (percentage of “rather important” and “very important”)
V178. Which statement is close to your point of view? (percentage of “There are absolutely clear
guidelines about good and evil. These always apply to everyone, whatever the circumstances”)
V181. Apart from weddings, funerals, and christenings, about how often do you attend religious
services these days? (percentage of “only on special holy days” to “more than once a week”)
07-House.qxd 3/19/2004 4:08 PM Page 117
V182. Independently of whether you go to church or not, would you say you are (percentage of “a
religious person”)
V190. How important is God in your life? 1 = not at all important. 10 = very important (percentage
of “7” to “10”).
Religious Dogma
(A higher score means stronger dogma) (Alpha = .96)
V185. Do you believe people have a soul? (percentage of “yes”)
V186. Do you believe the Devil exists? (percentage of “yes”)
V187. Do you believe in hell? (percentage of “yes”)
V188. Do you believe in heaven? (percentage of “yes”)
NOTE: GCR = Global Competitiveness Report (World Economic Forum, 1998); HDR = Human Development Report (United
Nations Development Program, 1998); WVS = World Values Survey, 1999.
THE IMPACT OF NATIONAL WEALTH The issue raised by these authors is important
and they both make an important case. GLOBE
In the second edition of his book Culture’s is sympathetic to both points and takes a rigor-
Consequences, Hofstede (2001) consistently ous approach to the issue. In subsequent
tested for the moderating effect of wealth when chapters, we will show correlations between
computing correlations between his cultural each cultural dimension and many of the
dimensions and many types of secondary data. His archival variables discussed earlier here.
rationale was that economic wealth can confound GLOBE takes the view that gross national prod-
such relationships. For example, he argued that uct per capita (GNP/capita) is a reflection of
increasing affluence causes higher levels of indi- a society’s natural resources as well as its effec-
vidualism in society. So, in correlating cultural tiveness in managing its external adaptation and
dimensions with any other type of data, his con- internal integration challenges (Schein, 1992).
cern was that a spurious correlation may be found As a result, it is plausible that it would be corre-
caused by the impact of wealth rather than culture, lated with many aspects of its culture. But the
or that the relationship may be significant at some relationship is not just a simple unidirectional
income levels but insignificant at other levels. one; although wealth can be accumulated as a
In his review of Hofstede’s book, Smith result of a society’s success, it can also help
(2002) cautioned against such an approach, facilitate its further success. National wealth has
arguing that national wealth could be an integral a reinforcing effect that can help facilitate the
part of a country’s culture. It is not necessarily relationship between culture and other national
an extraneous variable that can be easily taken features. A performance-oriented society can
out. Moderating for wealth may deprive us of a prosper and as a result can better educate its
better understanding of the complex dynamics people who will contribute more to their societies.
of culture formation and maintenance. The relationships among wealth, national culture,
07-House.qxd 3/19/2004 4:08 PM Page 118
Table 7.8 Correlations Between Cultural Dimensions and GNP per Capita (1998)
and other archival variables are so intertwined the archival variables that are significantly
that they cannot be easily isolated, and cause and correlated. The results are shown in Table 7.9.
effect relationships, although intuitively appeal- Sixteen out of 21 variables are significantly
ing, are hard to verify empirically. Therefore, correlated with GNP/capita, suggesting that the
we decided to follow a three-step process: archival variables of interest to us are generally
relevant and consequential because they are
1. We calculated the correlations between all related to national wealth.
cultural dimensions and GNP/capita to identify
the cultural dimensions that are significantly cor- 3. From the above two steps, it is clear that
related. The results are shown in Table 7.8. Six wealth, national culture, and archival variables
out of nine values and six out of nine practices are are in many ways interconnected. It is tem-
significantly correlated with GNP/capita. This is pting at this stage to calculate the correla-
strong evidence for our premise that GNP/capita is tions between cultural dimensions and archival
an integral part of a society’s culture. variables while partialling out GNP/capita. The
problem with this approach is that the results are
2. We calculated the correlations between not symmetric. If the correlation is still signifi-
all archival variables and GNP/capita to identify cant after controlling for GNP/capita, we can
07-House.qxd 3/19/2004 4:08 PM Page 119
Table 7.9 Correlations Between GNP per Capita and Archival Variables
Initiative .06
n = 39
Passiveness −.39*
n = 38
Stability −.52**
n = 39
(Continued)
07-House.qxd 3/19/2004 4:08 PM Page 120
interpret it to mean that the two are related even 19 of this book, we address these two questions
if we hold GNP/capita constant. But if the rela- in terms of a variety of cultural dimensions.
tionship is not significant, it does not necessar- Each chapter focuses on one cultural dimension
ily mean that the relationship is spurious. Due to such as performance orientation. To answer
the intercorrelations between the variables, par- the first question, the chapter will present
tialling out GNP/capita does not control just for the GLOBE project results related to the partic-
wealth. Given that wealth is a part and parcel of ular cultural dimension. To answer the second
a country’s culture, partialling it out takes many question, the chapter provides a detailed analy-
other aspects of the culture out of the relation- sis of the relationship between GLOBE findings
ship and we do not really know how to interpret and the other types of cross-cultural findings
the results. As a possible solution, we conducted described in this chapter.
a regression analysis with the archival data as It is important to note that the analyses
the dependent variable. Three independent vari- reported in the forthcoming chapters are cross-
ables were entered in three separate steps: The sectional and correlational and can only postu-
specific cultural dimension, GNP/capita, and late possible relationships, falling short of
their interaction. The purpose of this procedure reaching causal conclusions. We should also be
is to verify Hofstede’s (2001) approach, in cognizant of the potential deficiencies in corre-
which he compared the correlation coefficients lating different types of cross-national data and
between his cultural dimensions and his variables their implications for interpreting the results.
of interest in wealthy countries and in poor The data refer to complex concepts and are col-
countries. Such an approach is based on the lected in different time frames using different
assumption that GNP/capita has an interaction methodologies. Despite the potential shortcom-
effect with the cultural dimension. Our results ing, the reported analyses will provide a richer
showed that in more than 300 such regressions, and fuller understanding of the similarities and
only nine cases of significant interaction effects differences among cultures and their possible
were observed, leading to the conclusion that correlates.
they are random occurrences. We therefore,
have no evidence for a significant moderating
effect for GNP/capita. REFERENCES
To conclude, at the start of this chapter, we
suggested that the fundamental purpose of Hofstede, G. (2001). Culture’s consequences (2nd
cross-cultural research is to find the answers to ed.). Thousand Oaks, CA: Sage.
two questions: How are human communities House, R. J., Wright, N. S., & Aditya, R. N. (1997).
different or similar, and why? In Chapters 12 to Cross-cultural research on organizational
07-House.qxd 3/19/2004 4:08 PM Page 121
leadership: A critical analysis and a proposed Smith, P. B. (2002, January). Culture’s consequences:
theory. In P. C. Earley & M. Erez (Eds.), Something old and something new. Human
New perspectives in international industrial Relations, 55(1), 119–137.
organizational psychology (pp. 535–625). San The World Bank. (2000) World development indica-
Francisco: New Lexington. tors. Washington DC: Author.
International Institute for Management Development. World Economic Forum. (1979). Report on inter-
(1999). The world competitiveness yearbook. national competitiveness. Geneva, Switzerland:
Lausanne, Switzerland: Author. Author.
Inglehart, R., Basanez, M., & Moreno, A. (1998). World Economic Forum. (1998). The global compet-
Human values and beliefs: A cross-cultural source- itiveness report. Geneva, Switzerland: Author.
book. Ann Arbor: University of Michigan Press. United Nations Development Program. (1998).
Schein, E. H. (1992). Organizational culture and Human development report. New York: Oxford
leadership (2nd ed.). San Francisco: Jossey-Bass. University Press.
08-House.qxd 3/19/2004 4:11 PM Page 122
8
THE DEVELOPMENT AND
VALIDATION OF THE GLOBE
CULTURE AND LEADERSHIP SCALES
PAUL J. HANGES
MARCUS W. DICKSON
AUTHORS’ NOTE: The work reported in this chapter was partially funded by the National Science Foundation
under Grant No. 9711143.
122
08-House.qxd 3/19/2004 4:11 PM Page 123
then grouped into 10 categories and he developed Some assessment of the actual convergence
a theory of the interrelations among these value among people within an organization or a society
categories (Schwartz, 1992, 1994). On the basis is clearly required to demonstrate that the GLOBE
of his theory as well as prior work by other scales were designed appropriately. Further, the
researchers (e.g., Rokeach, 1973), Schwartz devel- convergent–emergent conceptualization of the
oped his survey. Consistent with the theory- GLOBE constructs suggests that the GLOBE
driven approach to scale development, Schwartz scales need to be designed so that they operate
specified several potential biases that could limit at an appropriate level of analysis. In other
the value of his scale (e.g., having items only words, the psychometric properties of GLOBE
assessing Western societies values; cultural scales should emerge at either the organiza-
response bias) and took steps to minimize the tional or societal level of analysis (depending on
influence of these factors (Smith & Schwartz, the desired focus of the scale). Therefore, some
1997). The advantages provided by Schwartz’ empirical verification that the GLOBE scales
values survey to the cross-cultural literature is a are operating at their targeted level of analysis is
clear example of the utility of developing scales also required. It should be noted that the GLOBE
according to the theory-driven approach. Thus, scales were designed to assess variation at a par-
it was this approach that we attempted to follow ticular level of analysis and so it should not be
when developing the GLOBE scales. surprising if different psychometric properties
are found if researchers use these scales at differ-
ent levels than the one they were designed for.
Construct Specification
Clearly, because the GLOBE scales were designed
As indicated previously, we first specified to assess convergent–emergent constructs, they
the general nature of the constructs that we were not designed to assess individual variation
wanted to measure before writing any items or within organizations or societies and should not
developing any of the GLOBE scales. This crit- be used for this purpose. We test these assump-
ical step, which unfortunately is frequently tions and hypotheses as part of our scale devel-
overlooked, determines how the items should opment and validation process.
be written as well as the kinds of statistical analy-
ses that need to be performed to assess the ade- Dimension Specification
quacy of the scales (Kozlowski & Klein, 2000).
and Item Generation
We propose that organizational culture, societal
culture, and culturally endorsed implicit leader- Now that the general nature of the GLOBE
ship theory are what multilevel researchers call constructs have been identified (i.e., convergent–
convergent–emergent constructs (Kozlowski & emergent constructs), the next task was to spec-
Klein, 2000). These constructs are convergent ify the exact cultural and leadership dimensions
because the responses from people within orga- that would be used to differentiate organizations
nizations or societies are believed to center and societies and to write items that assessed
about a single value usually represented by these dimensions. As indicated by House and
scale means. They are called emergent because Javidan in Chapter 2 of this book, we originally
even though the origin of these constructs are identified seven cultural dimensions that might
a function of the cognition, affect, and person- influence the type of behaviors and attributes
ality of the survey respondents, the properties desired of leadership in a society. These dimen-
of these constructs are actually manifested at sions were: (a) Uncertainty Avoidance; (b) Power
the aggregate- or group- (e.g., organization or Distance; (c) Individualism and Collectivism;
society) level of analysis. (d) Gender Egalitarianism; (e) Future Orientation;
Thus, by conceptualizing the GLOBE con- (f) Performance Orientation; and (g) Humane
structs as convergent–emergent, we are implicitly Orientation.
hypothesizing that respondents within organiza- A total of 371 culture items were originally
tions or societies will exhibit some similarity in written. Most of these items were derived through
their perceptions of organizational or societal interviews and focus groups held in several
culture as well as their perceptions of the CLTs. countries. To ensure that the content domain of the
08-House.qxd 3/19/2004 4:11 PM Page 125
Table 8.1 Example of Culture Item Quartets Across Organizational and Societal Levels and the Two
Culture Manifestations
culture construct was adequately covered, we assessing the cultural values of an organization or
categorized the GLOBE culture items into those society focused respondents’ attention on how
assessing more tangible attributes of culture (e.g., things should be (referred to as “Should Be”
current policies and practices) and those assessing items). The importance of specifying the appropri-
more intangible attributes (e.g., cultural norms and ate frame of reference for questionnaire items
values). This categorization helped identify prob- has been emphasized in both the levels of analysis
lems in having too many or too few items assess- (e.g., Klein, Conn, Smith, & Sorra, 2001; Klein,
ing tangible or intangible aspects of culture in our Dansereau, & Hall, 1994; Payne, Fineman, &
organizational and societal culture scales. Indeed, Wall, 1976; Schneider, 1990; Schriesheim, 1979;
we wrote new items so that our culture scales Yammarino, 1990) and the cross-cultural (e.g.,
followed a “parallel quartet” structure. That is, we Dickson, Aditya, & Chhokar, 2000; Hanges, Lord,
wrote four versions of each culture item so that & Dickson, 2000) literature. Thus, the basic struc-
both the tangible (i.e., cultural practices) and intan- ture of our items comprising parallel quartets was
gible (i.e., cultural values) aspects of culture were identical, but the frame of reference was varied
assessed across the two levels of analysis consid- according to the particular cultural manifestation
ered in the GLOBE project (i.e., organizational and level of analysis being assessed. Table 8.1
culture and societal culture). The items assessing contains an example of a quartet of parallel culture
the cultural practices of an organization or society items, showing essentially the same question in
focused respondents’ attention on how things are four forms: Organization Cultural Practices (As
(referred to as “As Is” items), whereas the items Is); Organization Cultural Values (Should Be);
08-House.qxd 3/19/2004 4:11 PM Page 126
Societal Cultural Practices (As Is); and Societal sorters into the categories for which they were
Cultural Values (Should Be). theoretically intended were retained for further
With respect to the leadership items, 382 items analysis. There were no dimensions of societal
were originally written. Consistent with the way or organizational culture for which a majority of
implicit leadership theory of individuals has items failed to meet this criterion. Thus, a suffi-
been measured in previous research (e.g., Foti & cient number of items were retained for the
Lord, 1987; Lord, Foti, & DeVader, 1984; Lord & measurement of each dimension.
Maher, 1991a, 1991b), the GLOBE leadership In addition to this sorting task, CCIs provided
items consisted of behavioral and trait descrip- Item Evaluation Reports in which they noted any
tors (e.g., autocratic, benevolent, nurturing, and items containing words or phrases that were
visionary) along with a brief definition of these ambiguous or could not be adequately translated
descriptors. The items were written to reflect a in the target country’s native language. CCIs also
variety of traits, skills, abilities, and personality identified questions that might be culturally inap-
characteristics potentially relevant to leadership propriate. Most of the items that were problem-
emergence and effectiveness. In generating the atic were dropped from further consideration.
items, our focus was on developing a compre- In some cases, we were able to rewrite items to
hensive list of leader attributes and behaviors eliminate potential problems but retain the intent
rather than on solely using a priori leadership and dimensionality of the original item.
scales. However, the initial pool of leadership Finally, the cross-cultural literature indicates
items was partially based on several extant that a systematic bias may occur if respondents
leadership theories, and thus included leader complete a survey that is not in their native
behaviors and attributes described in established language (Brislin, 1986). To prevent this, the
leadership theories. Items were rated on a questionnaires were translated into several lan-
seven-point scale that ranged from a low of guages. CCIs were responsible for having the
“This behavior or characteristic greatly inhibits survey translated from English into their native
a person from being an outstanding leader” to language. This was done by the CCI, by some
a high of “This behavior or characteristic con- other person fluent in both languages, or most
tributes greatly to a person being an outstanding often by a professional translator. The transla-
leader.” The definition of leadership provided to tion was then independently translated again,
our respondents along with several examples of from the native language of the culture back to
leadership items and the response scale used for English. This back-translation was then sent
these items are shown in Tables 8.1 and 8.2.2 to the authors of this chapter, who compared
the original English version of the survey with
the back-translation to verify the accuracy of the
Item Review Process
translation. A pragmatic approach (Brislin, 1976,
The societal and organizational culture items 1986) was taken in evaluating the adequacy
were screened for appropriateness by use of of the back-translations. In other words, we
three procedures: Q-sorting, item evaluation, were more concerned with the accuracy with
and translation and back translation. Leadership which the concepts were translated rather than
items were screened by item evaluation and the exact words being used in the translations.
conceptual equivalence of the back translation. When discrepancies between the original survey
The Q-sorting procedure consisted of sorting and the back-translations were encountered, the
the culture items into theoretical categories rep- CCI was notified and the discrepancy was dis-
resented by the a priori dimensions of culture cussed. If necessary, revisions in the translation
described above, first by seven PhD students in were made.
the Department of Psychology at the University As a result of editing or deleting items as a
of Maryland and subsequently by country co- result of the Q-sorting, item evaluation, and
investigators (CCIs) representing 38 countries. translation and back-translation processes, the
The sorters were not told the particular cultural item pools were reduced by approximately 50%.
dimension that each item was intended to mea- One benefit of subjecting our items to this mul-
sure. Items that were categorized by 80% of the tistep evaluation process is that the items that
08-House.qxd 3/19/2004 4:11 PM Page 127
survived are very likely to exhibit measurement retained items to individuals in their respective
equivalence across countries. In other words, countries who had full-time working experience
the ability of sorters who came from different as a white-collar employee or manager. Because
world cultures to agree on the allocation of the survey was lengthy, it was divided into two
items indicates that these sorters had similar parallel versions, A and B. Each version con-
interpretations of the items as well as the theo- tained approximately half of the leadership
retical dimensions underlying the items. items and half of the organizational and societal
Further, the fact that the remaining items also culture items. A total of 877 individuals com-
survived an explicit evaluation for cultural pleted the first pilot study survey.
inconsistencies and the translation and back- Several different statistical analyses were
translation process strongly suggests common- performed to assess the psychometric properties
alty of meaning of the questionnaire items and of the scales. Specifically, we conducted a series
scales across cultures. of exploratory factor analyses, reliability analy-
ses, and aggregation analyses (e.g., rwg (J) analy-
ses, intraclass correlations [ICC(1)], one-way
Pilot Studies
analyses of variance) on our scales. It should be
Two empirical pilot studies were conducted noted that because we wanted these scales to
to assess the psychometric properties of the measure organizational or societal level and not
resulting a priori culture scales and to empiri- individual level variation, we performed these
cally develop leadership scales. analyses on the means of the country item
responses for each scale. We refined our scales
on the basis of these analyses while trying to
Pilot Study 1
maintain, whenever possible, the isomorphic
The CCIs in 28 countries (see Table 8.3 for quartet structure of the culture scales described
these countries) distributed the survey of above. These statistical analyses, if considered
08-House.qxd 3/19/2004 4:11 PM Page 128
together, provide useful information about the of the psychometric properties of our scales.
construct validity of the culture scales. Data for Pilot Study 2 came from 15 countries
Separate factor analyses of each of the culture that did not participate in the previous pilot
scales indicated that they were all unidimen- study (see Table 8.3). In general, the psychome-
sional. A first-order exploratory factor analysis tric properties of all of the scales were con-
of the leader attributes items yielded 16 unidi- firmed by replication process. A total of 1,066
mensional factors that describe specific leader individuals completed one of the three versions
attributes and behaviors. (organizational culture items, societal culture
The rwg(J) analyses (James, Demaree, & Wolf, items, and leadership items) of the Pilot Study 2
1984, 1993), demonstrated that aggregation of questionnaires. Our results not only confirmed
the organizational and societal culture scales is the psychometric properties of the leadership
appropriate (average rwg(J) = .73, n = 54).3 Intraclass subscales and the culture scales,4 but they also
correlation coefficients [ICC(1)] and one-way verified through aggregation tests that we were
analyses of variance (Bliese, 2000; McGraw & justified in aggregating these scales to their
Wong, 1996; Shrout & Fleiss, 1979) for each of target level of analysis.
the scales indicated statistically significant
within-culture agreement and between-culture
differences. The societal culture scales exhib- PHASE 2: FINAL SAMPLE AND
ited low to moderate correlations with each CONFIRMATION OF MEASUREMENT SCALES
other. Thus, they provide independent and unique
information about societal cultures. Phase 2 of GLOBE was designed to formally
The leadership subscales substantially dif- test the hypotheses concerning culture and
fered in their relationship to one another. The leadership. In the present chapter, we use these
absolute correlation among the leadership data to provide further evidence concerning
subscales ranged from a low of .00 (e.g., Status- the psychometric properties of the GLOBE
Conscious with Calmness) to a high of .86 scales. It should be noted that additional items
(Status-Conscious with Procedural). Overall, were written during Phase 1 of GLOBE on the
20% of the interrelationships were statistically basis of the results of the pilot studies as well
significant. A second-order factor analysis of as focus groups and interviews conducted by
the 16 leadership factors yielded 4 unidimen- the CCIs. We added several leadership attribute
sional factors that describe classes of leader items as well as some culture items to the ques-
behaviors that represent global leader behavior tionnaire. These new leadership items were
patterns. written to ensure that our 16 original leadership
To summarize, at the end of Pilot Study 1, we subscales were not biased by including only
developed 16 factorially derived leadership sub- western leadership behaviors. Further, we
scales at the societal level of analysis that repre- wrote several items that described autocratic,
sent specific leader behaviors. These subscales self-aggrandizing, manipulative, and punitive
could be further classified into four second-order behaviors because it was suggested in the inter-
factorially derived leadership dimensions that views and focus groups that some societies
represent global leader behavior patterns. We might view these behaviors as enhancing leader
have seven organizational culture practices effectiveness. The Phase 2 data were used to
(As Is) scales, seven organizational culture val- identify additional leadership scales among
ues (Should Be) scales, seven societal culture these items with the final result being an expan-
practices (As Is) scales, and seven societal culture sion of the original 16 leadership subscales to
values (Should Be) scales. 21 subscales. See Table 8.4 for a listing of these
leadership subscales.
We also added several items to develop
Pilot Study 2
an alternative measure of the cultural dimension
The purpose of the second pilot study was to of collectivism. These items were added
replicate the psychometric analyses of the scales because it became clear that the collectivism
in a different sample, and assess the robustness culture scales derived from the pilot studies
08-House.qxd 3/19/2004 4:11 PM Page 129
Table 8.3 GLOBE Participating Countries for Pilot Studies 1 and 2 and Phase 2
(Continued)
08-House.qxd 3/19/2004 4:11 PM Page 130
NOTE: Countries identified with a * indicate that they provided data for the specified data collection period. Countries marked
with a * for Pilot Study 1 provided qualitative item information or survey data, or both.
08-House.qxd 3/19/2004 4:11 PM Page 131
Table 8.4 Leadership Prototype Scales: First Order Factors and Leader Attribute Items
did not include items relevant to collectivistic in press at that time (i.e., Gelfand, Triandis, &
in-group phenomena. The new items were Chan, 1996) indicated that individualism and
adapted from Triandis’s (1995) work on collec- collectivism might be independent dimensions.
tivism so that they conformed to our multiple The new collectivism items were added to
manifestation (i.e., practices and values) and the the Pilot Study 2 survey and an exploratory
multiple levels of analysis conceptualization of factor analysis on the Pilot Study 2 data con-
culture. The additional items focused on collec- firmed that our original collectivism scales
tivistic in-group phenomena such as group and the new Triandis-based collectivism scales
pride, loyalty, and cohesiveness. In addition, measured separate aspects of collectivism. We
the original Individualism–Collectivism scale confirmed the two collectivism scales by
placed Individualism and Collectivism at two factor analytically deriving two dimensions of
ends of one continuum, but research appearing Collectivism: Institutional Collectivism (the
08-House.qxd 3/19/2004 4:11 PM Page 132
original Individualism–Collectivism scales) and Form Alpha respondents were not significantly
In-Group Collectivism (consisting of items different from those for Form Beta respondents.
relevant to family and organization and based Thus, the individual leadership scale scores
on Triandis’ scales). for the two samples were combined to produce
Thus, at the beginning of Phase 2, we con- means on the leadership scales for all cultures.
firmed scales measuring eight cultural practices As a result of the independent assessment of
(As Is) and eight cultural values (Should Be) the organizational and societal variables, and
scales at both the organizational and societal because the mean leadership scale responses in
levels of analysis with which to measure the each sample in each culture were not different,
seven original culture dimensions plus the new the responses are relatively free of common
Collectivism scale based on Triandis (1995). source response bias.
the extent to which our scales were generalizable factor structure of the culture scales. We
across different respondents in an organization or primarily used the comparative fit index (CFI)
society (i.e., interrater reliability). to assess the fit of the factor structure to our
data. The average CFI for the organizational
Results for Phase 2 cultural practices (As Is) and organizational
cultural values (Should Be) scales were .93 and
The consequence of adding items and con- .94, respectively. The average CFI for the soci-
ducting additional factor analyses resulted in etal cultural practices (As Is) and societal cul-
expanding the original 16 leadership subscales tural values (Should Be) scales were .89 and .95,
to 21 subscales. These additional basic factors respectively. The low CFIs were primarily a
represent both positive and negative elements function of the Gender Egalitarianism cultural
of leadership (viewed from a conventional practices (As Is) and Gender Egalitarianism
Western perspective). cultural values (Should Be) scales, at both the
societal and organizational levels of analysis.
Aggregation Verification Using exploratory maximum likelihood factor
analysis on the group level variance–covariance
Overall, these results support aggregation.
matrix, we discovered that these scales could be
The average rwg(J) s for the organizational and
separated into two unidimensional factors.
societal cultural practices (As Is) scales are .69
These two factors appeared to be meaningful
and .85, respectively. The average rwg(J) s for the
with the first factor consisting of items assessing
organizational and societal cultural values
the assertiveness of an organization or society
(Should Be) scales were both .80. Finally, the
and the second factor reflecting the degree of
average rwg(J) for the leadership attribute sub-
equality between men and women in an organi-
scales was .78. See Hanges, House, and col-
zation or society. When these two scales were
leagues (2003) for more details, if desired.
treated separately, the internal consistency of
We also calculated ICC(1) for these scales
these subscales was substantially improved. We
and these values are shown in Table 8.5. It
therefore decided to treat these two subfactors
should be noted that the average ICC(1)
of Gender Egalitarianism separately for future
reported in the organizational literature is .12
analyses.
(James, 1982) with the typical values ranging
The average CFI for the leadership attribute
from .05 to .20 (Bliese, 2000). The average
subscales was .92. This result indicated sub-
ICC(1) for the GLOBE organizational cultural
stantial support for our leadership subscales.
practices (As Is) scales is .22, whereas the
However, for three of our scales (i.e., Charismatic
average ICC(1) for the organizational cultural
I: Visionary, Self-centered, and Decisive), the
values (Should Be) scales is .27. The average
CFIs were lower than desired (i.e., CFI = .86,
ICC(1) for the society cultural practices (As Is)
.85, and .78, respectively). We examined these
culture scales is .25, whereas the average
three scales and discovered that the low fit
ICC(1) for the society cultural values (Should
indices were primarily a function of a different
Be) scales is .27. The average ICC(1) for the
(and much weaker) factor structure operating at
leadership attribute subscales is .18. Given the
the within-society level of analysis. Because
value of ICC(1) typically found in the organiza-
these scales were created to measure culturally
tional research, the aggregation of the GLOBE
endorsed (i.e., emergent–convergent) constructs,
scales to their targeted levels of analysis is
we changed the pooled within-society factor
strongly supported.7 In summary, the Phase 2
structure. We discovered that the goodness of fit
data again support the appropriateness of aggre-
of these scales improved to acceptable levels if
gating these scales to the organizational and
a two-factor within-society structure was imposed.
societal level of analysis.
Thus, we interpreted these results as supporting
the between-society factor structure of the lead-
Multilevel Confirmatory Factor Analysis
ership and culture scales. For the three scales
Overall, the results of the Muthen multilevel that required a different factor structure at the
confirmatory factor analysis replicated the within-society level of analysis, the multilevel
Table 8.5 ICC(1), Internal Consistency, and ICC(2) Results for Organizational and Societal Culture Scales and the CLT Subscales: Phase 2 Data
08-House.qxd
134
GLOBE Cultural Dimension Scales ORGANIZATIONAL SOCIETAL
PRACTICES (AS IS) ICC(1) Internal Consistency ICC(2) ICC(1) Internal Consistency ICC(2)
3/19/2004
135
08-House.qxd 3/19/2004 4:11 PM Page 136
factor analysis confirmed the earlier discussion strongly suggest that by averaging the responses
that the properties of scales measuring emergent– of at least 45 respondents, our scales exhibited
convergent constructs operate at a particular sufficient reliability to differentiate organiza-
level of analysis.8 tions and societies on our culture and leadership
scales.
Reliability
Global Leadership Scales
As indicated earlier, we assessed the reliabil-
ity of our scales with respect to two random As we expected based on the pilot studies,
error sources. First, we calculated the internal there were significant interrelationships among
consistency of our scales at the organizational the leadership attribute scales. Thus, we con-
or societal level of analysis to assess the degree ducted a second-order maximum likelihood
to which our scales were free from error caused exploratory factor analysis. Using a number of
by item variability. To calculate these internal criteria to determine the number of factors to
consistency estimates, we first aggregated the extract (i.e., meaningfulness of the factor struc-
items comprising each scale to the organiza- ture; all retained factors account for more vari-
tional or societal level of analysis and computed ance than a single item; in other words, the
Cronbach’s alpha on the basis of these averaged eigenvalue of a retained factor is greater than
item responses. Table 8.5 shows the internal one, etc.), we settled on a six-factor solution. The
consistency estimates for our scales. The aver- factors that were identified are (a) Charismatic/
age Cronbach alpha for the organizational cul- Value-Based leadership, (b) Team-Oriented lead-
tural practices (As Is) and the organizational ership, (c) Autonomous leadership, (d) Humane-
cultural values (Should Be) scales were 0.61 Oriented leadership, (e) Participative leadership,
and 0.55, respectively. The average Cronbach and (f) Self-Protective leadership. We refer to the
alpha for the society cultural practices (As Is) 6 second-order factors as CLT leadership dimen-
and society cultural values (Should Be) scales sions and to the 21 subscales as basic CLT lead-
were .77 and .75, respectively. Finally, the aver- ership dimensions.12 The six global leadership
age internal consistency for the 21 leadership dimensions are found in Table 8.6. The internal
attribute subscales was .75. Although there were consistency reliability and interrater reliability
a few organizational culture scales that had lower for these CLT leadership dimensions (computed
Cronbach alphas than desired, the majority of our by using the linear composite reliability formula
scales exhibited adequate internal consistency.9 provided in Nunnally and Bernstein, 1994) are
Second, we explored the extent to which our also shown in this table. As seen in the table,
scales exhibited interrater reliability. In particular, the reliability of the scales designed to measure
we assumed that an average of 45 responses10 these global CLT dimensions are very acceptable
per organization or society were available11 and (i.e., average internal consistency reliability = .84;
computed ICC(2) (Shrout & Fleiss, 1979) for average interrater reliability = .95).
our scales to assess the extent to which our
scales exhibited interrater reliability. The ICC(2)s
Construct Validation
for our scales are also shown in Table 8.5. As
shown in this table, even with an assumed aver- The aforementioned results provide strong
age of 45 responses per organization or society, evidence that the GLOBE scales were construc-
the interrater reliability of our scales was quite ted properly, that they measured variation at
sizable. The average ICC(2) for the organiza- the desired level of analysis, and that they could
tional cultural practices (As Is) and cultural values be used to reliably make distinctions between
(Should Be) scales were .92 and .94, respec- organizations or societies. Although these results
tively. The average ICC(2) for the societal cul- document the psychometric properties of our
tural practices (As Is) and cultural values (Should scales, it should also be noted that these results
Be) scales were .93 and .95, respectively. The also provide some initial information about the
average ICC(2) for the leadership attribute construct validity of our scales. In this section,
subscales was .90. In summary, these results we more closely examine the construct validity
08-House.qxd 3/19/2004 4:11 PM Page 137
NOTE: The numbered, italicized topics are Global CLT leadership dimensions. They comprise first-order leadership attribute
(CLT) scales. Numbers following Global CLT names represent the range of country-level mean values on a seven-point scale
ranging from 1 (greatly inhibits) to 7 (contributes greatly) to outstanding leadership. Internal consistency reliability for the
Global CLTs were obtained by using the internal consistency of the first-order leadership attribute scales (Table 8.4) and
applying Nunnally and Bernstein’s (1994) formula for unit-weighted linear composite scores. Interrater reliability estimates
were obtained in a similar fashion except that the ICC(2) estimates were used instead of the internal consistency estimates for
these first-order leadership attribute scales.
several cross-cultural researchers have argued scales (i.e., range of correlations for the societal
that interpretation of the rank order of cultures cultural practices (As Is) scales: .93 to .98; aver-
based on average scale scores is problematic. age correlation for societal cultural practices
These researchers argue that some correction is (As Is) scales: .95; range of correlations for the
needed to minimize the influence of this bias. societal cultural values (Should Be) scales: .86 to
Although a multimethod approach is probably .98; average correlation for the societal cultural
the optimal manner to minimize the influence of values (Should Be) scales: .93). The magnitude
this bias (Triandis, 1994), a statistical standardiza- of these correlations implies that cultural-response
tion correction procedure has been developed that bias plays a small role in our culture scales.
is believed to remove cultural-response biases We also correlated the corrected and uncor-
from the original “uncorrected” questionnaire rected first- and second-order leadership attribute
responses (van de Vijver & Leung, 1997). In the scales. Once again, a substantial level of agree-
GLOBE project, we performed this statistical ment existed between the corrected and uncor-
correction procedure and correlated our original rected first-order (range of correlations: .85 to .99;
scales with these “response bias corrected” scales average correlation: .93) and second-order
to ascertain the extent to which response bias was (range of correlations: .87 to .98; average corre-
a problem with our scales. Further, we developed lation: .95) leadership scales. This magnitude
an extension of this procedure to help interpret of these correlations once again implies that
the rank ordering of the countries on the basis of cultural-response bias plays a small role in our
this correction. This extension of the traditional leadership scales.
response bias correction procedure and its ability
to identify specific societies that exhibit substan-
Nomological Network:
tial response bias is discussed in Appendix B.
Correlation of GLOBE’s and
Following Triandis’ (1994) description of the
Hofstede’s (1980) Culture Scales
current statistical correction procedure used in
the cross-cultural literature, we calculated each As we noted previously, three of the
individual’s mean and standard deviation across culture dimensions that we measured are
all items in the survey. It is critical when com- direct descendants of the culture dimensions
puting these means and standard deviations identified by Hofstede (1980). Specifically, we
that the items measure a wide range of con- developed scales to measure Uncertainty
structs. Only if this is true will an individual’s Avoidance, Power Distance, and Individualism–
mean and standard deviation reflect response Collectivism, with an additional scale to
bias tendencies as opposed to responses to some measure collectivism on the basis of Triandis’
meaningful construct. After computing these work. (Our Gender Egalitarianism scale was
values, an individual’s corrected item scores not intended to reflect the same construct as
are computed by subtracting that individual’s Hofstede’s Masculinity scale, though there
average response from the actual response for are some similarities between our Assertive-
that item and dividing this difference by that ness scales and Hofstede’s Masculinity scale.)
individual’s standard deviation. Corrected scale Although this construct similarity between
scores were then created by combining the some of the GLOBE cultural dimensions and
appropriate corrected items and these corrected Hofstede’s dimensions enables quick identifica-
scale scores were then aggregated to the society tion of hypotheses concerning which scales will
level of analysis. exhibit convergent validity, it is important to
To assess the extent to which the uncorrected remember that Hofstede’s scales do not differ-
scale scores were subject to cultural-response entiate cultural practices from cultural values.
bias, we correlated the corrected scale scores that Thus, before generating hypotheses about which
were aggregated to the society level of analysis GLOBE culture scales will converge with
with the aggregated uncorrected scale scores. Hofstede’s scales, it is important to examine
We first correlated the corrected and uncorrected Hofstede’s scales to ascertain which cultural
societal culture scales. These correlations revealed manifestation (i.e., practices or values) is being
a substantial degree of agreement between the measured.
08-House.qxd 3/19/2004 4:11 PM Page 139
With respect to power distance, Hofstede’s In summary, on the basis of this analysis we
scale appears to be focusing on organizational expected the following convergence:
cultural practices. Specifically, two of the three
items in Hofstede’s Power Distance scale used in • Hofstede’s Power Distance scale should be
the IBM study assess cultural practices (i.e., As positively related to GLOBE’s Power Distance
Is) whereas the remaining item appears to be a cultural practices (As Is) scale.
values-based question asking about the type of • Hofstede’s Uncertainty Avoidance scale
manager preferred by the respondents (p. 101). should be modestly positively related to
This mixing of cultural practices and values GLOBE’s Uncertainty Avoidance cultural
within one scale causes some problems in gener- values (Should Be) scale.
ating specific hypotheses about which GLOBE • Hofstede’s Individualism scale should be
scales will correspond to Hofstede’s measures. negatively correlated to GLOBE’s Collec-
However, given that two of the three items tivism I: Institutional Collectivism cultural
appear to measure practices, we decided that this values (Should Be) scale.
scale primarily assesses cultural practices.
With respect to uncertainty avoidance, We did not generate a specific hypothesis
Hofstede’s scale included three items, the first for the Masculinity scale because the GLOBE
of which is clearly a values item: “Company rules scales were not designed to directly build on
should not be broken even when the employee Hofstede’s work. Rather, the GLOBE scales were
thinks it is in the company’s best interest” an attempt to refine the concept into two separate,
(p. 164). The second item assesses a behavioral distinctive, and meaningful dimensions. Although
intention or expectation (i.e., the length of time we do not have a specific hypothesis, we explored
that an employee intends to remain with his or the relationships between the Hofstede Masculi-
her current employer), whereas the third item nity and GLOBE Gender Egalitarianism and
assesses an outcome (i.e., the frequency that a Assertiveness scales. In general, we expect that
respondent experiences feeling nervous or tense convergence will occur on the GLOBE cultural
at work). Unfortunately, these last two items values scales of this construct.
do not seem to fit clearly into either a practices We tested our hypotheses by using the country
or values framework. This scale clearly lacks scale scores reported in Hofstede’s (1980) book,
face validity with respect to the construct it is as well as later updates to these country rankings
intended to measure and the label assigned to (Hofstede, 2001), and correlating the country-
it. Interestingly, Hofstede himself noted that this level scores with the country-level scores from
scale was perhaps not the best measure of the the GLOBE project. These findings related to
construct, given that it was created from existing convergent validity are reported in Table 8.7.
items and that the concept was not well defined Consistent with our predictions, there was a
at the time of the scale’s development (1980, significant positive correlation between the
p. 163). Given that there is a single item that Hofstede Power Distance scale and our societal
clearly measures cultural values, we conclude Power Distance cultural practices (As Is) scale
that Hofstede’s uncertainty avoidance scale (r (41) = .61, p < .01). Further, Hofstede’s Power
possibly corresponds with GLOBE’s Uncertainty Distance scale did not significantly correlate
Avoidance cultural values (Should Be) scale, but with the GLOBE societal Power Distance cultural
only modestly. values (Should Be) scale (r (41) = −.03, p > .05).
Hofstede’s individualism scale and his Also consistent with our predictions, there
Masculinity scale appear to be assessing cultural was a significant positive correlation between
values, rather than cultural practices. Further, Hofstede’s Uncertainty Avoidance scale and
because Hofstede’s individualism scale places the GLOBE Uncertainty Avoidance cultural
individualism and collectivism on a single values (Should Be) (r (41) = .32, p < .05) scale.
dimension, we expect that this scale should Somewhat surprising, however, was the signifi-
meaningfully relate to our Collectivism I: cant negative correlation between Hofstede’s
Institutional Collectivism values measure of this Uncertainty Avoidance scale and the GLOBE
construct. cultural practices scale of this dimension
08-House.qxd 3/19/2004 4:11 PM Page 140
Table 8.7 Convergent Validity Coefficients Between GLOBE Scales and Hofstede Scales
Power Distance
Uncertainty Avoidance
Individualism
Masculinity
** = p < .01
(r (41) = −.61, p < .01). The negative correlation GLOBE finding that the cultural practice scales
between these two scales is consistent with are negatively correlated with the cultural
the general GLOBE finding that the cultural values scales.
practice scales are negatively correlated with Finally, with respect to the Hofstede
the cultural values scales. Masculinity scale, the only significant relation-
As we predicted, the Hofstede Individualism ship was with GLOBE’s societal Assertiveness
scale was significantly negatively related to cultural practices (As Is) scale (r (40) = .42,
our societal Collectivism I: Institutional Collec- p < .01).
tivism cultural values (Should Be) scale (r (41) In summary, there was evidence of conver-
= − .55, p < .01) and not to our societal gence between the two sets of scales on many
Collectivism I: Institutional Collectivism cultural but not all dimensions. It is important to keep
practices (As Is) scale (r (41) = .15, p > .05). in mind, however, that contrary to Hofstede’s
With respect to the GLOBE Collectivism II: research, the GLOBE project scales were devel-
In-Group Collectivism construct, the results in oped and psychometrically tested for construct
Table 8.6 show that Hofstede’s Individualism validity as the project progressed from incep-
scale significantly negatively correlates with the tion. In addition, GLOBE set out to assess both
GLOBE practices measure. Once again, this neg- the more objective as well as subjective aspects
ative correlation is consistent with the general of similar dimensions of culture. Indeed, as we
08-House.qxd 3/19/2004 4:11 PM Page 141
have shown, the Hofstede (1980) measures between some of the GLOBE scales and some
combine these two aspects of culture. The of the Schwartz scales, we expected several
GLOBE scales measure these two aspects of specific scales to be correlated. Specifically, we
culture separately and, as shown in the correla- predicted that
tions with Hofstede’s measures, the patterns of
relationships differ depending upon which • Schwartz’s Embeddedness scale would be
aspect of culture was being measured. Thus, positively related to the GLOBE societal
one contribution of the present culture measures Uncertainty Avoidance and Collectivism
is that, for each dimension of culture assessed, cultural values (Should Be) scales
both the practices and values aspects of culture • Schwartz’s Intellectual Autonomy scale would
are measured separately. be negatively related to the GLOBE societal
Uncertainty Avoidance and Collectivism cultural
values (Should Be) scales
Nomological Network:
• Schwartz’s Hierarchy dimension would be
Correlation of GLOBE Scales
positively related to the GLOBE societal Power
With the Schwartz Value Survey
Distance cultural values (Should Be) scale
In 1994, Schwartz extended his individual- • Schwartz’s Egalitarianism dimension would be
level taxonomy of human values to the society positively related to the GLOBE societal Gender
level to identify dimensions that differentiate Egalitarianism cultural values (Should Be) scale
cultures. Following Kluckhohn’s (1951) and but negatively related to the GLOBE societal
Rokeach’s (1973) conceptualization of culture, Assertiveness cultural values (Should Be) scale
Schwartz defined human values as desirable • Schwartz’s Mastery dimension would be
goals that people use as guiding principles in positively related to the GLOBE societal
their lives (Schwartz, 1992, 1994). An individ- Performance Orientation cultural values
ual’s priorities among a set of values is a func- (Should Be) scale
tion of that individual’s unique experiences as
well as the experiences shared by everyone in Schwartz (2001) collected mean ratings on
a given society. Schwartz developed a values his seven dimensions for 64 different countries.
survey that can be used to identify a set of These means were provided by averaging
values that can differentiate individuals from ratings provided by schoolteachers who were
one another (Schwartz & Bilsky, 1987) or that sampled from each culture. Schoolteachers
can differentiate societies from one another. were selected because they, to a substantial
With respect to the societal application of extent, convey societal values across genera-
Schwartz’s survey, seven dimensions were tions. Using the Phase 2 GLOBE data, Sagiv
identified that differentiate societies. His seven and Gupta (2002) found 48 countries in com-
ecological dimensions are Embeddedness mon with the Schwartz list. These researchers
(previously labeled Conservatism), Intellectual correlated the societal-level GLOBE scale
Autonomy, Affective Autonomy, Hierarchy, scores with the Schwartz data. Overall, they
Egalitarianism, Mastery, and Harmony (Schwartz, found good agreement between the GLOBE
1994, 2001; Schwartz & Melech, 2000). Brief cultural values scales and the Schwartz scales.
descriptions of these constructs are provided in The correlations from their study that corre-
Table 8.8. spond to the aforementioned hypotheses are
On the basis of Schwartz’s society-level val- reproduced in Table 8.8.
ues taxonomy, we expected stronger relation- As shown in this table, most of our hypo-
ships between Schwartz’s scales and the theses received some support. More specifi-
GLOBE societal cultural values (Should Be) cally, as predicted, Schwartz’s Embeddedness
scales. In general, we do not expect the GLOBE dimension was significantly positively related
cultural practices (As Is) scales will correlate to the GLOBE Uncertainty Avoidance cultural
with Schwartz’s dimensions because his scales values scale. Schwartz’s Intellectual Autonomy
are clearly assessing cultural values. On the dimension was significantly negatively related
basis of the apparent conceptual overlap to the GLOBE Uncertainty Avoidance cultural
08-House.qxd 3/19/2004 4:11 PM Page 142
values scale. Schwartz’s Hierarchy dimension other measurement approaches (Brewer &
was significantly positively related to the Hunter, 1989).
GLOBE Power Distance cultural values scale. Gupta, Sully de Luque, and House in Chapter
Schwartz’s Egalitarian Commitment dimension 9 describe GLOBE’s development of a set of
was significantly positively related to our soci- unobtrusive measures for the GLOBE culture
etal Gender Egalitarianism cultural values practices dimensions. As discussed in that
measure and negatively related to our societal chapter, the unobtrusive measures are based on a
Gender Assertiveness cultural values scale. content analysis of archival data to generate
It should be noted, however, that not all of measures of societal cultural practices. Gupta
our hypotheses were supported. Schwartz’s and colleagues report that the intercorrelations of
Embeddedness dimension was not significantly the unobtrusive measures and the GLOBE cul-
related to either of the GLOBE Collectivism val- ture practices scales range from a low of .51 for
ues scales. Schwartz’s Intellectual Autonomy the societal Humane Orientation cultural prac-
dimension was not significantly related to either tices (As Is) scale to a high of .65 for societal
of the GLOBE Collectivism values scales. Gender Egalitarianism cultural practices (As Is)
Finally, Schwartz’s Mastery dimension was not scale (Gupta et al., Chapter 9, this volume). In
significantly related to the GLOBE Performance other words, these results show that the GLOBE
Orientation cultural values scale. survey responses provided by middle managers
The pattern of the correlations obtained with in three industries correspond to archival sources
Schwartz’s scales provides additional conver- concerning societal practices. This reinforces the
gent validity evidence for our scales. The corre- conclusion that the construct validity evidence
lations between Schwartz’s scales and our for our scales is not bounded by survey mea-
scales indicate the generalizability of the con- surement methodology. Further, the GLOBE
structs measured by the GLOBE scales given society practices measures and the unobtrusive
the different development approach taken by measures that did not correspond to the GLOBE
Schwartz to develop his scales as well as the measures had lower correlations than those that
different sample (i.e., schoolteachers) used for did correspond to the GLOBE measures.
his country means.
Nomological Network:
Nomological Network: Correlation of Correlation of GLOBE Scales
GLOBE Scales With Unobtrusive Measures With World Values Survey Data
Although the previous analyses provided The construct validity of the GLOBE’s
additional information about the construct societal value scales was assessed by developing
validity of the GLOBE scales, it is important to new culture scales using the data obtained
realize that this evidence all concerned con- from Inglehart, Basanez, and Moreno’s (1998)
structs assessed using survey methodology. World Values Surveys (WVS) project. The
Some of the construct validity evidence may be WVS data consists of responses from approxi-
attributable to the common use of surveys mately 350 questions of human values (Inglehart
across the Hofstede, Schwartz, and GLOBE et al., 1998) ranging from issues involving
studies. Thus, further evidence on the construct family to issues regarding work to topics
validity of the GLOBE measures would be concerning political and social matters. Nine
obtained if we find meaningful relationships separate culture dimensions that were conceptu-
between our scales and nonsurvey measures of ally consistent with the GLOBE cultural values
these culture constructs. In an attempt to assess dimensions were developed. Information about
the convergence of culture constructs indepen- the development of these new scales using
dent of survey methodology, a set of unobtru- the WVS data is provided in Chapter 9. Project
sive measures (i.e., measures that use naturally GLOBE and the WVS data had a total of
occurring data) was developed. Use of unobtru- 39 societies in common.
sive measures has the advantage of being free Gupta and colleagues argue that examining
from the reactive sources of error that can affect the WVS data can be used to test the construct
08-House.qxd 3/19/2004 4:11 PM Page 143
Table 8.8 Convergent Validity Coefficients Between GLOBE Scales and Schwartz’s Scales
Embeddedness
Cultural emphasis on the maintenance of status quo,
minimization of disruptions to traditional order (social order,
family security, respect for tradition)
Intellectual Autonomy
Cultural emphasis on the promotion and protection of individual
pursuit of intellectual directions
Hierarchy
Cultural emphasis on the legitimacy of hierarchical and differential
allocation of financial and social resources
Egalitarianism
Cultural emphasis on the transcendence of selfish interests,
voluntary cooperation, and concern for the welfare of others
Assertiveness −.44**
Mastery
Cultural emphasis on active efforts to modify the social and natural
environment through action
NOTE: We have no hypotheses about Schwartz’s Harmony scale, which refers to the importance of being in tune with the
natural and social world.
** p < .01.
validity of the GLOBE culture scales. In is that unlike the GLOBE scales which were
particular, they argue that this comparison collected using self-report questionnaires, the
would be meaningful because the two projects WVS data were collected using face-to-face
had several methodological differences. For interviews. Further, the response format used in
example, unlike the GLOBE sample, which was the WVS varied across questions. Some ques-
limited to middle managers from three types tions asked for categorical information whereas
of industries, the WVS data came from others asked for respondents to answer using
all adult citizens (18 years or older) from an ordinal-level scale. Finally, the rationale for
each society. Inglehart and colleagues (1998) conducting these two studies differed. The
attempted to create representative samples for WVS data was collected to explore intergener-
each society in their sample. Another difference ation differences in the social, political, and
08-House.qxd 3/19/2004 4:11 PM Page 144
economic values. As discussed earlier in this conceptual model indicates that societal cultural
book, the GLOBE surveys were designed practices should also influence the cultural prac-
specifically for measuring societal cultural tices of an organization.
values and practices constructs. It is for these Given this rationale, additional information
methodological reasons that Gupta and col- for the construct validity of the GLOBE culture
leagues argue a meaningful analysis of the con- scales can be obtained by demonstrating that
vergent and discriminant validity of the GLOBE an objective characteristic (i.e., industrial sector
cultural values scales could be conducted using of an organization) affects the organizational
the WVS data. cultural practices scales but not the societal cul-
Comparing the correlations between the tural practices scales. To assess the hypothesis
GLOBE scales and the new scales that measure that the industrial sector affects the GLOBE
the same culture construct, the convergent practices scales, we separated the GLOBE data
validities were all statistically significant, by industry and correlated the rank order of the
ranging from 0.38 to 0.88 with the average of societies across the three industries using the
0.59. Comparing the correlations between the organizational and societal cultural practices
GLOBE scales and the WVS scales that mea- scales. Our hypothesis about the industry effect
sure different culture constructs, the average on the cultural practices scales would be sup-
discriminant validity was 0.24 and this average ported if the organizational cultural practices
coefficient was significantly less than the scales exhibited low correlations with (i.e., low
average convergent coefficient of 0.59. In other reliability due to sensitivity to industry differ-
words, these results show that the GLOBE ences) across the three industries but the soci-
survey responses provided by middle managers etal cultural practices scales showed high
in three industries correspond to interview correlations (i.e., high reliability due to insensi-
responses provided by a representative sample tivity to industry differences).
of individuals that were independently col- Consistent with expectations, the organiza-
lected. These results provide evidence for tional cultural practices scales were found to
the construct validity of the GLOBE cultural have a low average correlation of 0.38 (i.e.,
values scales. See Chapter 9 by Gupta and average financial–food correlation = 0.41; aver-
colleagues for a more detailed discussion of age financial–telecommunication correlation =
these results. 0.33; average food–telecommunication correla-
tion = 0.39) whereas the societal cultural prac-
tices scales had an average correlation of 0.73
Nomological Network: Sensitivity of
(i.e., average financial–food correlation = 0.88;
GLOBE Scales to Industry Effects
average financial–telecommunication correla-
As discussed earlier in this book, the GLOBE tion = 0.62; average food–telecommunication
conceptual model indicates that the industrial correlation = 0.68). Thus, even though the orga-
sector to which an organization belongs and the nizational and societal practices scales were
common kinds of pressures (e.g., the rate of composed of similar items that only differed in
technological change and the general level of the level of focus of the question (i.e., organiza-
environmental turbulence) encountered by orga- tion versus society), industry differences sub-
nizations in an industrial sector should influence stantially affected the organizational but not the
the kinds of practices adopted in that organi- societal cultural practices scales. These findings
zation. In contrast to organizational practices, provide additional evidence for the construct
however, the practices adopted on a society- validity of our scales.13
wide basis should be relatively unaffected by
the contingencies experienced by organizations
Nomological Network: Construct
in a particular industrial sector. Unless a society
Validity of the CLT Scales
is dominated by a single industry, broader fac-
tors (e.g., shared language, historical trends, The construct validity of the GLOBE CLT
and environmental conditions) influence soci- leadership dimensions cannot be directly
etal cultural practices. Indeed, the GLOBE assessed by a psychometric analysis of the
08-House.qxd 3/19/2004 4:11 PM Page 145
GLOBE data as it was with our culture perspective, as face saver and status consciousness
dimension scales. Given the nature of our may be important if viewed from a non-Western
research design investigating implicit leader- perspective.
ship theory, instead we can provide an indirect
assessment of the meaningfulness of global
CLT leadership dimensions by examining the CONCLUSIONS
relationship of our dimensions to prior leader-
ship literature. Of the six global CLT leadership In summary, as part of the GLOBE project,
dimensions, four are closely related to prior we have developed scales that assess organiza-
leadership constructs found in the extant leader- tional and societal culture and culturally shared
ship literature. The GLOBE Charismatic/Value- implicit theories of leadership. These scales
Based CLT dimension contains three subscales were found to be aggregatable, reliable, and uni-
that embody current aspects of charismatic lead- dimensional. Evidence for construct validity of
ership: developing a vision, inspiring others, the culture scales was provided from several
and engaging in self-sacrificial behaviors. sources.
Integrity and decisiveness, also found in the We do, of course, recognize that there are
Charismatic/Value-Based CLT dimension, have limitations to our research. First, we have inten-
been linked to effective leadership and are con- tionally chosen to focus on perceptions of orga-
sistent with charismatic leadership theory. Our nizational leadership, rather than political or
second major CLT dimension, Team Oriented, moral leadership. Implicit leadership theory pre-
reflects a current interest in the leadership of dicts that the behaviors seen as effective for
teams found in contemporary leadership organizational leaders will be different from the
textbooks (e.g., Yukl, 2002). This dimension behaviors seen as effective for other types of
includes several subscales related to the suc- leaders (Lord & Maher, 1991b; Hanges, Lord,
cessful leadership of teams: collaborative team & Dickson, 2000) or from effective followers
orientation and team integration. It also con- (Leslie and Van Velsor, 1998), and thus our
tains two scales that concern administrative measures may be less applicable to follower
and diplomatic skills that would be consistent perceptions or nonbusiness leadership arenas.
with team effectiveness. The Humane Oriented We suspect that many of our findings will be
CLT leadership dimension may be considered found to be generalizable to other domains. This
directly related to “leadership consideration,” speculation remains to be tested empirically.
a major leadership behavior identified in the A second limitation of the research presented
classic leadership Ohio State and Michigan here is that it is monomethod: All scale data
studies (Bass, 1990). Our fourth major CLT presented were collected using questionnaires.
dimension, Participative leadership, also has a However, the correlations of the scales with
long and important tradition within the leader- unobtrusive measures suggest that this potential
ship literature. issue is not a major concern here. Further, the
In contrast to the previously mentioned data are not subject to common source bias as
CLT dimensions that might be expected given is often the case in monomethod research, as
the leadership literature, we found two dimen- the societal culture and leadership items were
sions that are probably not typically associated completed by different people.
with leadership, at least within the “Western– A third limitation is that we only have
oriented” leadership literature. The first dimen- preliminary information about the construct
sion, Autonomous, reflects an individualistic, validity evidence of the leadership scales.
independent, and unique aspect of leadership. Unlike the societal culture scales, there are
Our second unexpected CLT dimension, Self- no a priori cross-cultural implicit leadership
Protective, would likely be viewed negatively scales that were available to correlate with
from a Western perspective, and has not been our scales. Clearly, further validation of these
previously part of the leadership literature. scales is needed. In a forthcoming anthology,
Nevertheless, aspects of this dimension can eas- chapters from different societies are presented
ily be identified within an “Eastern” leadership in which descriptions of culturally endorsed
08-House.qxd 3/19/2004 4:11 PM Page 146
leadership theories are developed by using in respondents across our societies. For example, in
alternative methodologies (e.g., media analy- contemporary Germany, the term Führer (“leader”) is
ses, focus group discussions). These authors used to describe one certain politician in German
compare the leadership profiles generated by history closely associated with the Holocaust (i.e.,
these alternative methods with the profiles Adolf Hitler), and thus Führer is generally perceived
provided by our scales. The overall result of negatively. Rather, business leaders in Germany are
approximately 25 country specific chapters generally called Führungskraft, Führungsperson, or
provides evidence of construct validity of our Manager. Interestingly, in contemporary Austria, the
leadership scales. German word Führer seemed not to have the same
Finally, it cannot be repeated often enough: association with the Holocaust. In order to have all
These scales were primarily designed to mea- our respondents describe the behaviors and trait
sure convergent–emergent constructs. That is, descriptors of business leaders, the country co-inves-
our primary goal when developing these scales tigators decided that it was necessary to develop a
was to differentiate between organizational and general definition of leadership. Our jointly agreed-
societal cultures. They were not specifically on definition is shown in Table 8.1, and this defini-
designed to measure differences within cultures tion appeared at the beginning of the leadership items
or between individuals. Thus, our scales are section of the survey. It was hoped that this definition
most immediately useful to cross-cultural rather oriented our respondents to a common leader concept
than intracultural researchers. without biasing their responses by inadvertently sug-
gesting that certain types of behaviors or traits are
more desirable than others.
ENDNOTES 3. Research has shown that rwg(J) can produce
values less than zero or greater than 1 (Lindell &
1. An illustration of this problem can be found Brandt, 1997). We recoded any of these “out of
in the team literature with the concept of group bound” values to zero before computing our average
potency (Guzzo, Yost, Campbell, & Shea, 1993). rwg(J) s, as recommended by James, Demaree, and
Group potency refers to the belief of team members Wolfe (1993).
that their team has the competencies to successfully 4. It should be noted that the Pilot Study 2 con-
complete its tasks. Early studies attempting to mea- firmatory factor analyses were performed on the total
sure group potency borrowed items from the individ- variance–covariance matrices (i.e., total or pan-
ual focused self-efficacy literature. Specifically, they cultural level of analysis) instead of the organiza-
asked each team member to rate his or her own tional or societal level of analysis. We performed the
individual task competence. Individual responses to analyses at this level because aggregating the scales
these items were then aggregated to the team level of to the societal level of analysis would result in too
analysis and these average self-efficacy measures few observations (i.e., 12 to 13) to permit use of con-
were believed to be a good indicator of group potency. firmatory factor analytic procedures. Performing
Unfortunately, subsequent research indicated that this the confirmatory factor analysis on total variance–
assumption was invalid (Yost, 1993). Indeed, it is covariance matrices provides a conservative test of
possible to have a team comprised of individuals who the dimensionality of our scales (Muthen, 1989). As
have very healthy beliefs of their own capabilities indicated by Kreft and De Leeuw (1998), the total
(i.e., strong average self-efficacy scores) but who variance–covariance matrix (i.e., the pancultural
have no confidence in the capabilities of the other matrix) contains both between-group level as well
team members (i.e., low group potency) (Yost, 1993). as the within-group level information. The more
This example illustrates that the bias caused by using within-group information contained in our total vari-
the wrong type of item (individualized self-efficacy ance–covariance matrices, the less likely it is that
items) cannot be eliminated even though the data was the between-group factor structure obtained in
properly manipulated (i.e., aggregated to the team Pilot Study 1 would be replicated. However, if the
level of analysis). dimensionality of the scales were replicated in Pilot
2. It became clear fairly early in the GLOBE pro- Study 2, then support for the invariance of the scale
ject, that the term leader activated different concepts could be declared.
08-House.qxd 3/19/2004 4:11 PM Page 147
5. Although obtaining samples from at least two for these three scales does not mean that these scales
industries in each society was our goal, four of the lack measurement equivalence across societies.
societies (i.e., Denmark, El Salvador, Namibia, and Rather, as noted in Endnote 4, lack of fit caused by
Venezuela) provided data from only one industry. measurement nonequivalence cannot be eliminated
6. Cross-cultural researchers typically test the by modifying the factor structure imposed on the
equivalence of a scale by performing a multigroup pooled within-society variance–covariance matrix.
confirmatory factor analysis in which the equivalence Rather, evidence for measurement equivalence across
of a within-society factor structure is compared across societies for these scales is provided because the
multiple societies. Lack of measurement equivalence is same two-factor structure was imposed on the pooled
declared if different factor structures are found across within-society covariance matrix and the fit of the
societies. We did not perform our confirmatory factor model exceeded accepted standards.
analysis in this manner because the convergent– 9. Although the low internal consistency for
emergent nature of our constructs directly implies that some of our organizational culture scales was disap-
the similarity of any within-group factor structure is pointing, we decided to keep these scales in our
not the critical issue. The important issue for scales study. The literature on criterion-referenced tests has
measuring convergent–emergent constructs is the documented the utility of scales exhibiting little, if
appropriateness of a factor structure that accounts for any, internal consistency (e.g., Nunnally & Bernstein,
between organization or societal variation. 1994). All of our scales exhibited some level of inter-
It should be noted, however, that Muthen’s nal consistency. Thus, even the organizational culture
multilevel confirmatory factor analysis provides some scales exhibiting low internal consistency might
information about the equivalence of the within-group prove to be useful for predicting when certain leader-
factor structure. In this statistical procedure, the within ship attributes are culturally endorsed. Clearly, these
variance–covariance matrix is estimated by pooling results suggest that the constructs measured by these
the within-society matrices from each society and scales are slightly broader at the organizational level
imposing a factor structure on this pooled within-group of analysis. Future research should add items to these
matrix. Within-society measurement equivalence is scales to improve the coverage of the content domain
declared with this procedure if (a) the imposed factor at the organizational level of analysis.
structure adequately fits the pooled within-society 10. Generalizability analyses conducted on Pilot
matrix or (b) any lack of fit can be eliminated by mod- Study 2 data indicated that we needed an average of
ifying this common within-society factor structure. 45 respondents per organization or society to have
Lack of fit caused by measurement nonequivalence reliable scaling of our organizations or societies. Our
cannot be eliminated by modifying the factor structure CCIs used this number of respondents as the desired
applied on the pooled within-society variance–covari- minimum sample size from each organization for
ance matrix. However, it should be recognized that the Phase 2 data collection. However, it is important to
most direct assessment of measurement equivalence as note that the Pilot Study 2 data only permitted soci-
it is traditionally conceived in the cross-cultural etal level comparisons and so this target number of
literature would be to test for societal differences in respondents overestimated the actual number needed
the within-society factor structure. Unfortunately, the to achieve reliable organizational level data. As
multilevel confirmatory factor analysis procedure shown in Table 8.4, the organizational level scales
cannot test for these interactions. exhibited acceptable levels of reliability even though
7. It is important to remember when examining we did not have an average of 45 respondents per
ICC(1) values that it indicates the percentage of total organization.
scale score variance attributable to between organiza- 11. It should be noted that we had more than
tion–society differences. It is not surprising therefore, an average of 45 respondents for our society-level
that given all the differences that can occur between scales. Thus, the ICC(2) estimates reported in Table
people within organizations or societies, ICC(1)’s in 8.4 are conservative estimates for the society-level
the 5–20% range indicate fairly powerful effects of culture and leadership scales.
the overall organization or society. 12. We performed the following steps to create
8. It should be noted that imposing a different the six global CLT dimensions. First, all of the first-
factor structure at the within-society level of analysis order CLT scales were standardized. Standardization
08-House.qxd 3/19/2004 4:11 PM Page 148
equalized the variance of the first-order scales and correlation = 0.77). While these results are in the
enabled us to create six unit-weighted composite hypothesized direction (i.e., average correlation for
scores by simply adding the first-order CLT scales the organizational cultural values scales is smaller
specified in Table 8.5 for each global leadership than the average correlation for the societal cultural
dimension. Without this standardization, the summa- values scales), the magnitude of this effect is sub-
tion of the first-order CLT scales would have created stantially reduced.
differentially weighted composite scores in which the
first-order scales with larger variances would have
more influence on the composite (i.e., have a greater REFERENCES
weight) than scales with smaller variances. Second,
we computed the mean and variance for our six Bass, B. M. (1990). Bass and Stogdill’s handbook of
global CLT dimensions by following the linear com- leadership: Theory, research, and managerial
posite formulas specified in Nunnally and Bernstein applications (3rd ed.). New York: Free Press.
(1994). Specifically, these formulas indicate that the Bliese, P. D. (2000). Within-group agreement, non-
mean of a unit-weighted linear composite is the sum independence, and reliability: Implications for
of the means of the unstandardized first-order scales data aggregation and analysis. In K. J. Klein &
used to create that composite. The standard deviation S. W. J. Kozlowski (Eds.), Multilevel theory,
of a linear composite is obtained by adding the vari- research, and methods in organizations:
ance of the unstandardized scales and twice the Foundations, extensions, and new directions
covariances among these scales and taking the square (pp. 349–381). San Francisco, CA: Jossey-Bass.
root of this sum. Third, we multiplied the global CLT Brewer, J., & Hunter, A. (1989). Multimethod
dimension by the calculated standard deviation and research: A synthesis of styles. Newbury Park,
added the calculated mean to create unstandardized, CA: Sage.
unit-weighted global CLT dimensions. Finally, we Brislin, R. W. (Ed.). (1976). Translation: Application
divided these global CLT dimensions by the number and research. New York: John Wiley/Halsted.
of scales used to create them so that the range of the Brislin, R. W. (1986). The wording and translation of
global CLT dimensions was on the original 1 to 7 research instruments. In W. J. Lohner & J. W.
scale. Berry (Eds.), Field methods in cross-cultural
13. Although one might develop a similar research (pp. 137–164). Beverly Hills, CA: Sage.
hypothesis for the organizational and societal cul- Cascio, W. F. (1998). Applied psychology in human
tural values scales, it should be remembered that resource management (5th ed.). Englewood
the data collection strategy used in GLOBE proba- Cliffs, NJ: Prentice Hall.
bly minimizes the effect of industry on the organiza- Dickson, M. (1997). Universality and variation in
tional cultural values scales. Specifically, the CCIs organizationally cognitive prototypes of effec-
identified organizations that represented one of the tive leadership. Unpublished doctoral disserta-
three specified industries and were representative tion, Department of Psychology, University of
of their societal culture. The CCIs were specifically Maryland.
instructed that they could not include multinational Dickson, M. W., Aditya, R. N., & Chhokar, J. S.
corporations. Thus, the strategy used to identify the (2000). Definition and interpretation in cross-
organizations in GLOBE probably increased the cultural organizational culture research: Some
overlap between organizational and societal cultural pointers from the GLOBE research program. In
values. Indeed, the results showed that the average N. Ashkanasy, C. Wilderom, & M. Petersen
correlation for the organizational cultural values (Eds.), Handbook of organizational culture and
scales was 0.74 (i.e., average financial–food correla- climate (pp. 447–464). Thousand Oaks, CA:
tion = 0.80; average financial–telecommunication Sage.
correlation = 0.70; average food–telecommunication Dyer, N., Sipe, W. P., & Hanges, P. J. (1997).
correlation = 0.74) whereas it was 0.77 for the societal Multilevel confirmatory factor analysis:
cultural values scales (i.e., average financial–food Demonstration of Muthen’s technique. Academy
correlation = 0.79; average financial–telecommunication of Management Best Paper Proceedings,
correlation = 0.76; average food–telecommunication 391–394.
08-House.qxd 3/19/2004 4:11 PM Page 149
Foti, R. J., & Lord, R. G. (1987). Prototypes James, L. R., Demaree, R. G., & Wolf, G. (1984).
and scripts: The effects of alternative methods Estimating within-group interrater reliability
of processing information. Organizational with and without response bias. Journal of
Behavior and Human Decision Processes, 39, Applied Psychology, 69, 85–98.
318–341. James, L. R., Demaree, R. G., & Wolf, G. (1993).
Gelfand, M. J., Triandis, H. C., & Chan, K. S. (1996). R wg(J): An assessment of within group interrater
Individualism versus collectivism or versus agreement. Journal of Applied Psychology, 79,
authoritarianism? European Journal of Social 306–309.
Psychology, 26, 397–410. Klein, K. J., Conn, A. L., Smith, D. B., & Sorra, J. S.
Guzzo, R. A., Yost, P. R., Campbell, R. J., & Shea, (2001). Is everyone in agreement? Exploring
G. P. (1993). Potency in groups: Articulating a the determinants of within-group agreement
construct. British Journal of Social Psychology, in survey responses. Journal of Applied
32, 87–106. Psychology, 86, 3–16.
Hanges, P. J., Dyer, N., & Sipe, W. P. (1999, April). Klein, K. J., Dansereau, F., & Hall, R. J. (1994).
Multilevel confirmatory factor analysis: Levels issues in theory development, data col-
Demonstration of Muthen’s technique. In lection, and analysis. Academy of Management
S. Palmer, G. Ziets, & P. Bliese (Chairs), Review, 24, 308–324.
Advances in multilevel research: New tech- Kluckhohn, C. (1951). The study of culture. In
niques and methodological issues. Symposium D. Lerner & H. D. Lasswell (Eds.), The policy
conducted at the meeting of the Society of sciences (pp. 86–101). Stanford, CA: Stanford
Industrial and Organizational Psychology, University Press.
Atlanta, GA. Kozlowski, S. W. J., & Klein, K. J. (2000). A multi-
Hanges, P. J., House, R. J., Dickson, M. W., level approach to theory and research in organi-
Dorfman, P. W., & GLOBE. (2003). The devel- zations: Contextual, temporal, and emergent
opment and validation of scales to measure soci- processes. In K. J. Klein & S. W. J. Kozlowski
etal and organizational culture. Manuscript (Eds.), Multilevel theory, research, and methods
submitted for publication. in organizations: Foundations, extensions, and
Hanges, P. J., Lord, R. G., & Dickson, M. W. (2000). new directions (pp. 3–90). San Francisco, CA:
An information-processing perspective on lead- Jossey-Bass.
ership and culture: A case for connectionist Kreft, I., & De Leeuw, J. (1998). Introducing multi-
architecture. Applied Psychology: An Inter- level modeling. Thousand Oaks, CA: Sage.
national Review, 49, 133–161. Leslie, J. B., & Van Velsor, E. (1998). A cross-
Hofstede, G. (1980). Culture’s consequences: Inter- national comparison of effective leadership
national differences in work-related values. and teamwork: Toward a global workforce.
London: Sage. Greensboro, NC: Center for Creative Leadership.
Hofstede, G. (2001). Culture’s consequences: Lindell, M. K., & Brandt, C. J. (1997). Measuring
Comparing values, behaviors, institutions and interrater agreement for ratings of a single
organizations across nations (2nd ed.). Thousand target. Applied Psychological Measurement,
Oaks, CA: Sage. 21, 271–278.
Hui, C. C., & Triandis, H. C. (1989). Effects of Lord, R. G., Foti, R., & DeVader, C. (1984). A test
culture and response format on extreme of leadership categorization theory: Internal
response style. Journal of Cross-Cultural structure, information processing, and leader-
Psychology, 20, 296–309. ship perceptions. Organizational Behavior and
Inglehart, R., Basanez, M., & Moreno, A. Human Performance, 34, 343–378.
(1998). Human values and beliefs: A cross- Lord, R. G., & Maher, K. J. (1991a). Cognitive
cultural sourcebook. Ann Arbor: University of theory in industrial and organizational psychol-
Michigan. ogy. In M. D. Dunnette & L. M. Hough (Eds.),
James, L. R. (1982). Aggregation bias in estimates of Handbook of industrial and organizational psy-
perceptual agreement. Journal of Applied chology (2nd ed., Vol. 2, pp. 1–62). Palo Alto,
Psychology, 76, 219–229. CA: Consulting Psychologists Press.
08-House.qxd 3/19/2004 4:11 PM Page 150
Lord, R., & Maher, K. J. (1991b). Leadership and Schwartz, S. (1992). Universals in the content
information processing: Linking perceptions and structure of values: Theoretical advances
and performance. Boston: Unwin-Hyman. and empirical tests in 20 countries. In M. Zanna
Messick, S. (1981). Constructs and their vicissitudes (Ed.), Advances in experimental social
in educational and psychological measurement. psychology (Vol. 25, pp. 1–65). New York:
Psychological Bulletin, 89, 575–588. Academic.
McGraw, K. O., & Wong, S. P. (1996). Forming Schwartz, S. (1994). Beyond individualism-
inferences about some intraclass correlation collectivism: New cultural dimensions of val-
coefficients. Psychological Methods, 1, 30–46. ues. In U. Kim, H. C. Triandis, C. Kagitçibasi,
Muthen, B. O. (1989, June). Latent variable model- S. C. Choi, & G. Yoon (Eds.), Individualism
ing in heterogeneous populations. Presidential and collectivism: Theory, method and applica-
address to the Psychometric Society, Princeton, tions (pp. 85–119). London: Sage.
New Jersey. Schwartz, S. (2001, April). Antecedents and conse-
Muthen, B. O. (1990, June). Mean and covariance quences of national variation on three cultural
structure analysis of hierarchical data. Paper dimensions. Keynote address at the meeting
presented at the Psychometric Society, Princeton, of Institute for Research on Intercultural
New Jersey. Cooperation, Tilburg, The Netherlands.
Muthen, B. O. (1994). Multilevel covariance struc- Schwartz, S., & Bilsky, W. (1987). Toward a univer-
ture analysis. Sociological methods & research, sal psychological structure of human values.
22, 376–398. Journal of Personality and Social Psychology,
Nunnally, J. C., & Bernstein, I. H. (1994). Psycho- 53, 550–562.
metric theory (3rd ed.). New York: McGraw-Hill. Schwartz, S., & Melech, G. (2000). National differ-
Payne, R. L., Fineman, S., & Wall, T. D. (1976). ences in micro and macro worry: Social, eco-
Organizational climate and job satisfaction: A nomic, and cultural explanations. In E. Diener &
conceptual synthesis. Organizational Behavior E. Suh (Eds.), Subjective well-being across
and Human Performance, 16, 45–62. cultures and nations (219–256). Boston: MIT
Roberts, K., & Boyacigiller, N. (1984). Cross- Press.
national organizational research: The grasp of Shrout, P. E., & Fleiss, J. L. (1979). Intraclass corre-
the blind man. In B. M. Staw & L. L. Cummings lations: Uses in assessing rater reliability.
(Eds.), Research in organizational behavior Psychological Bulletin, 86, 420–428.
(Vol. 6, pp. 423–475). Greenwich CT: JAI. Smith, P. B. (2002). Culture’s consequences:
Rokeach, M. (1973). The nature of human values. Something old and something new. Human
New York: Free Press. Relations, 55, 119–124.
Sagiv, L., & Gupta, V. (2002, August) Schwartz Smith, P. B., & Schwartz, S. H. (1997). Values. In
theory of cultural dimensions of values. In J. Berry, M. Segall, & C. Kagitçibasi (Eds.),
V. Gupta & R. J. House (Chairs), Reconciling Handbook of cross-cultural psychology (Vol. 3,
the findings of five cross-cultural research pro- pp. 77–118). Boston, MA: Allyn & Bacon.
grams: Hofstede, Trompenaars, World Values Stening, B. W., & Everett, J. E. (1984). Response
Survey, Schwartz and GLOBE. Symposium styles in a cross-cultural managerial study.
conducted at the meeting of the Academy of Journal of Social Psychology, 122, 151–156.
Management, Denver, CO. Triandis, H. C. (1994). Cross-cultural industrial
Schneider, B. (1990). The climate for service: An appli- and organizational psychology. In H. C.
cation of the climate construct. In B. Schneider Triandis, M. D. Dunnette, & L. M. Hough
(Ed.), Organizational climate and culture (pp. (Eds.), Handbook of industrial and orga-
393–412). San Francisco: Jossey-Bass. nizational psychology (2nd ed., Vol. 4,
Schriesheim, C. A. (1979). The similarity of individ- pp. 103–172). Palo Alto, CA: Consulting
ual directed and group directed leader behavior Psychologists Press.
descriptions. Academy of Management Journal, Triandis, H. C. (1995). Individualism and collec-
22, 345–355. tivism. Boulder, CO: Westview.
08-House.qxd 3/19/2004 4:11 PM Page 151
van de Vijver, F., & Leung, K. (1997). Methods and Yost, P. R. (1993). Individual versus group
data analysis for cross-cultural research. performance: I can’t but we can. Unpublished
Thousand Oaks, CA: Sage. master’s thesis, University of Maryland, College
Yammarino, F. J. (1990). Individual- and group- Park.
directed leader behavior descriptions. Educational Yukl, G. A. (2002). Leadership in organizations
and Psychological Measurement, 50, 739–759. (5th ed.). Upper Saddle River, NJ: Prentice Hall.
09-House.qxd 3/19/2004 4:12 PM Page 152
9
MULTISOURCE CONSTRUCT
VALIDITY OF GLOBE SCALES
VIPIN GUPTA
ROBERT J. HOUSE
n this chapter we assess the construct validity of societal practices to clarify the thematic
AUTHORS’ NOTE: The authors wish to thank Klaus Krippendorff for his review and helpful comments on the
early part of this study. In addition, we gratefully acknowledge the dedicated support provided by Sally Chan,
James Zale, Danielle Rizk, Tom Spies, and Julia Kang at various times during the preparation of this manuscript.
Without their assistance this research would not have been possible. We are grateful to Paul Hanges for his
invaluable guidance, suggestions, and reviews of this chapter.
152
09-House.qxd 3/19/2004 4:12 PM Page 153
We begin the discussion in the first part of the system, such as the World Bank. These agencies
chapter by examining the validity of the GLOBE systematically collate a wide variety of data from
societal-cultural practice constructs using unob- various national sources and try to provide these
trusive measures derived from a content analysis in a comparable format for policy and academic
of Culturgrams. In the second part of the chapter, analysis. In addition, several researchers and
we then examine the validity of the GLOBE soci- research organizations publish reference data
etal cultural value constructs using questionnaire on civil liberties, corruption, climate, and other
scales derived from the WVS. variables (Parker, 1997). We obtained such
archival data from several well-recognized cross-
cultural sources: Kurian (1997), Sullivan (1991),
UNESCO (1997), the World Bank (1997), and
DEVELOPING UNOBTRUSIVE Parker (1997). Selected items that showed signif-
MEASURES OF THE GLOBE icant correlations with the GLOBE aggregated
SOCIETAL CULTURAL PRACTICES SCALES society practices data are given in Table 9.1.
Table 9.1 Correlation of Archival Data with Aggregated Society Practice Scale Scores1
NOTE: N varies from 40–54. ***p < 0.001. **p < 0.01. *p < 0.05.
1
There is no effect of GNP, per capita income, population size of country, or year of independence on the correlations between
the unobtrusive measures and their isomorphic-questionnaire-based measures. However, selected subsamples such as more
or less wealthy countries, indicated differences in correlations suggesting boundary conditions for some of the unobtrusive
measures.
09-House.qxd 3/19/2004 4:12 PM Page 155
Test for
Correspondence
PHENOMENA
Known Content
Method Analysis
DATA
non-GLOBE societies. Krippendorff (1980) humane culture, (c) whether visiting was a
suggests involving experts who are familiar with warm and sincere gesture (more humane) or
the theoretical linkages between the data and the was an expected ritualistic act (more impersonal
concepts of study in this search for contextual and therefore not counted as humane), and
knowledge. Thus, two members of the research (d) whether other implicit or explicit concepts
team inspected the text data, each developing throughout the text contradicted or diminished
construct-inference examples independently. the importance of visiting.
From this we identified culturally relevant Examining implicit concepts allowed for
concepts related to each of the nine GLOBE nonobvious concepts to surface. In addition,
constructs.5 when analyzing concepts for Institutional
Collectivism, the macro or societal level was a
central feature of this dimension. Implicit con-
Strategy Development cepts indicating Institutional Collectivism in the
and Definition of Measures Culturgram text included such phenomena as
group-oriented activities that were encouraged
Developing Strategies or sponsored by formal organizations such as
Several strategies for conducting content government agencies, foundations, or educational
analyses have been suggested (Roberts, 1997). institutions. Examples of unobtrusive indicators
We chose conceptual analysis for our investiga- of Institutional Collectivism include high num-
tion. Conceptual analysis, or thematic analysis bers of sponsored team sports, social clubs, or
(Stone, 1997), commonly focuses on the detec- recreational centers, all of which imply a form of
tion of specific words or concepts, as well as the institutional collectivism other than families.
identification of patterns or themes in the text. These examples illustrate that making
Carley (1994) defines a concept “as a single implicit concepts explicit and then quantifying
idea regardless of whether it is represented by a these concepts is important because it results
single word or phrase” (p. 726). in the construction of a translation rule. These
Two ways of analyzing concepts have been rules established our confidence in the coding
noted in the research: explicit concept analysis and ensured coding consistency throughout
and implicit concept analysis (Carley, 1994). the text (Carley, 1994). Translation rules guide
Actual words and phrases that appear in the coders to record various concepts in a precise
text represent explicit concepts. Implied words manner. The coders would discuss the con-
and phrases that occur in the text indicate cepts—in this case the cultural activity related
implicit concepts. By using explicit concepts to the identified themes—and develop instruc-
alone, researchers may overlook more indirect tions (or rules) on how to note the occurrence of
nuances in their sources of content analysis the concept. Other important elements of this
data. Implicit analysis allows greater ability to stage of the content-analysis process are defin-
extract deep meaning from the text. Thus, we ing measurements6 and defining categories7.
used both implicit and explicit concept analysis The unobtrusive items for each of the GLOBE
in our research. society practices are presented in Table 9.2.
Examining embedded concepts within the
text proved to be an important advantage in
Coding for Unobtrusive Indicators
this analysis. For example, “visiting one
another” was identified as an important indica- The next step in the content analysis process
tor of the Humane Orientation dimension. In the involved the actual coding of the text. Coding
Culturgram text, the implicit messages regard- rules were specified and manual coding began.
ing visiting included such considerations as Manual coding was essential for this analysis
(a) whether the importance of visiting was the in that much of the coding extracted implicit
primary variable mentioned in the section on information. Extensive training in identifying
the topic, implying a more humane culture, unobtrusive indicators was provided for the
(b) whether the author was emphatic about coders to assure that they made appropriate
the importance of visiting, implying a more qualitative judgments about the contents of
09-House.qxd 3/19/2004 4:12 PM Page 157
the Culturgram text. All unobtrusive indicators This enhanced the reliability of the study
were rated on a 5-point scale, with 5 represent- because it promoted reproducibility of the find-
ing high scores on the indicators. Using this type ings through the use of multiple raters.
of scale required conceptualizing the source From the Culturgrams, a sample of 10
material on a continuum (Krippendorff, 1980), countries not in the GLOBE sample was coded
with the final dimension scores consisting of on each of the nine GLOBE practice dimen-
the overall averages of the dimension indica- sions. The average initial results were compared
tors. Two experts, each of whom had a PhD in across the 10 societies on all unobtrusive
management, coded all societies independently. indicators. In both the first and second rounds
09-House.qxd 3/19/2004 4:12 PM Page 158
Table 9.3 Interrater Correlation and Internal Consistency Reliability Estimates for Unobtrusive
Measures
* Based on the Spearman-Brown Prophecy formula (Nunnally & Bernstein, 1994). ** p < 0.01.
of preliminary coding based on cultures not of the averaging of rater responses and are also
in the GLOBE study, items with lower coding provided in Table 9.3.
agreement between coders were discussed, One way to assess validity of unobtrusive
other discrepancies were sorted out, and recod- measures is to determine if the measured data
ing was conducted until an acceptable level of are related to other analyses of the construct
agreement was achieved. in ways predicted by the theory (Bowen and
This rigor in coding the sample texts clarified Bowen, 1999). Two primary aspects of validity
coding rules and weaved information into the are convergent validity and discriminant valid-
classification process (Weber, 1990). After the ity (Campbell & Fiske, 1959; Weber, 1990). For
coders reached an acceptable level of agreement, this part of our study, convergent validity
the reliability of the coding and the validity of is represented as the agreement, or correlation,
the scales could be measured. between the unobtrusive measures and the
corresponding GLOBE measures to assess soci-
etal-level cultural practices through text analysis
Assessing Quality and Accuracy
and survey questionnaire responses, respec-
In this study, reliability was assessed as the tively. A dimension has discriminant validity if
consistency with which multiple raters coded it is relatively distinct from other criteria vari-
the Culturgram texts. Table 9.3 gives the inter- ables. Validity may best be clarified through
rater correlation for each of the nine unobtrusive the use of a multitrait–multimethod matrix
measurement scales. The scale score for each (Campbell & Fiske, 1959). For this study, the
dimension consists of scores from two raters. multiple traits consisted of the nine cultural
The average interrater correlation was 0.70 dimensions. The two methods consisted of the
across nine unobtrusive measure scales, which questionnaire-based method (GLOBE societal-
is within the accepted reliability range level practice scales) and the content-coding
(Nunnally & Bernstein, 1994). In addition, method (unobtrusive measures of culture).
using the Spearman-Brown Prophecy formula Table 9.4 provides the complete multitrait–
we produced reliability estimates for each of multimethod matrix. Table 9.5 provides a
the average unobtrusive scores. These consist summary of this information.
Table 9.4 The Multitrait-Multimethod Matrix of Societal Practice Measures
09-House.qxd
A1 B1 C1 D1 E1 F1 G1 H1 I1 A2 B2 C2 D2 E2 F2 G2 H2 I2
3/19/2004
Method 1
A1 * (.90)
B1. .10 (.86)
4:12 PM
Method 2
A2. .56* −.06 .31* .13 .10 .04 −.03 −.26* .35** (.72)
B2 .15 .59** .04 .07 −.12 .26* .03 .15 .09 .05 (.75)
C2 .63** .11 .62** .26* .12 −.13 .15 −.46** .47** .62** .08 (.80)
D2 −.13 −.04 −.25 .51** .03 .17 −.23 .08 −.30* .25** −.42** .07 (.88)
E2 .30* .20 .27* −.06 .58** −.11 .02 −.22 .26* .42** −.38** .47** .41** (.67)
F2 −.47** −.17 −.59** −.21 −.17 .58** −.44** .61** −.62** −.11 .13 −.40** .27** .11 (.77)
G2 .10 .04 .10 .10 .03 −.27* .65** .01 .04 −.30* −.10 −.07 −.14* −.03 −.23 (.66)
H2 −.49** −.09 −.43** −.26** −.21 .37** −.26* .62** −.36** −.33** .21 −.44** −.15** −.30* .61** −.32* (.80)
I2 .51** .15 .54** .22 .08 .30* .29* −.60** .60** .58** −.08 .75** .01 .38** −.59** −.06 −.48** (.88)
NOTE: Letters A, B, C, D, E, F, G, H, and I refer to statuses on the dimensions of culture. Subscripts 1 and 2 refer to the 2 methods used in this study. Method 1 refers to the Unobtrusive
items and Method 2 refers to the GLOBE items. Validity coefficients are the diagonal set of boldface numbers; reliability coefficients are the numbers in parentheses along the principal
diagonal. For this study, the reliability coefficient for the unobtrusive items was calculated by aggregating two raters to form the scores for the dimensions, then using the Spearman-Brown
159
Prophecy formula. A = Performance Orientation. B = Assertiveness. C = Future Orientation. D = Humane Orientation. E = Institutional Collectivism. F = In-Group Collectivism. G = Gender
Egalitarianism. H = Power Distance. I = Uncertainty Avoidance. * p < 0.05 ** p < 0.01
09-House.qxd 3/19/2004 4:12 PM Page 160
1 2a 2b 3 4 5 6
GLOBE Coefficients
3/19/2004
Orientation
Assertiveness 0.59 0.11 0.11 0.12 0.18 0.86 0.75
Future 0.62 0.29 0.32 0.40 0.36 0.96 0.80
Orientation
Page 161
NOTES: Summary conditions for construct validity: 1> 2(a), 2(b), 3, & 4; 5>3; 6>4. d
U (Discriminant coefficient) = Average absolute correlation of an unobtrusive measure with
a other eight unobtrusive measures.
G,U (Convergent coefficient) = Correlation of the GLOBE measure with the unobtrusive
measure. e
G (Discriminant coefficient) = Average absolute correlation of a GLOBE measure with
b other eight GLOBE measures.
G,U (Discriminant coefficient) = Average absolute correlation of a GLOBE measure with
the nine unobtrusive measures. f
U (Reliability) = Reliability of an unobtrusive measure.
c g
U,G (Discriminant coefficient) = Average absolute correlation of an unobtrusive measure G (Reliability) = Reliability of a GLOBE measure.
161
with the nine GLOBE measures.
09-House.qxd 3/19/2004 4:12 PM Page 162
the GLOBE practice scores demonstrate that the of women, and vice versa (Kellehear, 2001).
managerial responses reflect the society in which Accordingly, unobtrusive measures are best
the respondents are embedded, not merely the used in conjunction with other variables.
cultures of middle managers. With respect to the use of archival data or
Each research method used to study culture data from questionnaires, interviews, and census
has its strengths and limitations. A general reports for cultural analysis, other limitations
issue associated with the study of culture is the are notable. Archival data are often gathered
inherent multilayered complexity of the subject for alternative purposes, which then define the
matter. As a result, the research methods used scope of samples and the choice of variables.
must balance depth and sensitivity with breadth; These data can help infer relationships among
in this manner, researchers can detect both variables within and across cultures. However,
fine cultural nuances and develop a sense of the the specific inferences derived from secondary
overall cultural context. Clearly, assessing the analysis may not be relevant for each subgroup
culture dimensions independently, in isolation within or across cultures.
from the other dimensions, may lead to prema-
ture conclusions and mistaken assumptions of QUESTIONNAIRE VALIDATION MEASURES
culture. Examining the dimensions in tandem,
FOR GLOBE SOCIETAL VALUE SCALES
however, may provide an enhanced appraisal of
cultures studied.
The construct validity of the GLOBE’s societal
With respect to the use of unobtrusive mea-
value scales based on “Should Be” response
sures as a research method, two issues surface
formats was assessed by correlating the GLOBE
as to how intervening variables may affect the
societal-level value scales with other indices
relationships between the unobtrusive measures
created from the World Values Surveys (WVS)
and the GLOBE scores. The first issue addresses
(Inglehart, Basanez, & Moreno, 1998). We used
whether intervening variables may exist when
selected questionnaire items of the WVS,
the unobtrusive measure is a good measure of
described below as “outcropping measures.”
the construct in question. The second issue
In this section, we seek to assess the validity
focuses on what happens when the unobtrusive
of the nine GLOBE societal cultural value
measure itself is not capturing everything it
dimensions.
should and, as a result, is not fully reflective of
The World Values Survey includes responses
the construct. For instance, a study of recycling
across societies covering a range of economic,
behavior across cultures using post hoc behav-
political, and cultural issues, from approxi-
ior data could overlook the fact that in some
mately 350 questions relevant to human values
societies there tends to be considerable in-home (Inglehart et al., 1998). Questions in the WVS
recycling, such as use of empty bottles and cans range from questions involving self and family,
for storage purposes. By focusing on specific to issues regarding work and jobs, to topics con-
manifestations of recycling behavior, it might cerning political and social matters. The World
be difficult to obtain a true and appropriate Values Surveys were administered to respon-
appraisal of environmental and waste sensitivity, dents in more than 70 societies, of which 39
which may be considered a proxy for future overlap with the 62 societies in the GLOBE
orientation. Clearly, considerable time and sample. The WVS raw data are publicly avail-
effort must be spent to ensure that the unobtru- able. For these reasons, we considered the WVS
sive measures used are appropriate indicators of data to be adequate data from which we could
the constructs being examined. construct scales relevant to the GLOBE societal-
Another limitation of unobtrusive measures level value scales listed in Chapter 2 of this
relates to selective recording of data. Researchers volume.
with various interests, biases, and experiences
may focus on certain objects and relation-
The World Values Survey Methodology
ships. For instance, female observers of men
may document different features of communica- The WVS survey methodology for measuring
tion, gestures, and clothing than male observers values differs substantively from that of GLOBE
09-House.qxd 3/19/2004 4:12 PM Page 163
in five major ways. First, the sampling universe respondent. Thus, all aggregates assumed the
of the World Values Surveys consisted of all underlying construct to be nominal. Because
adult citizens, ages 18 and older. Representative the values covered by the WVS concerned a
samples were used in most cases. To accomplish broad array of human beliefs, we expected sub-
this, a random selection of locations was made in stantial differences in the variances among the
each society, and then individuals were sampled items. Therefore, we standardized all the items
in each location. The available World Values to a mean of 0 and a standard deviation of 1, so
Survey data also contain weights to make the that an equal weight was given to all items used.
samples replicate the national population para- The GLOBE instrument contained items in a
meters more closely8. Project GLOBE uses a standard 7-point response format. To construct
more concentrated sample of middle-level man- the WVS cultural scales that reflected GLOBE
agers from three industries (food processing, societal-level values scales, we used a simple
financial services, and telecommunication). aggregated mean of items selected to corre-
Second, the data collection method varied spond to the GLOBE scales.
from that of Project GLOBE. The World Values A fifth and final difference between the
Surveys were carried out through face-to-face two survey sources concerns the purpose of
interviews. GLOBE surveys were given to the the surveys. The World Values Surveys were
respondents to complete, in some cases in the intended to test the intergeneration differences
presence of the investigator and some cases not. in the social, political, and economic values,
A third difference between the two research pro- arising as a result of the processes of postindus-
jects was that the timeframe for data collection trialism. The GLOBE surveys were designed
differed substantially for each. The WVS data specifically to develop and measure middle
have been collected at four different times: in the managers’ reports of selected societal cultural
early 1980s, in the early 1990s, in the late 1990s, values and practices and other related cultural
and in the early 2000s. All the nations sampled in and leadership constructs.
the early 1980s were sampled again in one of the Given the substantive differences of these
two 1990s questionnaire administrations9. We sources, we determined that the WVS data could
elected to use the combined data from the two be used as a source to establish convergent and
1990s samples for the purposes of validation discriminant validity of GLOBE’s societal value
because Inglehart & Baker (2000) showed that scales. Consistent with our tests of unobtrusive
although there were some shifts in the values of measures, we used the validation procedures
societies over the two 1990s questionnaire recommended by Campbell and Fiske (1959).
administrations, those shifts were not drastic.
GLOBE data, in contrast, were collected over a Hypotheses for the Development
single period of time, between 1995 and 1997. of the Outcropping Measures of the
A fourth difference between the two projects
GLOBE Societal Value Dimensions
was that unlike the survey instrument used for
data collection in Project GLOBE, the WVS We created hypotheses for the development
instrument contains items in a variety of of WVS scales for the nine GLOBE societal
nonstandardized formats. Both nominal and value dimensions. These value dimensions are
ordinal scales were used in the WVS data col- defined in Chapter 2 and are discussed in detail
lection; some questions have yes–no answers, in Chapters 12 through 19. The theoretical con-
whereas others have a 3-point, 5-point, 7-point, nections outlined in the following sections
or 10-point response format. For the purposes of describe the inference linkages between the
the GLOBE study, we took an average of the GLOBE societal value dimensions and the
responses of each item for each society. In some selected WVS questionnaire items. The indirect
questions, the respondent had to choose a first and implicit inference linkages concerning the
and second preference from a group of four WVS items are based on “outcroppings” asso-
items. We assigned a score of 2 to the first pref- ciated with the core GLOBE value dimen-
erence, a score of 1 to the second preference, sions. Outcroppings are implicit and usually
and a score of 0 to the remaining items, for each unexpected effects of theoretical formulations.
09-House.qxd 3/19/2004 4:12 PM Page 164
The concept of “outcroppings” can be illustrated Hypothesis 1: The societal value of performance
with the following example: orientation is associated with norms that encour-
age a preference to choose one’s work or job
Complexity Theory, as an (unexpected) outcropping
of Chaos Theory in physics, has shown many Assertiveness Values
levels of proof that suggest that “reverse entropy” and Political Involvement
(or “Centropy” as J. J. Hurtak calls it) also exists.
Centropy would be the natural tendency for The Assertiveness value dimension is
simple things to become more complex, hence the associated with the following preferences, among
name “complexity theory.” (Wilcock, 1999) others: strong expression, articulation, and com-
munication of one’s thoughts, feelings, beliefs,
In this example, the original chaos theory and rights, both in political and social forums
gave rise to an additional insight, which was (see Chapter 15 by Den Hartog). An outcropping
not expected by physicists at the time of of the conceptualization of assertiveness is that
formulation of chaos theory. The outcropping assertiveness-oriented societies will have a high
of complexity theory serves to provide new degree of political activism among their members
insights and to further validate chaos theory. (Lange & Jakubowski, 1976). The GLOBE
In this section, we review several outcrop- construct of Assertiveness implies, but does not
pings associated with the GLOBE societal level specifically postulate, that in assertiveness-
value dimensions. We use selected items from the oriented cultures individuals throughout the
WVS that we believe have inferential linkages socioeconomic ladder are aware of the value of
to the GLOBE value dimensions as a source of participation in politics. In addition, the norm of
illustrating outcroppings of the GLOBE value assertiveness enables and encourages individuals
dimensions. These linkages were not originally in these cultures to express themselves politi-
anticipated when the GLOBE value dimensions cally. Thus, political assertiveness is an outcrop-
were formulated. ping of the GLOBE construct of Assertiveness.
From the above comments, it is suggested that
greater political involvement is more frequently
The GLOBE Societal Value
found in assertive cultures. The WVS included
Dimensions of Culture several items regarding societal political involve-
Performance Orientation Values ment. Given the arguments regarding assertive-
and Choice in Work ness, we advance the following hypothesis:
The Performance Orientation value dimen- Hypothesis 2: Societal norms of assertiveness
sion reflects the extent to which a society is are associated with greater political activity on
reported to encourage and reward performance the part of societal members
excellence and improvement. This dimension is
described in more detail in Chapter 12 by
Future Orientation Values
Javidan. Performance Orientation values encour-
and Spiritual Orientation
age a preference for challenge and being in con-
trol of one’sm destiny. An outcropping of the The GLOBE Future Orientation value dimen-
Performance Orientation dimension, although sion is described and discussed in Chapter 13
not explicitly stated, implies a norm of prefer- by Ashkanasy, Gupta, Mayfield, and Trevor-
ence for having jobs in which people control Roberts. It is related to the concept of short-term
what they do and when they do it. Extending the versus long-term orientation. It is also associated
logic a bit further, it is likely that performance- with the distinction between materialistic ver-
oriented cultures place a high value on work sus spiritual orientation (Cervantes & Ramirez,
choice. The WVS included one item for freedom 1992). Spiritual orientation is an outcropping of
of choice to work and one item that refers to GLOBE’s conceptualization of future orienta-
motivation of people to work. Given the argu- tion. In future-oriented cultures, the material
ments regarding performance orientation, we and spiritual realms are more integrated (Cross,
advance the following hypothesis: 2001). This integration is also an outcropping
09-House.qxd 3/19/2004 4:12 PM Page 165
of the GLOBE definition of future orientation are likely to be based on group rather than
because it was not expected at the time the individual performance. Personal independence
Future Orientation dimension was formulated. has low priority in institutionally oriented
From the above comments, it is clear that spir- collective societies (Ryff, 1989). The notion of
itual orientation is expected to be more frequently autonomous individuals, living free of society
found in cultures that score high on future orien- while living in that society, is contrary to the
tation. Thus we advance the following hypothesis: norms of societies that embrace institutional
collectivism (Shweder & Bourne, 1984).
Hypothesis 3: Societies that score high on the Societies characterized by lower institutional
value of future orientation will be characterized by collectivism tend to embrace a preoccupation
strong spiritual orientation among their members with self-reliance and independent personality
(Bellah, Madsen, Sullivan, Swindler, & Tipton,
Humane Orientation Values 1985). A common guiding principle in Asian
and Public Morality societies that score high on institutional collec-
tivism is that “The nail that sticks out gets
The Humane Orientation value construct is
pounded down”; in the Netherlands, a common
discussed in Chapter 18 by Kabasakal and
idiom is, “The head that rises above the ditch
Bodur. The norms of societies valuing humane
gets shot off.”
orientation are concerned with improvement of
From the above comments, we conclude
the human condition. One of the distinguishing
that a high degree of respect for legitimated
norms of humane-oriented societies is public
societal institutions is more frequently found
morality (Kurtz, 2001). Public morality is
in institutionally collective cultures. The WVS
expressed through laws and norms that empha-
included one item regarding a high priority on
size and reinforce moral behavior.
respect for legitimated societal institutions.
Thus we would expect a sense of public
On the basis of the above discussion regard-
morality to be stronger and more frequently
ing institutional collectivism, we advance the
found in humane-oriented cultures. We identi-
following hypothesis:
fied three items in the WVS concerning public
morality. The above discussion suggests the Hypothesis 5. Societies reported to value insti-
following hypothesis: tutional collectivism have members who place a
high priority on societal respect
Hypothesis 4: Societies that are reported to
value humane orientation will be characterized In-Group Collectivism Values
by norms that encourage and reinforce a sense and Pride in Family and Nation
of public morality
The In-Group Collectivism value dimension
encompasses how individuals relate to an in-
Institutional Collectivism Values
group as an autonomous unit and how individu-
and Societal Respect
als attend to responsibilities concerning their
The Institutional Collectivism value con- in-group (see Chapter 16). In-group collectivism
struct reflects inducements and rewards for norms reflect both pride in membership of group
collective behavior and norms, rather than members and general affective identification
incentives and rewards for the enactment of toward family, group, community, and nation
individual freedom and autonomy (see Chapter (Triandis, Bontempto, Villareal, Asai, & Lucca,
16 by Gelfand, Bhawuk, Nishii, and Bechtold). 1988).
Institutional collectivism emphasizes shared In strong in-group collectivistic societies,
objectives, interchangeable interests, and respect there is an emphasis on collaboration, cohesive-
for socially legitimated institutions (Chatman, ness, and harmony. Responsibility and identifi-
Polzer, Barsade, & Neale, 1998). cation with the group begins with the immediate
In organizations, institutional collectivism group, and then gradually extends externally.
likely takes the form of strong team orientation Put differently, in-group collectivism represents
and development. To the extent possible, tasks a strong sense of group identity and may extend
09-House.qxd 3/19/2004 4:12 PM Page 166
communities” could have been theoretically sample of countries common in both studies.
related to our definition of political involvement Table 9.7 provides the multitrait–multimethod
(a correlate of Assertiveness) or theoretically matrix. The correlations between constructs
related to our definition of sociotechnical gover- measured by a single method are termed as the
nance (a correlate of Uncertainty Avoidance). heterotrait–monomethod triangles. Of the two
Such items were excluded if the distinctiveness triangles, the first represents the selected WVS
of an item could not be determined. measures and the second the GLOBE variables.
Close attention was given to interpreting In addition, the heteromethod block is composed
the GLOBE societal value constructs for con- of the validity diagonal (bold values in the table),
tent. For instance, the scale of Sociotechnical and two heterotrait–heteromethod triangles.
Orientation used to validate the construct of The matrix can be grouped into three parti-
Uncertainty Avoidance values was assessed tions: diagonals, triangles, and blocks. There are
using only items that suggested social and tech- two heterotrait–monomethod triangles, which
nical approaches for managing uncertainty. The give correlations among measures that share the
final selection of the items was based on mea- same method of measurement, one for the ques-
sures with convincing face validity as corrobo- tionnaire-based method, and the other for the
rated by the GLOBE principal investigator. The unobtrusive measurement method (Campbell &
scale score for each dimension is composed of Fiske, 1959). The diagonals of these triangles
scores from the two raters. As a result of this contain the estimates of the reliability of each
item identification process, we obtained a con- measure in the matrix and are termed reliability
servative interrater agreement ranging from diagonals (monotrait–monomethod). In addition,
0.75 to 1.00. Of the nine GLOBE value valida- the matrix also contains a square heteromethod
tion scales plus the “nonapplicable” category, block, which gives correlations between con-
we had an average interrater agreement of structs measured by two different methods. This
0.82. Table 9.6 provides the items for the nine square block can be partitioned into a validity
outcropping measures scales. We assessed the diagonal (bold values in the table), and two
unidimensionality of these scales by conducting heterotrait–heteromethod triangles (Campbell
a principal components factor analysis. The & Fiske, 1959). The coefficients in the validity
factor loadings for these scales are shown in diagonal represent correlations between the mea-
Table 9.6. The Cronbach alpha reliability sures of the same trait using different methods
coefficient of all the scales exceeds the standard (i.e., the GLOBE questionnaire method and out-
of 0.70. cropping measurement method). Because the two
At first glance, several of the items selected methods are intended to measure the same con-
for validation of the GLOBE values scales may cept or trait, the coefficients in the validity diag-
not appear to directly affect the dimension of onal are expected to be strongly correlated, and
interest. However, for all items we developed are termed as monotrait–heteromethod correla-
either direct or indirect inference linkages, tions. Further, because the coefficients in the two
which may have been explicit or implicit, thus heterotrait–heteromethod triangles share neither
making these items relevant to the dimensions method nor trait, those coefficients are expected
through the rationale described earlier. to be very low (Campbell & Fiske, 1959).
The coefficients in the validity diagonal
represent correlation between each GLOBE
Findings
variable and the corresponding outcropping
The test of construct validity can be done using measures. Convergent validity exists when the
the multitrait–multimethod matrix (Campbell & coefficients in the validity diagonal are high and
Fiske, 1959). The nine cultural constructs repre- statistically significant. The correlation between
sent different traits. The two methods of measure- the hypothesized pairs ranges from 0.38 to 0.88,
ment, GLOBE variables and items from the and are all statistically significant at p < 0.05 for
WVS, represent different sources. The rater each of the nine validity coefficients.
scores from the WVS were correlated with the Discriminant validity is achieved if several
GLOBE societal level values scores for the criteria are met. First, the coefficients in the
09-House.qxd 3/19/2004 4:12 PM Page 169
Table 9.6 Operational Scales for the Outcropping Measures to Compare With GLOBE Values Scales
Factor
1. Choice in Work (alpha = 0.76) Loadings
V66. Feel much freedom of choice and control over the way life turns out −.81
V124. Our present society must be valiantly defended against all subversive forces −.82
V172. In this country, people in need are poor because of laziness and lack of will power −.84
V192. Claiming government benefits to which one is not entitled is generally justifiable −.85
V193. Avoiding a fare on public transport is generally justifiable −.90
V194. Cheating on taxes if one has a chance is generally justifiable −.87
V78. In general, it is not important to have a job that is respected by people −1.00
(Continued)
09-House.qxd 3/19/2004 4:12 PM Page 170
Factor
7a. Gender Parity (alpha = 0.92) (reverse scored) Loadings
V61. When jobs are scarce, men should have more rights to a job than women −.92
V101. On the whole, men make better political leaders than women do −.95
V103. A university education is more important for a boy than for a girl −.92
V104105a. The goal of a high level of economic growth should be given top priority
over the next 10 years in this nation .81
V108109b. Progress toward a less impersonal and more humane society should be given
top priority in this nation −.86
V125. We need larger income differences as incentives for individual effort .63
V128. Competition is harmful and brings out the worst in people, as opposed to
stimulating people to work hard and develop new ideas −.72
V63. Even when jobs are scarce, employers should not give more priority to home
nationals over immigrants −.67
V73/74-1. A good income so that there are no worries about money is an important
factor for people looking for a job .72
V113. Against a change in the way of life to more emphasis on the development of −.83
technology
V126. Government ownership of business and industry should be increased, as opposed
to private ownership .70
V127. People should take more responsibility for providing for themselves rather than
having the government provide for everyone −.73
V131. One should not be cautious about making major changes because you never
achieve much in life unless you act boldly −.72
heterotrait–heteromethod triangles are not as the measures exceeds the coefficients in the
high as those in the validity diagonal. In addition, heterotrait–monomethod triangles (Campbell &
the coefficients in the heterotrait–monomethod Fiske, 1959).
triangles are also less than the coefficients in Using heterotrait-heteromethod triangles, one
the validity diagonal. Last, the reliability of may compute the average absolute correlation of
Table 9.7 The Multitrait-Multimethod Matrix of Societal Value Measures
09-House.qxd
A1 B1 C1 D1 E1 F1 G1 H1 I1 (J) A2 B2 C2 D2 E2 F2 G2 H2 I2
Outcropping Measures
3/19/2004
A1 * (.76) .32*
B1. .37** (.90) .08
4:12 PM
GLOBE
A2. .38* −.01 .38 .59** .21 .41* .27 −.32 −.03 .36* (.90)
B2 .19 .39* −.15 .00 −.02 .07 −.22 .03 .23 −.20 −.06 (.53)
C2 .13 −.33 .59** .49 .34* .56** −.19 −.19 .39* −.01 .69** .03 (.76)
D2 .25 −.07 .24 .70** −.04 .26 .24 −.38* −.12 .38* .61** −.12 .38** (.70)
E2 .08 −.07 .42** .16 .51** .30 .00 −.22 .23 .15 .43** −.22 .50** .07 (.77)
F2 .26 −.25 .44** .35* .17 .60** .18 −.21 .14 .23 .70 −.02 .66** .27* .36** (.66)
G2 .25 −.12 .23 .26 −.07 .25 .67** −.35 −.57** .51** .28* −.29* −.16 .28* .00 .21 (.88)
H2 −.19 .18 −.22 −.48** −.16 −.18 −.33* .60** .09 −.27 −.61** .29* −.36** −.62** −.37** −.42** −.51** (.74)
I2 .00 −.19 .34* .08 .43** .44** −.58** .27 .88** −.29 .25 .19 .64** .02 .45** .35** −.49** .04 (.85)
NOTE: Letters A, B, C, D, E, F, G, H, and I refer to statuses on the dimensions of culture. Subscripts 1 and 2 to refer to the 2 methods used in this study. Method 1 refers to the Outcropping
Measure items and Method 2 refers to the GLOBE items. Validity coefficients are the diagonal set of boldface numbers; reliability coefficients are the numbers in parentheses along principal
diagonal. For criteria variable, N = 73; for GLOBE, N = 62; for multitrait-multimethod, N = 39. Gender Parity Scale used as Validation Measure for Gender Egalitarianism.
171
*A = Performance Orientation. B = Assertiveness. C = Future Orientation. D = Humane Orientation. E = Institutional Collectivism. F = In-Group Collectivism.
G = Gender Egalitarianism. H = Power Distance. I = Uncertainty Avoidance. (J) = Unity in Diversity. * p < 0.05 ** p < 0.01
09-House.qxd 3/19/2004 4:12 PM Page 172
each outcropping measure (W) with the nine help minimize the costs of unexpected change.
GLOBE variables (G). This average is labeled as Sociotechnical orientation is one way in which
the discriminant coefficient (W,G). Similarly, members of a society can manage uncertainty
we compute the average absolute correlation of collectively. Several other approaches for man-
each GLOBE variable with the nine outcropping aging uncertainty may exist, and the signifi-
measures. This average is labeled as the discrim- cance of sociotechnical orientation may itself
inant coefficient (G,W). Table 9.8 shows that vary for the sampled societies as well as for the
for each of the nine variables, both types of nonsampled societies. Further, the relevance of
discriminant coefficients are smaller than the sociotechnical orientation for managing uncer-
validity diagonal coefficients. The overall aver- tainty might grow or fall over the years across
age discriminant coefficient is 0.24 for both any given set of societies. Therefore, it would be
(G,W) and (W,G), which is significantly less important to confirm that the outcropping mea-
than the average convergent coefficient validity sures are indeed relevant for the societies in the
(i.e., diagonal coefficient) of 0.59. target sample before using them independently
Similarly, using heterotrait–monomethod or as tests of convergent validity in other cross-
triangles, the discriminant coefficient (G) was cultural studies. Because we did not have evi-
computed as the average correlation of each dence of the relevance of the outcropping
GLOBE measure with the other eight GLOBE measures for each society studied, we may have
measures; and a discriminant coefficient (W) made a conservative error. Despite this possibil-
as the average correlation of each outcropping ity, we were able to demonstrate both conver-
measure with the other eight outcropping mea- gent and discriminant validity of the GLOBE
sures. The overall average discriminant coeffi- scales using the outcropping measures.
cient (G) is 0.33, and overall average criteria In addition, it would be fruitful to examine
discriminant coefficient (W) is 0.26, both of data from the ongoing fourth wave of the World
which are also significantly less than the aver- Values Survey, which expands the sample of
age convergent coefficient shown in the first societies and also provides more updated infor-
column of Table 9.8. mation, and reaffirm the findings on the basis of
Last, the reliability of both GLOBE measures the second and third waves of the survey. Data
and outcropping measures exceeds the discrimi- from other independent cross-cultural studies
nant coefficient computed from heterotrait– may also be used for further validation, and for
monomethod triangles. Thus, we conclude that refining the domain of meaning that should be
the GLOBE societal value constructs have con- empirically associated with the GLOBE societal
vergent as well as discriminant validity. value constructs. Last, because the coders were
Using the construct validation procedure, able to identify only a single item from the WVS
we established the convergent as well as measures for the Institutional Collectivism con-
discriminant validity of the GLOBE societal struct, it may be advisable to identify an alterna-
value constructs. In addition, we showed that tive theme, or use an alternative data source.
the outcropping measures indeed capture the Despite these conservative limitations, we were
domain of meaning theoretically intended by able to demonstrate convergent and discriminant
the GLOBE societal value constructs. validity of the GLOBE societal scales.
Our findings must be viewed in the light of
potential limitations of the outcropping mea-
sures. These measures capture a part of the CONCLUSIONS
overall domain of meaning intended by the
respective value constructs. The potential In conclusion, we found that GLOBE measures
domain of meaning of each value construct is of cultural practices and values can be validated
quite broad-based. For instance, the value of by the use of independent measures from other
uncertainty avoidance would be related to a sources. The aggregated reports of middle-level
range of indicators associated with the need for managers concerning societal values and prac-
security, as well as other indicators that might tices correlated, as predicted, with a range of
help resolve the uncertainty about unknown, or theoretically predicted national level variables.
09-House.qxd
1 2a 2b 3 4 5 6
GLOBE Coefficients
3/19/2004
Orientation
Assertiveness 0.39 0.11 0.15 0.17 0.15 0.90 0.53
Future 0.59 0.33 0.30 0.22 0.43 0.95 0.76
Orientation
Page 173
NOTES: Summary Conditions for Construct Validity: 1 > 2(a), 2(b), 3, & 4; 5 > 3; 6 > 4 d
W (Discriminant coefficient) = Average absolute correlation of an outcropping measure
a with other eight outcropping measures.
G,W (Convergent coefficient) = Correlation of the GLOBE measure with the outcropping
measure. e
G (Discriminant coefficient) = Average absolute correlation of a GLOBE measure with
b other eight GLOBE measures.
G,W (Discriminant coefficient) = Average absolute correlation of a GLOBE measure with
the nine outcropping measures. f
W (Reliability) = Reliability of an unobtrusive measure.
c g
W,G (Discriminant coefficient) = Average absolute correlation of an outcropping measure G (Reliability) = Reliability of a GLOBE measure.
173
with the nine GLOBE measures.
09-House.qxd 3/19/2004 4:12 PM Page 174
Thus, the triangulation of traditional survey, descriptive background of the major historical,
questionnaire validation, and unobtrusive meth- political, and social events of the country; a depiction of
ods provide a valid and balanced perspective of the topography and climate; some general characteris-
societal culture unavailable by each perspective tics of the people; a description of population demogra-
alone. In addition, these alternative data sources phy; a review of the languages spoken, especially
helped to uncover cultural values and practices of noting the prevalence of English spoken; a statistical
societies not included in the GLOBE or the WVS percentage itemization of types of religion; an overview
questionnaire responses. Further, the unobtrusive of general attitudes; and considerations of appearance.
and outcropping measures can be used to identify Other sections incorporated in these four-page reports
and outcropping changes in cultural values and are: examination of common customs and courtesies;
practices, thereby helping the creation of indices description of traditional greetings and gestures; infor-
of cultural change. These indices could be used to mation regarding visiting, eating, and lifestyle habits;
make appropriate adjustments in the GLOBE and review of beliefs about family, dating, and mar-
cultural values and practices scores for future riage. In addition, these texts comprise information on
use. Such adjustments would allow more reliable analysis of social and economic issues; diet and recre-
assessment of the contemporary cultural values ation; major holidays; instructive details regarding
and practices, and help develop effective strate- commerce, society, government, the economy, educa-
gic and policy decisions rooted in cultural trends. tion, transportation, and communication, as well as a
discussion of health-related risks to citizens and visitors
in the country. Last, each Culturgram report contains
ENDNOTES information for travelers to each country, as well as
useful addresses to obtain further details.
1. Construct validity may be assessed through 5. The following is an illustration of our process
testing the convergent hypothesis and the discriminant of developing a rationale for constructs. The con-
hypothesis (Rossi, Wright, & Anderson, 1983). The struct of uncertainty avoidance is related to reliabil-
convergent hypothesis is that items within the domain ity, durability, and breadth of concepts. The medical
of meaning, although derived from different methods, sector is one of the critical domains predicated on
correlate together because they all reflect the same reliable and stabilizing domestic initiatives. This is
underlying construct or ‘true’ score. The discriminant so because many diseases tend to be climate and
hypothesis is that items from one domain will not society-specific and are subject to mutation over
correlate with items from another domain (Rossi, time. Thus, durable and broad-based home research is
Wright, & Anderson, 1983, pp. 100–101). Each essential to developing preventive and curative solu-
construct is intended to measure a common domain of tions. On the basis of this inference clarification
meaning. The degree to which researchers have repre- process, the item “medical facilities are effective in
sentatively sampled from that domain of meaning keeping the society essentially healthy and disease-
connotes content validity of the scale. free” was formed to represent uncertainty avoidance.
2. We selected Culturgrams reports (1999) as After this process of linking the GLOBE societal
our data text source in our content analyses because practice construct to the inferences, the project direc-
they include descriptions of daily customs and life tor reviewed the final constructs. It was determined
in societies. These reports provide comparable and that six of the nine dimension constructs demonstrated
consistent, four page descriptions of cultures of more than adequate face validity, but three of the nine
170 societies around the world. dimension constructs lacked convincing face validity.
3. We discuss these under the following topic At this point, the research team secured feedback from
headings: (a) applying the framework for content other experts to reformulate the construct-inference
analysis, (b) searching for suitable data, (c) searching process for the three problematic GLOBE dimensions
for contextual knowledge, (d) developing a strategy of humane orientation, assertiveness, and societal col-
and defining measurements, (e) coding for unobtru- lectivism. Numerous meetings were held before the
sive indicators, and (f) assessing quality and accuracy. team agreed on the construct-inference definitions for
4. Each Culturgram follows a standard format these dimensions of culture.
with sections of information that include a map indicat- 6. Concurrent with developing strategies for
ing the continental location and principal cities; a content analysis, it is important to formulate the units
09-House.qxd 3/19/2004 4:12 PM Page 175
of analysis, which is sometimes called unitizing categories. Two basic decisions are suggested when
(Krippendorff, 1980). In our content analysis, the developing category definitions: whether the categories
sampling unit was the four-page Culturgram report for should be (a) mutually exclusive and (b) broad or
each country that corresponded with the GLOBE pro- narrow (Weber, 1990). Mutually exclusive cate-
ject countries. The context units define the segment of gories were required in this analysis to ensure that
the text to be examined in order to characterize a our variables were not confounded. Recording
recording unit. These units set physical limits on the units needed to be classified in a single category.
contextual information that may comprise the recod- Simultaneous classification in two or more cate-
ing units. Context units do not need to be independent, gories would have resulted in violation of fundamen-
and may overlap and contain many recording units. In tal statistical assumptions and would have rendered
our study, the context units were sections of the text unreliable results (Weber, 1990). This decision
into which the Culturgram were divided. These sec- process is illustrated below.
tions of the texts described aspects of the culture such When coding for Power Distance the item
as history, general attitudes, family, and economy. For “Economic growth tends to go hand-in-hand with the
example, the Humane Orientation culture dimension unemployment, poverty, and stratification of society
evaluates such issues as concern for others, friendli- by income,” we identified referent units relating to
ness, sensitivity toward others, and generosity. the effects of unequal distribution of income and
By contrast, recording units are seldom defined private consumption expenditure in the society.
in terms of physical boundaries. The division Evidence of unemployment and poverty were found
between recording units is reached through a in the text regarding the economy, and occasionally
descriptive effort (Krippendorff, 1980). In this these indications were embedded in the discussion,
content analysis, the recording units were the theo- which required a thoughtful reading for implicit
retically driven ideas relating to the nine cultural information. To assess the stratification of society by
dimensions. To illustrate, the Denmark Culturgram income, we looked for references to gaps between
section on general attitudes included the statements rich and poor that were specific in the text and gener-
“Danes are known for their tolerance of other people ally uncomplicated to code. A second decision in the
and diverse points of view. They admire individuals development of category definitions involved how
with a friendly attitude, a sense of humor, intelli- broad or narrow the categories would be. Decisions
gence, sociability, personal stamina, integrity and an were made to code some concepts very specifically
open mind. . . . A love for understatement, rather and some more broadly. For example, the In-Group
than exaggeration, prevails” (p. 78). Thus, recording Collectivism item “Most people lack access to the
units in this section could be coded higher on opportunities for personal advancement” reflects a
Humane Orientation, given the reference to friendli- social system in which intellectual autonomy is not
ness and tolerance, and lower on Assertiveness facilitated. The Human Development Index value for
Orientation, given the reference to nonassertive each country was used as a narrowly defined category
behavior. After designating these units, we then to assess the collectivism category. In contrast, the
defined and delineated the actual coding of the concept of Uncertainty Avoidance was defined
units. Referential units were used in this analysis. broadly. This concept is related to the use of infor-
Referential units indicate how a unit is represented, mation technologies, supporting ample feedback, and
defining the ideas to which an expression refers. building capabilities for information technology
The referential unit denotes a similar expression in usage. It should be noted that all initial decisions
different ways, often defined by specific notions, regarding strategy development and measurement
events, persons, acts, or objects, (Krippendorff, defining issues were dependent on the information
1980). For example, the Performance Orientation available in Culturgrams as identified at the time
item, “The society maintains a diversity of reli- of exploratory analysis of non-GLOBE sample
gions” was coded in terms of the dominant religion, societies.
as well as openness to other religions such as in 8. We chose not to use their weight system
terms of state laws (secular or not), and recent because of the subjectivity involved in the choice of
growth of new religious ideologies in the society. parameters used for weighting.
7. It was necessary in this content analysis to 9. The data for the early 2000s cycle were not
take note of issues pertaining to the definition of publicly available as we conducted this research.
09-House.qxd 3/19/2004 4:12 PM Page 176
Ryff, C. D. (1989). Happiness is everything or is it? Triandis, H. C., Bontempto, R., Villareal, M. J.,
Explorations on the meaning of psychological Asai, M., & Lucca, N. (1988). Individualism and
well-being. Journal of Personality and Social collectivism: Cross-cultural perspectives on
Psychology, 57, 1069–1081. self-in-group relationships. Journal of Perso-
Sechrest, L., & Belew, J. (1983). Nonreactive nality and Social Psychology, 54, 323–338.
measures of social attitudes. Applied Social United Nations Education, Science and Culture
Psychology Annual, 4, 23–63. Organization. (1997). Statistical yearbook.
Shapiro, G., & Markoff, J. (1997). A matter of defini- New York: Author.
tion. In C. W. Roberts (Ed.), Text analysis for the Webb, E., Campbell, D. T., Schwartz, R. D., &
social sciences: Methods for drawing statistical Sechrest, L. (1966). Unobrusive measures: Non-
inferences from texts and transcripts. Mahwah, reactive research in the social sciences.
NJ: Lawrence Erlbaum. Chicago: Rand McNally.
Shweder, R. A., & Bourne, E. J. (1984). Does the Webb, E., Campbell, D. T., Schwartz, R. D., &
concept of the person vary cross-culturally? In Sechrest, L. (2000). Unobtrusive measures.
R. A. Shweder & R. A. LeVine (Eds.), Culture Thousand Oaks, CA: Sage
theory: Essays on mind, self, and emotion Weber, R. P. (1990). Basic content analysis (2nd ed.).
(pp. 158–199). New York: Cambridge University Newbury Park, CA: Sage.
Press. Wilcock, D. (1999, March 8). Convergence.
Stone, P. J. (1997). Thematic text analysis: New agen- Retrieved from http://www.dprins.demon.nl/
das for analyzing text content. In C. W. Roberts convergence/9913.html
(Ed.), Text analysis for the social sciences: World Bank. (1997). World development indicators
Methods for drawing statistical inferences from 1997 CD-ROM. Washington, DC: United
texts and transcripts (pp. 171–189). Mahwah, Nations.
NJ: Lawrence Erlbaum. Yan, W. F., & Gaier, E. L. (1994). Causal attribu-
Sullivan, M. J., III. (1991). Measuring global values: tions for college success and failure: An Asia-
The ranking of 162 countries. New York: American comparison. Journal of Cross-Cultural
Greenwood Press. Psychology, 25, 146–158.
10-House.qxd 3/19/2004 5:30 PM Page 178
10
REGIONAL AND
CLIMATE CLUSTERING
OF SOCIETAL CULTURES
VIPIN GUPTA
PAUL J. HANGES
n this chapter, we establish two types of similarities and differences among these
AUTHORS’ NOTE: Peter Dorfman put in long hours on editing and commenting on the various versions of this
chapter, but politely refused the offer to be a coauthor. We also thank Felix Brodbeck and Michele Gelfand for
their very helpful comments on an earlier draft of this chapter, and Bhakti Gupta for her assistance in preparing
figures for this chapter.
178
10-House.qxd 3/19/2004 5:30 PM Page 179
present the CLTs for each societal cluster and and interventions. Clusters may be used to guide
show how specific leadership prototypes are the sampling strategy for cross-cultural research
empirically linked to GLOBE cultural dimen- to ensure that an adequate sampling of cultural
sions. That is, we consider these cultural clus- variability is included in the samples. Researchers
ters to be helpful for summarizing the culturally can also test the generalizability of empirical
endorsed leadership prototypes that prevail in findings obtained in one culture to other cul-
the societies associated with specific cultural tures. Thus, clusters provide us with important
clusters. Third, the GLOBE data provides an information regarding societal variation and are
opportunity to empirically validate the appropri- a coherent and convenient way to summarize
ateness of societal clusters, particularly cluster- intercultural similarity as well as intercultural
ing procedures that are based on cultural values differences. Both the practical and theoretical
and practices, in contrast to clusters based on implications for clustering are presented in
concrete variables such as language, geography, more detail later in this chapter.
and modernity. Grouping countries (cultures The information in this chapter follows the
or societies) on the basis of constructs such same order we used to construct the GLOBE
as work goals, values, or beliefs has proved clusters. First, we reviewed the extant literature
difficult to validate. The vast database of the to determine the criteria other scholars have
GLOBE project provides a unique opportunity used to cluster societies. To better understand
to not only develop clusters, but also to empi- the dominating cultural forces responsible for
rically assess the adequacy of our clustering societal development, we conducted historical
scheme. and cultural analyses of their influence on dif-
The fourth and fifth reasons for clustering ferent regions around the world. We integrated
societies concern the practical and theoretical the extant literature on previous clustering
benefits that should result from an adequate attempts with this in-depth review of historical
clustering of societies. In terms of practical and cultural forces to establish our 10 regional
benefits, clusters provide a useful framework clusters. The thesis that these societal clusters
for managing the complexities of multicultural reflect greater differences in cultural practices
operations. Multinational corporations have and values across societies from different clus-
the difficult job of operating in nations that ters than societies within the same cluster was
have different sets of cultural expectations, tested using discriminant analysis. Through this
beliefs, and values. Likewise, clusters may analysis we confirmed the viability of our regional
provide useful information for working with clusters.
diverse nationalities or cultures within a specific
country. Practices, policies, and procedures
Forces Influencing Regional Clustering
that work quite effectively in one culture may
dramatically fail or produce counterproductive Scholars have used three major forces to
behavior in another culture. Information group countries into similar clusters: (a) geo-
relevant to clusters should assist the selection graphic proximity (Furnham, Kirkcaldy, &
and cultural training of managers who work Lynn, 1994), (b) mass migrations and ethnic
in global environments. In addition, cluster social capital (Portes & Zhou, 1994), and (c)
information can help managers understand the religious and linguistic commonality (Cattell,
viability of policies, practices, technologies, and 1950). Of these three forces, geography has long
human resources as they are applied across cul- been identified as having a major influence on
tural boundaries. Finally, societal clusters may culture. For example, Hofstede (1980) reported
also help regional managers determine limits geographical latitude to be the single largest fac-
and boundaries for moving human resources tor explaining intersocietal variation in cultural
and products across societies. dimensions such as power distance. Geography
Cluster-based information can assist in can influence culture because the physical
theory development. Judicious sampling within topography of a region can limit the interactions
and across societal clusters can test potential among people, the nature of the physical climate
boundary conditions for management theories in a region can influence the goals of people
10-House.qxd 3/19/2004 5:30 PM Page 180
living in that region, and the economic prosperity Indeed, researchers espousing this view believe
of a region can be affected by its physical geog- that societal clusters really reflect differing
raphy. Mass migrations often extend the social levels of economic development and once these
capital—the cohesive interrelationships and differences have been taken into account, these
networks developed to facilitate social action— researchers believe that societies will not
from one society to places far beyond its original meaningfully cluster.
geographical boundaries (Portes & Zhou, 1994). Nevertheless, contrasting evidence suggests
For instance, mass migrations from the United cultures tend to be quite different even among
Kingdom during the second half of the past societies at similar levels of economic develop-
millennium have been a critical factor in the ment. Thus, Trepo (1973) reported how the
creation of an ethnic Saxon culture across geo- transfer of management by objectives (MBO)
graphically distant nations such as Australia, from the United States to France was unsuc-
South Africa, and the United States (Ronen & cessful. In particular, it was difficult to reconcile
Shenkar, 1985). Within the same geographic the significance of hierarchy in French culture
region and level of economic development, with joint participation of supervisor and sub-
religion and language distinguish societal clus- ordinate to reach decisions regarding the sub-
ters. For example, one can differentiate between ordinate’s performance. Trompenaars (1993)
religious–linguistic societal clusters of Germanic and Hofstede (1980) provide many additional
(Germans) and Latin Europe (Romans). The examples that support the differentiation of cul-
shared religious and linguistic institutions may tures even among countries of similar economic
generate similar paradigms of metaphysical development.
philosophies of science that guide beliefs, In summary, several factors contribute to
behaviors, and values in each society (Berry, shared cultures among people of different socie-
Poortinga, Segall, & Dasen, 1992; Smith & ties. Economic forces may moderate some of
Bond, 1994). these factors, but it may be difficult to change
Although geographical, ethnic, and religious– fundamental characteristics of beliefs and
linguistic forces contribute to the formation of behaviors of people within different societies.
societal clusters, economic forces may some- In such cases, one may find distinct cultures
times cut across these factors and therefore of societies at any level of economic develop-
reduce their influence among these clusters. The ment, and also cutting across different degrees
economic view holds that cultural differences of modernity. In addition to these factors, sev-
are minimal in economically developed soci- eral researchers have indicated that social and
eties, which tend to share common manage- psychological variables such as attitudes, values,
ment models to realize economic efficiency and work goals can be used to cluster societies
(Hyden, 1983). Furthermore, cultural patterns (e.g., Haire, Ghiselli, & Porter, 1966; Ronen &
in developing countries reflect idiosyncratic Shenkar, 1985). In the next section of this
and nonuniversalistic influences, but economic chapter, we review the previous empirical
development generates a global convergence in attempts to cluster societies.
the values and behaviors from developed
societies to developing ones (Dunlap, Harbison, Empirical Studies
Kerr, & Myers, 1975).
on the Clustering of Societies
To mitigate some of these influences in
emerging markets and facilitate a universal There has been approximately a half-century
global culture, several multilateral institutions of effort to identify clusters of societies using
such as the International Monetary Fund have the analysis of international-level data (Cattell,
sponsored management workshops that are 1950). One of the most well-known and most-
based on the belief that diffusion of Western referenced clustering of societies was proposed
management models would contribute to eco- by Ronen and Shenkar (1985). Their research
nomic development (Kerrigan & Luke, 1987). The is a good example on how work-related values
economic view thus implies that grouping of and attitudes can be used to group countries
societies into cultural clusters may be artifactual. together in a meaningful way. We will save the
10-House.qxd 3/19/2004 5:30 PM Page 181
discussion of this study until last because of technical, and service personnel in 25 nations,
its direct relevance to the GLOBE project. The and collected data on 14 work goals. The eco-
earliest impetus to clustering research can be nomic and religion–language patterns were still
traced to the pioneering works of Toynbee reflected in the resulting clusters; however, some
(1947) and Cattell (1950). Toynbee identified anomalies surfaced as Latin American nations
21 distinct living or extinct cultural patterns were mapped together with Sweden and Israel.
across civilizations. Of the 21 patterns that he Brazil, Japan, and India did not group in a mean-
identified, only five clusters are still surviving: ingful interpretable way with other societies.
Western, Orthodox Christian, Islamic, Hindu, In another study, Ronen and Kraut (1977)
and Far Eastern. Cattell (1950) analyzed about studied the importance of 15 work goals using a
80 variables to construct 12 factor dimensions sample of 4,000 technicians in 15 nations.
that measured various psychological, sociologi- Though the number of nations in the sample was
cal, demographic, and economic characteristics small, the clusters had strong religion–language
of the societies within his sample. His societal basis. Hofstede’s (1976) study, which involved
clusters included (a) Catholic Homeland, 315 middle-level managers representing 26
(b) Catholic Colonial (including Latin American nationalities on a set of 12 scales measuring per-
countries), (c) Eastern European, (d) Nordic, sonal and interpersonal values, indicated that the
(e) Islamic, (f) East Baltic, (g) Hamitic (including national groupings on the basis of values could be
Arab societies), and (h) Oriental (India and interpreted in terms of religion, language, and
China). In his study, the larger developed nations, geography. Hofstede (1980) replicated these
such as France, Germany, the United Kingdom, results using a survey of IBM managers from a
the United States, the Soviet Union, and Japan, much larger sample of nationalities and found
could not be clustered with any of these and that Japan was isolated from the Far Eastern
emerged as independent units in the analysis. In a cluster of nations, indicating some moderating
review paper, Woliver and Cattell (1981) noted role of economic factors.
that 12–20 factors are necessary for full cluster More recently, Furnham, Kirkcaldy, and Lynn
descriptions, and termed these as “syntality” fac- (1994) surveyed 12,000 students from 41
tors. In contrast to the individual culture dimen- countries from 5 continents (South America,
sions, which indicate how societies differ from North America, Europe, Africa, and Asia-
one another, syntality factors identify commonal- Pacific). Results of this study highlighted distinct
ities among clustering of societies. Cattell used Western and Eastern cultures. For example,
secondary data on the organized behaviors of the Schwartz (1999) reported the findings for grade-
group as a whole (e.g., the number of treaties a school teacher samples from 44 nations and uni-
group makes, the group’s gross national product, versity student samples from 40 nations, whereby
the frequency of a group’s involvement in war, nations were plotted along with the structure of
and the group’s plan for welfare) as variables that seven value types on a two-dimensional graph.
loaded on his syntality factors. For both samples, there emerged a distinct
Most more recent work on clustering has East Europe cluster, a West Europe cluster, an
included psychological variables such as emp- English-speaking cluster, a Far East cluster,
loyee attitudes and goals, or management prac- an Islamic cluster, and a Latin America cluster.
tices variables such as perceived roles and Separately, Inglehart and Baker (2000) plotted
management styles, rather than socioeconomic 65 societies on two-dimensions of cross-cultural
data to cluster countries. Using 3,641 respon- variations using data from the 1990–1991 and
dents in 14 countries, Haire, Ghiselli, and Porter 1995–1998 World Values Surveys. The first dimen-
(1966) surveyed Maslow need satisfaction, atti- sion was Traditional/Secular-Rational authority
tude toward democratic managerial practices, and and the second was Survival/Self-Expression. They
cognitive descriptions of the managerial role. identified a Protestant Europe cluster, an English-
The resulting societal clusters showed strong speaking cluster, a Latin America cluster, an African
economic patterns and religion–language subpat- cluster, a South Asia cluster, a Catholic Europe
terns within each economic grouping. Later, Sirota cluster, an Orthodox cluster within a broader
and Greenwood (1971) surveyed 13,000 sales, ex-Communist cluster, and a Confucian cluster.
10-House.qxd 3/19/2004 5:30 PM Page 182
Other studies have examined the similarities model, with its Germanic and Nordic variants,
and differences among European countries with was clearly differentiated from the Southern
respect to culture and leadership attributes. Smith, European model on the basis of Latin (Catholic)
Dugan, and Trompenaars (1996) analyzed data on traditions. In Northern Europe, the Germanic
personal values and behavioral intentions of model was distinguished by comanagement or
10,000 managers and employees from 43 nations. codetermination. Responsibilities were clearly
They concluded that a fundamental divide exists and precisely defined in a military way, and
between Eastern and Western Europe, and noted there was little scope for individuals to use
that the footprint of history that appears to leave personal discretion, as is frequently the case in
the sharpest imprint at present is not the legacy of Latin European countries. The Nordic model
the Roman Empire, but that of the Soviet Union. shares democratic and participatory spirit with
Western European countries tend to value the Germanic model, but is also highly egali-
achieved status: that is, social standing and pres- tarian. Each person seeks opinions from others
tige reflecting the ability of an individual to and must justify all proposals and accept coun-
acquire an established position in society as a terproposals. The Latin European model, in
result of individual accomplishments. In contrast, contrast, inclines people to tightly control infor-
Eastern European countries tend to value ascribed mation, secrets, and power, and also endorses
status: that is, social standing or prestige that is hierarchy.
the result of inheritance or hereditary factors. Perhaps the most referenced and enduring
Smith (1997) analyzed a subset of these data that research findings regarding the empirical clus-
excluded Central and East European countries. tering of societies in the organizational literature
The North European countries in the Anglo and was that proposed by Ronen and Shenkar (1985).
Nordic cluster and Germany valued participation, Their pioneering work used previously pub-
equality, and utilitarian involvement. The South lished data that was subsequently subjected to
European countries included in the Latin a statistical procedure known as “smallest space
European and the Near East cluster, along with analysis” (Guttman, 1968). The results of the
Austria valued reliance on supervisors, hierarchy, cluster analysis procedure were presented in the
and loyal involvement. Zander (1997) also found form of a figure that grouped countries together
different leadership style preferences between in terms of their similarity on work-related vari-
Northern and Southern European countries: ables. For instance, Latin American and Latin
Those in the north preferred coaching and those European countries were closely related to each
in the south preferred directing. other as were the Germanic and Nordic countries.
Brodbeck and colleagues (2000) recently ana- They also proposed Arabic, Near Eastern, Anglo,
lyzed the GLOBE database for leadership proto- and Far Eastern clusters. Brazil, Japan, India,
types of 22 European nations and found 6 basic and Israel were considered independents and
clusters and 2 meta-clusters: Anglo, Nordic, and remained separate from other identifiable clus-
Germanic countries in the first metacluster of ters. Their results provide a useful example and
Northern and Western Europe; and Latin structure for understanding the relationship
European, Arab (Near East), and Central and between these worker attitudes and values
Eastern European countries in the second meta- and country groupings. They paid particular
cluster of Southern and Eastern Europe. The for- attention to studies containing the kind of data
mer group favored interpersonal directness and used in the present GLOBE project (including
proximity far more than the latter. work-related values and attitudes of respondents
In a study of 15 countries representative within each society) and it is for this reason that
of Western Europe, Tixier (1994) compared dif- we have saved the discussion of the Ronen
ferent management and communication styles and Shenkar study until last. However, Ronen
across Europe and found several exceptions and Shenkar (1985) had to rely on prior studies
to the “usually accepted cultural clusters of that were rather limited in the number and
northern European, southern European, Latin, sample of societies. Further, these studies did
Anglo-Saxon, Nordic, and Germanic countries” not have a sufficient representation of societies
(Tixier, 1994, p. 8). Still, the Northern European from Eastern Europe, Africa, and Asia. Because
10-House.qxd 3/19/2004 5:30 PM Page 183
GLOBE provides data from a large number of Our overall goal was to adequately cluster 61
organizations in societies with significant societies participating in the GLOBE study. As
regional differentiation, we are in a position to indicated by the previous review of the “cluster-
replicate and extend Ronen and Shenkar’s ing literature,” there is no perfect or widely
(1985) clustering of societies. Further, we can accepted clustering of countries. This is proba-
empirically validate our proposed clustering bly desirable given the different purposes for
scheme as well as others previously proposed. undertaking this effort. Different societal clus-
In sum, the prior empirical studies suggest that ters may be equally valid given the uses for
religion–language, geography, and ethnicity, and which the process was developed. GLOBE
work-related values and attitudes are relevant researchers used the results of previous empiri-
factors in the clustering of societies. Economic cal studies, other factors such as common lan-
development may also be associated with signif- guage, geography, and religion, and, perhaps
icant changes in the society’s work culture, and most importantly, historical accounts when con-
push it apart from other societies that do not structing the final GLOBE clusters.
have a similar level of economic development (as Table 10.1 presents a list of countries
in the case of Latin America and Latin Europe). organized in societal clusters by our proposed
Still, as we should expect, societies that share schema. As a result of our analysis, we propose
similar religion–language, geography, and ethnic- that 61 GLOBE societies can be grouped into
ity tend to look alike in regards to fundamental 10 distinct clusters. For ease of discussion, these
characteristics. It should be noted, however, that ten clusters are described as belonging to the
intercultural similarities may hide significant following geographic regions of the world: Europe,
within-society heterogeneity. For instance, Ronen the Americas, Africa, and Asia. What follows is
and Shenkar (1985) note: a discussion of the rationale for proposing these
specific clusters.
Many countries are not homogenous: they consist of
various populations. They may differ according to Europe
language (French and Flemish in Belgium; French,
German, and Italian in Switzerland, etc.); according We propose that European societies should
to climate and differing proximity to other countries be grouped into Anglo, Latin Europe, Nordic
(e.g. Northern and Southern Italy); or according to Europe, Germanic Europe, and Eastern Europe
urban/rural and other differences. (p. 441) clusters. The rationale for this grouping is
explained below.
GLOBE researchers remained cognizant of
intraculture variability particularly with respect Anglo Cluster
to countries such as South Africa, where vastly
different societal cultures exist within the same Anglo Cluster includes England (and
country. We have identified our societal clusters societies dominated by the English), Australia,
by considering the history of the societies under South Africa (White sample), Canada, New
consideration as well as the religious, linguistic, Zealand, Ireland, and the United States of
and economic similarities. Further, the develop- America. This cluster is based on several fac-
ment of the GLOBE societal clusters were tors including ethnic and linguistic similarities,
informed by the results of the aforementioned and migration patterns originating centuries ago
empirical studies. from areas now identified as Northern Europe.
As Stenton (1971) noted, during the Roman
Empire several Germanic people had migrated to
Britain since at least the 3rd century A.D. Rome
Developing Regional
withdrew its army from Britain in A.D. 410. In the
Clusters of GLOBE Societies
ensuing struggle for power, Germanic Oisc
The process and results of the GLOBE gained kingship in the latter half of the 5th cen-
societal culture clustering for major geographi- tury, bequeathing the name of the Oiscingas on
cal regions in the world are presented below. the Kentish (British) royal household. During
10-House.qxd 3/19/2004 5:30 PM Page 184
this period, as supported by archeological Martin Luther (1483–1546) taught that Roman
evidence and recorded in the Chronicles of Anglo Catholicism teaches Salvation from Original Sin
Saxons (Swanton, 2000), there were significant through faith alone, and that Original Sin could
migrations—craftsmen, warriors, and others— not be overcome by one’s own actions or merits.
from Jutland (northern Denmark), Angeln (as Under Roman Catholicism, one could only have
the border region of Denmark and Germany was faith that God would grant salvation, and that such
called), and Saxony (Germany), in addition to salvation was possible only through sacraments
Frisians from the Netherlands and Franks from performed by a priest, and that true priest must
northern France and Central Germany, espe- take a vow of celibacy (Ozment, 1986). Luther
cially to the eastern half of Britain (Swanton, promoted use of vernacular instead of Latin, and
2000). Over time, the boundaries between the empowerment of the laity to have a direct rela-
Angles and Saxons, the two dominant groups of tionship with God. In 1529, the Germanic reforms
migrants, blurred to give rise to Anglo-Saxon were termed as Protestant, when Reformed dele-
(English) culture in Britain, which had percep- gates in Germany issued a Protestatio. The Latin
tible influence on the local Celtic culture, lan- Europe cluster, which rejected the German-
guage, and institutions (for further references, minded empowering spirit of the Protestant
see Williamson, 1999). Over the second millen- reforms, includes Italy, Portugal, Spain, France,
nium, Anglo culture diffused to Ireland, the Switzerland (French speaking), and Israel.
United States, Canada, South Africa, Australia, With the exception of Israel and French-
and New Zealand, as the English migrants pene- speaking Switzerland, this clustering of countries
trated these societies. We therefore propose that is consistent with the findings of Ronen and
the following GLOBE societies be grouped into Shenkar (1985). We conceptualized the Latin
the Anglo cluster: United Kingdom, United Europe cluster to be based on more than
States, Canada, South Africa (White sample), embracing the Catholic religion, as indicated by
Ireland, Australia, and New Zealand. This clus- Israel’s inclusion as part of this cluster. Although
tering of countries is consistent with the findings many Jews converted to Catholicism due to reli-
of Ronan and Shenkar (1985). gious repression by the church, others migrated
to Eastern Europe to escape from such repres-
sion. These migrants, who later founded Israel,
Latin Europe Cluster retained their social and business ties with the
Latin European region and are therefore
The Latin Europe cluster consists of included as part of the Latin Europe cluster.
the regions influenced by Roman culture
(Latourette, 1965, Chapter 8). As Latourette
(1965) emphasizes, Nordic Europe Cluster
The cluster of Nordic Europe is related to the
the Catholic Reformation had its mainsprings and historical concept of Scandinavia. According to
its most cogent expressions in what might be called the Random House Encyclopaedia, Scandinavia
Latin Europe—the countries in Western Europe is the “region of northern Europe consisting
which had been most extensively incorporated in of the kingdoms of Sweden, Norway and
the Roman Empire. To be more precise, Protes- Denmark; culturally and historically Finland
tantism became the prevailing religion among the and Iceland are often considered part of this
Germanic peoples who, with one exception, had not area” (quoted at http://www.faqs.org/faqs/
settled within the former confines of the Roman nordic-faq/part2_NORDEN/preamble.html).
Empire. The exception—Great Britain—was more Scandinavia was first a geographical term refer-
apparent than real, for here the Anglo-Saxon and ring to the Scandinavian peninsula—Sweden
then the Scandinavian invaders had never been under and Norway. This was the island of “Scandinauia”
Roman rule and they had wiped out all cultural that Roman historian Pliny the Elder identified
traces of the Roman occupation. (Retrieved from in 67 C.E. to be situated in the sea at the edge
http://www.religion-online.org/cgi-bin/relsearchd. of the world, north of Germania (Pliny, 1991).
dll/showchapter?chapter_id=579) Later, political unions, territorial expansion, and
10-House.qxd 3/19/2004 5:30 PM Page 185
cultural interactions stretched the peninsular the communist regime of Eastern Europe for
concept to include Denmark, Finland and about 40 years, traditional German values such
Iceland. In Nordic languages, the word Scandi- as orderliness, straight forwardness, honesty,
navien means the countries that were the ancient and loyalty remained an integral part of the East
land of the Norsemen, an ethnic concept that German society. The traditional cultural leaders,
includes Denmark, Norway, and Sweden (and including poets, novelists, philosophers, and
sometimes Iceland), but excludes Finland, religious leaders also continued to play a major
which forms a part of Scandinavia only on a role in public life and in social welfare. As such,
broader cultural basis (Einola & Turgeon, 2000). the cultural patterns remained largely stable,
The term “Nordic countries” itself origi- facilitating reintegration with the West later (see
nated from French “Pays Nordiques,” and ini- also Sowards, 1996, for additional explanations).
tially referred to the “northern” countries in We propose that the following GLOBE societies
general and could be applied even to nations be included in the Germanic Europe cluster:
such as Canada. Subsequently, the term Nordic the Netherlands, Austria, Switzerland, former
became established to refer exclusively to the West Germany, and former East Germany. This
five Scandinavian countries, with their model clustering is again consistent with the findings
of welfare state and common history, culture, of Ronen and Shenkar (1985), except for
and religion, and similar languages (Einola & Switzerland, which is an additional nation in
Turgeon, 2000). The Nordic language term our study.
Norden (Pohjola or Pohjoismaat in Finnish) also
refers to these five Scandinavian countries,
Eastern Europe Cluster
which since 1956 have cooperated in the Nordic
Council. The Nordic Council recognizes the dat- The most common explanation of the Eastern
ing of the formal cooperation among the nations Europe cluster is based on Soviet hegemony.
of the region to 1397, when Margrete, the queen The Soviet factor, though obviously important,
of Denmark and Norway, was crowned the does not reflect other relevant forces, such as
queen of Sweden (Einola & Turgeon, 2000). geography and pre-Soviet history. Sowards
Accordingly, we propose to group the following (1996) has recommended an alternative formu-
GLOBE societies into the Nordic Europe clus- lation of the Eastern European region:
ter: Finland, Sweden, and Denmark. This clus-
tering of societies is consistent with the findings Since 1945, it has been deceptively easy to define
of Ronen and Shenkar (1985). Eastern Europe in terms of Soviet Communist dom-
ination. By this method, Eastern Europe is synony-
mous with the “satellites” set up by Stalin after the
Germanic Europe Cluster
Second World War—it is the region on the far side
The Germanic Europe cluster consists of of the Iron Curtain. Now that the Iron Curtain has
societies that continue to use the German lan- disappeared, it is easier to see the flaws in this
guage. Language, however, is not the primary approach, but for students of the area it always
reason for their clustering. The distinctive offered problems. For example, Greece was fre-
Germanic culture can be traced to at least 5400 quently excised from the area, on the grounds
B.C. (Schake, 1998). Tacitus (A.D. 55–117), who that it was not a Communist country, and lumped
was Roman, published a pamphlet Germenia on into something called the “Mediterranean.” But
the people of Germany, and described these to assert that Greece and Spain share more in
people as possessing a love of freedom and their historical backgrounds than do Greece and
fighting. Even after the Germanic people were Romania, one has to forget a great deal of Greek
converted to Roman Catholicism, they continued history. (Retrieved from http://www.lib.msu.edu/
to practice traditional religious practices and sowards/balkan/lecture2.html)
beliefs. Subsequently, during the postmedieval
times, the Protestant form of Christianity McNeill (1964) also highlighted that the
evolved as a distinct Germanic interpretation of history of Eastern Europe has been shaped by
life. Though former East Germany was under an exceptional form of a fundamentally Asian
10-House.qxd 3/19/2004 5:30 PM Page 186
culture, that of the steppe nomad (found in the supported by the tradition of civil law from
vast tracts of land in southeastern Europe and Spain and Portugal. The civil law, in contrast to
Asia). The “great war captains” continuously the precedent-based common law in Anglo
created the “steppe empire” in the region to nations, is founded on general deductive princi-
pull together separate pastoral groups, and con- ples. As such, the rules are subject to constant
stantly imposed controls over the land, labor, reinterpretation depending on particular facts.
produce, and freedom of movement of the region’s In fact, there exists a Brazilian saying, “For
peasant families. The presence of mountains friends, everything; for strangers, nothing; and
and forests provided social stability and promoted for enemies, the law” (Rosenn, 1988, p. 143).
group cohesiveness. As such, the societies in the Finally, Latin American paternalism derives its
Balkans and Central Europe, Eastern Europe, roots from the Iberian monarchy, the Catholic
and Central Asia share significant similarities. Church, and the extended patriarchal family.
We propose that the following GLOBE socie- The family’s boundaries are guided by com-
ties be included in the Eastern Europe2 cluster: padrazgo (i.e., coparenting), in which a child’s
Hungary, Russia, Kazakhstan, Albania, Poland, godparents move beyond friendship to formal-
Greece, Slovenia, and Georgia. This cluster of ize a closer bond to the family in the baptismal
countries has not been adequately sampled in ceremony. There is a fairly low level of trust of
prior studies and so it is not surprising that this those who are not part of their family or close
cluster really has not emerged in the prior research. friends. We propose that the following GLOBE
Thus, this is a newly proposed clustering of societies be classified into the Latin America
societies. cluster: Costa Rica, Venezuela, Ecuador, Mexico,
El Salvador, Columbia, Guatemala, Bolivia,
Brazil, and Argentina. This cluster of countries
The Americas
corresponds to the prior research showing a
Whereas the United States and Canada have Latin American societal culture distinguishable
been included in the Anglo cluster discussed from that of the rest of the world. Therefore,
above, the other GLOBE societies in the Americas unlike Ronen and Shenkar (1985), we include
share a distinct Latin culture, as explained below. Brazil as part of this cluster, rather than as an
independent nation.
Latin America Cluster
Africa
Catholicism has a dominant influence on the
societies in the Americas, with the exception of We propose that African societies be classi-
the United States and Canada. In addition, these fied into two clusters: Middle East and Sub-
societies “share a common Roman law heritage, Saharan Africa.
a common Iberian colonial past, and present-
day patterns of social organization” (Rosenn,
Middle East Cluster
1988, p. 128). There also tends to be an empha-
sis on Spanish and Portuguese languages. The The most prominent basis of Middle East
culture is characterized by the values of person- culture lies in the Near Eastern civilization of
alism, particularism, and paternalism (Osland, North Africa and West Asia that has existed
De Franco, & Osland, 1999). First, personalism since at least pre-Sumerian times in the 4th and
refers to simpatica, or a sense of connection and 3rd millennia B.C. The Arab conquest of Africa’s
avoiding direct affronts to personal dignity northern rim began within years after the death
(Albert, 1996). At the workplace, supervisors of the Prophet Muhammad more than 1,300
are expected to attend family functions of the years ago. The North African region encom-
employees, and have a regard for what would passes the Great Sahara Desert and Nile River
happen to the family if a poorly performing Valley (Danowitz, Nassef, & Goodman, 1995),
employee were fired. Second, particularism refers and includes Mauritania, Western Sahara,
to the legitimacy of using personal connections Morocco, Algeria, Tunisia, Libya, and Egypt.
for one’s particular benefits. This concept is The culture of this Northern Rim came to be
10-House.qxd 3/19/2004 5:30 PM Page 187
shaped by Islamic moral and legal code (the African societies experienced with respect
Koran and the Sharia), the Arabic language (the to domains such as religion, language, and
key to understanding Islamic scriptures), and customs. Further, due to limited interethnic
the geographic features the Nile River and the marriages and interaction, there exists a vast
Sahara Desert. In these societies, the scope of diversity in ethnicity, religion, language, and
government and law was limited, for Islam alone customs in the southern and central African
had the total regulatory force. A clear distinction Rim. Historically, this rim was difficult to
existed between what was legal and what was access and was economically rather isolated
moral (such as in dress code), and moral inappro- from Egypt and other countries that formed the
priateness was strongly criticized and punished core of Arab civilization. The Arab societies
as a crime against Islam and the society. The pan- looked toward the southern African region
regional influence of Islam was solidified under primarily for slave trading, thereby further iso-
the Ottoman Turk Empire of medieval times and lating those societies. Subsequently, although
subsequent European imperialism (Karsh & facilitating Christian missionary efforts to set up
Karsh, 1999). Anderson (1995) emphasizes: schools and social interest groups in the region,
the British also persisted in their “divide and
European imperialism came relatively late to the rule” policy by empowering and coercing
Middle East and North Africa and was charac- indigenous chiefs to regulate local affairs to
terized by a preoccupation with the geostrategic— promote British commercial and economic
as opposed to economic—value of the region. To interests in the colonies. The history of slavery
ensure low-cost access to the region and to guar- and distrust remains strongly entrenched in the
antee their global strategic interests, the European psyche of the indigenous southerners. Stead
powers first attempted to identify and support (1996) notes that since the introduction of slav-
compliant local political authorities in protec- ery in the Cape Colony (dating back to 1658),
torates of varying degrees of formality. When the the divisions among White, Black, Indian, and
regional political system collapsed with the demise Colored people (persons of mixed racial ances-
of the Ottoman Empire in the First World War, the try) emerged and deepened as non-White people
European powers reconfigured the political and were denied many human rights. Racial separa-
economic landscape of the region, inventing new tion was greatly exacerbated after 1948 when
states and imposing European-style administra- the National Party came to power in South
tions, thereby creating a regional system that would Africa and pursued its policy of apartheid, or
secure European political interests. (p. 29) separateness. Racial segregation remains to be
bridged socially and mentally throughout the
We propose that the following GLOBE region, even after formal abolition of apartheid
societies be included in the Middle East cluster: during the 1990s.
Qatar, Morocco, Turkey, Egypt, and Kuwait. A distinctive philosophical concept in Sub-
The clustering of these nations is consistent Saharan Africa cluster is Ubuntu, which refers
with the Arabic cluster of Ronen and Shenkar to humaneness that individuals and groups dis-
(1985), though our study sampled only one play for one another (Mbigi & Maree, 1995).
nation (Kuwait) included in their Arabic cluster. Ubuntu is the “foundation for the basic values
Further, we departed from Ronen and Shenkar that manifest themselves in the ways” Sub-
(1985) by including Turkey as part of this clus- Saharan African people believe and behave
ter. We placed Turkey in the Middle East clus- toward each other and everyone else they
ter due to the strong historical and cultural links encounter (Mangaliso, 2001, p. 24). The con-
dating to the Ottoman period connecting Turkey cept is summed up in the phrase umntu ngumntu
with the other societies in the Middle East. ngabanye (a person is a person through others),
which implies that the relationship and recogni-
tion by others is at the core of a person’s iden-
Sub-Saharan Africa Cluster
tity. Consequently, the Sub-Saharan African
The Sub-Saharan African societies did not cluster is characterized by the norms of reci-
experience the kind of homogenization northern procity, suppression of self-interest, the virtue
10-House.qxd 3/19/2004 5:30 PM Page 188
of symbiosis, and human interdependence river (now dry), with the rest distributed mainly
(Mangaliso, 2001). We propose that the follow- on its key tributaries such as the Indus. Archeo-
ing GLOBE societies be included in the Sub- logically attested historical evidence shows
Saharan Africa cluster: Namibia, Zambia, India’s strong cultural influence in Southeast
Zimbabwe, Nigeria, and South Africa (Black Asian nations, dating to the first millennium A.D.
sample). It should be noted that prior empirical (Vallibhotama & Saraya, 1996). The early king-
work has not really identified this cluster before. doms in Southeast Asia were organized along the
However, for the reasons stated above, we Indian societal lines; that is, using Buddhist or
believe that there is justification for believing Brahmin officials as their advisors.
that these societies will cluster together.
The local rulers, who came into contact with the
foreigners mainly through trade, felt attracted to
Asia
certain elements of Indian civilization, which they
Prior empirical research has identified a subsequently adopted and adapted to their own
single Asian cluster (e.g., Ronen & Shenkar, needs . . . As far as “Sinization” is concerned, this
1985). However, in their discussion, Ronen was quite a different process. It was confined to
and Shenkar suggest that this single cluster northern Viet Nam . . . . (Vallibhotama & Saraya,
may be artifactual in that prior empirical 1994, p. 417)
research did not have an adequate sampling of
these societies. In GLOBE, we hypothesize that A distinct cultural feature of the region,
the Asian societies (excluding Central Asia and spanning from ancient Persia to the modern
Middle East) can be grouped into two different Philippines, was the rather peaceful and interac-
clusters: Confucian and Southern Asian. These tive coexistence of diversity over long periods.
two clusters have quite distinct worldviews. The As Levi-Strauss (1951) noted, people of widely
teachings and works of Confucius, and later differing religious beliefs coexisted peacefully
Buddha, had a distinct historical influence on in spite of these differences. For instance, the
the Confucian cluster; whereas those in the southern Asians have shown over history their
Southern Asian cluster shared Brahmanism and ability to live together despite different beliefs.
Buddhist, Islamic, and Christian influences. In the region, one frequently finds Islamic
Historically, China and India have had distinct mosques next to Hindu temples with families
spheres of influence in Confucian Asian and of gods and goddesses, and Buddhist pagodas
Southern Asian societies respectively. Huxley or Christian churches or Sikh Gurudwaras with
(1997) in his review of ancient legal systems, images of their teachers. These “complementary
for instance, suggests that during the third and forms of faith seem irreconcilable yet they
second centuries before the Common Era (B.C.E) co-exist peaceably” (Levi-Strauss, 1951). The
there flourished three distinct schools of natural region historically shared common Vedic
law originating from India, China, and the culture, and later absorbed Islamic and British
Hellenistic world, thus recognizing that the dis- influences within its boundaries.
tinctiveness of the Indian and Chinese cultures In the societies of the region, there were
dates back to these historical times. continuous efforts to realize a rich interaction
and balance among spirituality, psychology,
philosophy, morality, politics, economics, and
Southern Asia Cluster
society. The societies in the Southern Asia clus-
Radioactive dating of archeological materials ter, in particular Iran, also shared strong and
indicates a continuous development of cultural sustained interactions with the Arab cluster.
sequence in the Greater Indus Valley, covering Several of these societies, especially Malaysia
eastern Iran, southeastern Turkmenistan, Balu- and Indonesia, have enjoyed considerable
chistan, Afghanistan, Pakistan, and western migration of Chinese over the second millen-
India, since at least 5,000 B.C. (Thapar & Moghul, nium A.D. and have maintained strong economic
1996). More than 1,200 of the 1,600 known links with China, South Korea, Taiwan, and
settlements of the valley were on the Saraswati Japan over the recent decades. Further, due to
10-House.qxd 3/19/2004 5:30 PM Page 189
their peculiar colonial situation under the Confucian influence extended to those Eastern
Portuguese culture, Mexican-American influ- societies located within the China cultural orbit
ences have been quite prominent in the namely Korea, Japan, Hong Kong, Taiwan and
Philippines. In regard to the Philippines, Singapore, not to mention overseas Chinese com-
Inglehart, Petterson, and Puranen (1994) note munities everywhere. In China, itself, Confucianism
that although geographically close to China and has provided the indispensable mainstay of a sys-
Taiwan, she was not historically shaped by the tem of education that is more than two thousand
Confucian model and is culturally closer to the years old. (Retrieved from http://department.
South Asian societies and Latin American monm.edu/classics/Speel_Festschrift/brett.htm)
societies, with whom the Philippines shares a
history of Hispanic colonization. On the whole, Some distinctive Confucian teachings
despite the diversity of external influences, an included the emphasis on learning through a
interesting feature shared by all Southern Asian hierarchical, family modeled institution, which
societies is their propensity to successfully taught principles such as diligence, self-sacrifice,
assimilate external and modern influences with and delayed gratification. The Confucian model
their indigenous cultures (see Gupta, Surie, & of a family included the firm but compassionate
Chhokar, 2002, for additional references and father, the loyal child who can never fully
evidence). We propose that the following repay a deep debt to the parents, and the ances-
GLOBE societies be included in the Southern tors who are to be respected and worshipped
Asia cluster: Iran, India, Indonesia, Philippines, (Wei-ming, 1996). Therefore, we propose the
Malaysia, and Thailand. following GLOBE societies be grouped into the
Confucian cluster: Taiwan, Singapore, Hong
Kong, South Korea, China, and Japan. Table
Confucian Asia Cluster 10.1 and Figure 10.1 provide a display of the ten
The Confucian cluster is defined by the regional clusters of societies.
strong historical influence of China and
Confucian ideology. Japan, though geographi-
cally isolated from China, has had rich cultural EMPIRICAL TEST OF SOCIETAL CLUSTERS
interactions with China. Even currently, Japan
shares close affinity with the Chinese culture. We statistically tested the empirical validity of
As Lowe (1998) noted, the proposed clustering presented in Table 10.1.
Discriminant analysis is a statistical technique
Hofstede’s results also differentiate between East that develops a linear function from a set of
Asian societies on his “etic” dimensions with, for variables (i.e., societal culture dimensions in
example, Japan showing a unique cultural profile, the present study) believed to be important in
predicting differences in structure and behaviour differentiating group membership (i.e., the
with its immediate “Confucian” neighbours. The hypothesized societal clusters) of the data
differences in structure may be reliably attributed observations (i.e., GLOBE societies). We there-
to differing institutional environments, but what fore use this technique to test the hypotheses
all these Confucian societies do appear to have that our hypothesized GLOBE societal cluster-
structurally in common is a reliance upon “net- ing, as summarized in Table 10.1, is supported
works” which are co-ordinated through the mech- by our data. We will use this technique to statis-
anism of trust, and this common structure can be tically test the extent to which our classification
more reliably attributed to common cultural influ- is supported by the data.
ences. (p. 328) Because discriminant analysis builds a linear
function by using the empirical data, we ran-
Using the extensive World Values Survey, domly split our data at the individual level into
Inglehart and Carballo (1997) also identify two halves (i.e., developmental sample and
coherent value system patterns of secularism holdout sample). The developmental sample
deriving from common societal objectives in the consisted of all the GLOBE societies and was
Confucian cluster. Brett (1997) observed that the used to build the linear function connecting
10-House.qxd 3/19/2004 5:30 PM Page 190
Southern Asia
ia La
n As Philippiness tin
a Am
ci Indonesia
fu Ec er
on EI ua ic a
C ore Malaysia
S d
g ap ong India Co alv or
n
Si ng K an l a
Thailand Bo umb dor
Ho aiw na a B liv ia
T hi e Iran Gu ra ia
C Kor z
Ar atem il
t h n
u a Co gen ala
So Jap Ve sta tina
st
n R
Nor
e Ea
M ezu ica
Kuw ey
Denm
ex ela
Egy ait
dic Eu
Finla
ico
k
Middl
Swed
Qat co
Mo pt
Tu r
roc
ar
ark
nd
rope
en
Hung a
Greec
U.S da
Alban a
Slove
Aus .A.
Irela lia
Polan
Kaza
le)
Russi
Georg an
ca
Can
En nd
E a stern
tra
Sou gland
ary
e
land
(Wh th Afri
New Samp
ni
i
khst
d
a
Zim
ia
Zea
lo
ite
Na babw
E uro
An g
m ria ds
Za ibia e
mb ust erlan
pe
N A th nd
S ig ia Israel
(B outh eria e Ne erla ast
lac Af
k S ric
Italy T witz ny-E est
h
S ma -W
am a Switzerland r y
ple (French Speaking) Ge rman
) e
G
Spain
Su
b-
Sa Portugal pe
ha u ro
ra France E
Af ic
ric an
a e rm
G
L a ti n E u ro p e
the GLOBE culture dimensions to the societal Collectivism, (f) In-Group Collectivism, (g) Gen-
clusters. Although the holdout sample consisted der Egalitarianism, (h) Power Distance, and
of different respondents, it also contained data for (i) Uncertainty Avoidance.
all the GLOBE societies and this sample was used In the developmental sample (i.e., the sample
to test the robustness of our discriminant analysis. used to develop the linear discriminant func-
We used the societal practices (As Is) and soci- tion), five discriminant functions captured 92.8%
etal values (Should Be) scores on nine GLOBE of the variation among the GLOBE societal
cultural scales as the variables used to predict clusters. For this sample, 59 of the 61 societies
the group membership of our societies in were classified accurately into the hypothesized
our discriminant analysis. The nine GLOBE clusters (see Table 10.1), yielding 96.7% classi-
cultural dimensions are: (a) Performance Orien- fication reliability of the discriminant functions.
tation, (b) Assertiveness, (c) Future Orientation, Post hoc analysis revealed that only two
(d) Humane Orientation, (e) Institutional countries were not classified as predicted: Costa
10-House.qxd 3/19/2004 5:30 PM Page 191
Rica and Guatemala. It turns out that in the amounting to 59% accuracy rate for the
developmental sample, both of these countries discriminant functions. Post hoc analyses
had a greater likelihood of being classified into revealed some reasons for this drop in accuracy.
the Latin Europe cluster than into the originally Specifically, Table 10.2 provides summary data
hypothesized Latin America cluster. on the average probability of classification of
Although the results of the discriminant societies into their hypothesized cluster, and
functions on the development sample yielded the average probability of classification into
strong supportive results for the GLOBE soci- the best alternative cluster. As can be seen in
etal clustering, it is important to assess the this table, for societies in 8 of the 10 clusters,
robustness of this result because the linear func- the average probability of classification into
tions developed by the discriminant analysis their hypothesized cluster exceeded 0.75. The
might overfit the developmental data and thus, two exceptions were the Germanic Europe
the adequacy of the reclustering accuracy could (0.50) and Sub-Saharan Africa (0.53) clusters.
be inflated. Thus, we examined the adequacy Germanic European societies had a large 0.40
of the discriminant analysis linear functions to probability of being classified into the Nordic
classify the GLOBE societies in the holdout Europe cluster. Sub-Saharan African societies
sample. We achieved the following results had 0.16 probability of classification into the
using this technique. In the holdout sample, 36 Middle East cluster. Thus, geographical proximity
of the 61 societies were classified as predicted, and associated cultural interactions may be a
10-House.qxd 3/19/2004 5:30 PM Page 192
Table 10.3 Cultural Clusters Classified on Societal Culture Practices (As Is) Scores
Cultural Dimension High-Score Clusters Mid-Score Clusters Low-Score Clusters Cluster-Average Range
Performance Confucian Asia Southern Asia Latin America 3.73–4.58
Orientation Germanic Europe Sub-Saharan Africa Eastern Europe
Anglo Latin Europe
Nordic Europe
Middle East
Assertiveness Germanic Europe Sub-Saharan Africa Nordic Europe 3.66–4.55
Eastern Europe Latin America
Anglo
Middle East
Confucian Asia
Latin Europe
Southern Asia
Future Orientation Germanic Europe Confucian Asia Middle East 3.38–4.40
Nordic Europe Anglo Latin America
Southern Asia Eastern Europe
Sub-Saharan Africa
Latin Europe
Humane Orientation Southern Asia Middle East Latin Europe 3.55–4.71
Sub-Saharan Africa Anglo Germanic Europe
Nordic Europe
Latin America
Confucian Asia
Eastern Europe
Institutional Nordic Europe Anglo Germanic Europe 3.86–4.88
Collectivism Confucian Asia Southern Asia Latin Europe
Sub-Saharan Africa Latin America
Middle East
Eastern Europe
In-Group Southern Asia Sub-Saharan Africa Anglo 3.75–5.87
Collectivism Middle East Latin Europe Germanic Europe
Eastern Europe Nordic Europe
Latin America
Confucian Asia
Gender Eastern Europe Latin America Middle East 2.95–3.84
Egalitarianism Nordic Europe Anglo
Latin Europe
Sub-Saharan Africa
Southern Asia
Confucian Asia
Germanic Europe
Power Distance Southern Asia Nordic Europe 4.54–5.39
Latin America
Eastern Europe
Sub-Saharan Africa
Middle East
Latin Europe
Confucian Asia
Anglo
Germanic Europe
Uncertainty Nordic Europe Confucian Asia Middle East 3.56–5.19
Avoidance Germanic Europe Anglo Latin America
Sub-Saharan Africa Eastern Europe
Latin Europe
Southern Asia
NOTE: Means of high-score clusters are significantly higher (p < 0.05) than the rest, means of low-score clusters are significantly
lower (p < 0.05) than the rest, and means of mid-score clusters are not significantly different from the rest (p > 0.05).
10-House.qxd 3/19/2004 5:30 PM Page 194
Table 10.4 Cultural Clusters Classified on Societal Culture Values (Should Be) Scores
Cultural Dimension High-Score Clusters Mid-Score Clusters Low-Score Clusters Cluster Average Range
NOTE: Means of high-score clusters are significantly higher ( p < 0.05) than the rest, means of low-score clusters are significantly
lower (p < 0.05) than the rest, and means of mid-score clusters are not significantly different from the rest (p > 0.05).
10-House.qxd 3/19/2004 5:30 PM Page 195
Performance Orientation
Latin America
7
6
Sub-Saharan Africa Anglo
5
Eastern Europe
As Is Should Be
Assertiveness
Latin America
7
Eastern Europe
As Is Should Be
Future Orientation
Latin America
7
Eastern Europe
As Is Should Be
Humane Orientation
Latin America
7
Eastern Europe
As Is Should Be
Institutional Collectivism
Latin America
7
Eastern Europe
As Is Should Be
In-Group Collectivism
Latin America
7
Eastern Europe
As Is Should Be
Gender Egalitarianism
Latin America
7
Eastern Europe
As Is Should Be
Power Distance
Latin America
7
Eastern Europe
As Is Should Be
Uncertainty Avoidance
Latin America
7
Eastern Europe
As Is Should Be
Europe, Confucian Asia, Southern Asia, Middle weak practices scores for Power Distance and
East, and Sub-Saharan Africa clusters) are Assertiveness, and strong practices scores for
noticeably different from each other. Institutional Collectivism and Gender Egalitari-
Within the Western region, the Germanic anism. Smiley (1999), for instance, notes that
Europe cluster shows high practices scores for Nordics tend to be modest, punctual, honest,
Performance Orientation, Uncertainty Avoidance, and high-minded, and rich people generally
Future Orientation, and Assertiveness, but low dress, eat, and travel in the same style as the
practices scores for Humane Orientation, Insti- prosperous middle class, all of which reflect
tutional Collectivism, and In-Group Collectivism. underplaying of assertive, familial, and mascu-
Put differently, societies in the Germanic Europe line authority and emphasis on certainty, social
cluster rely on more assertive, and individualistic unity, and cooperation.
approaches, which are futuristic, well-defined, The Latin Europe cluster is distinguished
result oriented, and often harsh, reflecting the by weak practices scores for Institutional
technocratic orientation of the Germanic socie- Collectivism and Humane Orientation, indicating
ties, considered to be “a reaction against the Hitler the affective autonomy orientation of Latin
era. The very word ‘leader’ is ‘Führer’ in German European societies. Latin America is character-
with all that that denotes” (de Vries, 2000, ized by high practices scores for In-Group
retrieved from http://matilde.emeraldinsight. Collectivism, and low scores for Performance
com/vl=8159516/cl=63/nw=1/rpsv/now/ Orientation, Uncertainty Avoidance, Future
archive/june2000/spotlight.htm). The Nordic Orientation, and Institutional Collectivism. In
Europecluster, which is culturally most similar other words, Latin American societies tend to
to the Germanic Europe cluster, shows strong enact life as it comes, taking its unpredictability
practices scores for Uncertainty Avoidance and as a fact of life, and not overly worrying about
Future Orientation, and weak practices scores results. There is less concern with institutional
for In-Group Collectivism. However, it also has collective goals than with family bonds. Finally,
10-House.qxd 3/19/2004 5:30 PM Page 200
Anglo practices scores tend to be higher for and must not be approached using instrumental
Performance Orientation, but lower for In-Group means. Instead, the family forms the basis for
Collectivism. These traits indicate high goal insurance against future contingencies, and
orientation of Anglo societies, where rewards looks toward the male members for primary
tend to be based on merit and achievement goals protection and support. Finally, managers in the
take precedence over the family bonds. Sub-Saharan Africa societies report stronger
The respondents in the five Western clusters practices of Humane Orientation. This attribute
perceive their societal practices as lower in In- highlights the humanistic model of these societies
Group Collectivism, Humane Orientation, and that live a difficult and rural kind of life.
Power Distance, but higher in Uncertainty
Avoidance and Future Orientation. In contrast to
Metaconfiguration of Clusters
the Eastern region, they scored stronger in values
for Gender Egalitarianism, but scored weaker in We used Multidimensional Scaling Proce-
values for Uncertainty Avoidance, Future Orien- dure to assess the degree of overall distance
tation, Power Distance, and Assertiveness. These among the 10 cultural clusters, across the nine
characteristics are typical of industrialized— values and the nine practices dimensions of
structured and predictable—societies, but ones societal culture. Based on several alternate mea-
that rely on rationalism not authority. The sures of computing average distance in societal
respondents in the five Eastern clusters perceive culture dimensions, we identified a metaconfig-
their societal practices to be high in Humane uration of the distances in the cultural practices
Orientation and In-Group Collectivism, and are and values of the 10 cultural clusters. This meta-
less concerned with Gender Egalitarianism. configuration of GLOBE cultural clusters is
Respondents in these societies report higher given in Figure 10.11.
values and practices of Power Distance; and The vertical North–South axis of the meta-
although they report significantly lower practices configuration separates the five Western clusters
of uncertainty Avoidance and Future Orientation, from the five Eastern clusters. Four quadrants
they report stronger values of Uncertainty are identifiable. Quadrant I consists of Middle
Avoidance and Future Orientation. East, Sub-Saharan Africa, and Eastern Europe
Within the Eastern clusters, Eastern Europe, clusters. Quadrant II consists of Latin America
with its strong links to the rest of Europe, is and Latin Europe clusters. Quadrant III consists
distinguished by the practices of higher Gender of Anglo, Germanic Europe, and Nordic Europe
Egalitarianism and In-Group Collectivism, but clusters. Quadrant IV consists of Confucian Asia
lower Performance Orientation, Uncertainty and Southern Asia clusters. Anglo and Sub-
Avoidance, and Future Orientation. In contrast, Saharan Africa clusters appear in the center of
the Confucian Asia societal cluster is charac- the configuration. As seen from Tables 10.3 and
terized by practices of higher Performance 10.4, Anglo and Sub-Saharan Africa clusters
Orientation, Institutional Collectivism, and In- have mid-level scores on most of the practices
Group Collectivism. The goals are collective and and values dimensions of GLOBE.
family-oriented, and rewards are significant for We also examined the extent to which
performance toward meeting collective goals. individual’s values and practices in their soci-
Southern Asian societies report significantly etal culture are influenced by societal clusters,
higher practices of In-Group Collectivism as distinguished from the independent society
and Humane Orientation. These practices ref- (i.e., nation). Table 10.5 presents eta squares of
lect a deep community orientation, which is the society and cluster effects on individual val-
the hallmark of these societies. Middle Eastern ues and practices in their societal culture. Eta
societies report lower Uncertainty Avoidance, square measures the proportion of variance
Future Orientation, and Gender Egalitarianism accounted by the fixed societal or cluster
in their practices than did other clusters, but effects. The results indicate that societal cluster
greater practices of In-Group Collectivism. The effects account for more than two-thirds of the
Middle East cluster represents societies that intersociety differences in values as well as
believe the future unfolds with the will of Allah practices of Uncertainty Avoidance, Future
10-House.qxd 3/19/2004 5:30 PM Page 201
Latin Eastern
America Europe
Sub-
Latin Middle East
Saharan
Europe
Africa
WEST EAST
Nordic Confucian
Europe Asia
Orientation, and Institutional Collectivism. In This suggests that the societal cluster is an
other words, the clusters captured shared soci- appropriate and relevant unit of analysis, and that
etal attributes of the kind of goals pursued (indi- the GLOBE cluster classifications are reliable
vidual or collective), the frame of the goals indicators of reports of worldwide cultural
(short-term or futuristic), and the structure of attributes. Next, we further explore the practical
the goals (rule-based or uncertainty-oriented). (i.e., managerial implications) and theoretical
These societal attributes also discriminate advantages of our clustering of societies.
among several clusters. Tables 10.3 and 10.4,
for instance, indicated that managers in Latin
Managerial Implications
American societies report lower practices, but
higher values, for collective goals, futuristic ori- Human resource staffing: The use of regional
entation, and rule-based structures. In contrast, clusters should be of great help for the selection
Nordic European societies report lower prac- and training of managers who work in global
tices and higher values for individual goals, environments. To the extent that regional clus-
shorter-term orientation, and uncertainty- ters capture important cultural distinctions
oriented structure. both within and among clusters, the process of
Cluster effects for Power Distance and selecting managers to work in multicultural
Humane Orientation values account for only a environments should be simpler. Rather than
quarter of the intersocietal variations. As shown developing a separate selection procedure for
in Table 10.4, the range for the cluster means on 60-plus countries, it is obvious that fewer
these two dimensions is very low. On the whole, procedures would be necessary. Criteria can
cluster effects account for 54% of the inter- be developed for use in countries within a spe-
societal variation in values and 65% of the inter- cific cluster. In addition, where validity studies
societal variation in practices. Thus the majority would be extremely difficult to carry out due
of societal effects on an individual’s values and to limited samples in particular countries, the
practices are explained by the cluster factor. sample size can be increased dramatically by
10-House.qxd 3/19/2004 5:30 PM Page 202
Table 10.5 Percentage of Variance in Individual Values (Should Be) and Practices (As Is) About Societal
Cultures, Accounted by Society and Cluster Effects
GLOBE Cultural Dimensions Society Effect Cluster Effect Cluster and Society
Societal Should Be
Societal As Is
NOTE: Percentage of variances in table are eta2 obtained from a multivariate analysis of variance. All eta2 were significant at
p = 0.001.
using a grouping of country samples within managers must deal with people of vastly different
regional clusters. backgrounds. By grouping countries with similar
backgrounds, a limited set of training programs
Human resource training: Interpersonal conflict can be developed to assist managers in acquiring
based on cultural misunderstanding is likely to the skills necessary for effective leadership. For ins-
occur in multicultural environments in which tance, if a particular multinational firm promotes
10-House.qxd 3/19/2004 5:30 PM Page 203
delegation and participatory management among related hypothesis, the greater the cultural
managers, training should be provided for distance between societal clusters, the higher
employees who originate in high power distance the likelihood of failure. These, of course are
cultures. Managers and employees alike from empirical questions that can be tested and have
these cultures will likely have much less experi- both practical and theoretical implications.
ence with a participatory, value-based, or team-
oriented management philosophy.
IMPLICATIONS FOR THEORY BUILDING
Effective management: Regional clusters have
implications for managers operating within Clustering countries in the GLOBE project can
cultures. Culture dimension scores provide play an important role in organizational theory
managers with an initial assessment of the cul- development. As discussed below, the GLOBE
ture and hints on how to behave in that culture. taxonomy of cultures can assist researchers in
Using a behavioral perspective of culture, the several ways.
shared constraints of members of a certain group
limit the behavior of group members (see Bond & Sampling cultures for cross-cultural studies: It is
Smith, 1996). The event-management leadership hard to overemphasize the importance of ade-
research by Smith and colleagues illustrates this quate sampling for cross-cultural studies (van de
point. In their 25-country study, Smith and Vijver & Leung, 1997). A major issue for cross-
Peterson (1994) asked managers how much they cultural research concerns selection of countries
relied upon eight sources of meaning (e.g., for- that are required to test a specific theory. As
mal organization rules, national norms) for eight noted by van de Vijver and Leung (1997), sam-
categories of events (e.g., handling poor subor- pling may take the form of systematic, random,
dinate performance). Results indicated that event or convenience sampling, with the last least pre-
management processes were related to cultural ferred. The GLOBE cultural groupings provide
dimensions. For instance, managers in collec- researchers with a guide to systematically sam-
tivist, high power distance countries (e.g., Iran, ple respondents on the basis of specific variables
Korea, and the Philippines) showed an aversion of interest. The GLOBE clusters may be used as
to using subordinates as a source of guidance. a guide to ensure that an adequate sampling of
This type of information should assist managers cultural variability is included in the samples. It
when operating in common cultures. As another may be possible to select fewer country samples
example using the results of this GLOBE pro- if the researcher is confident that the country
ject, effective managerial leadership requires an selection represents a significant range of cul-
understanding of the desired leadership attrib- tural variables and variability on each cultural
utes from each culture. As an obvious example, variable. The sampling issue is critical to ruling
leaders that originate from Humane Orientation out explanations other than culture, a problem
societal cultures are expected to embody humane inherent in cross-cultural research.
personality attributes and behave accordingly.
Ruling out rival explanations: How should
Reducing expatriate failures: The rate of expa- researchers avoid the trap of blindly inferring
triate failure is a subject of considerable debate. that cross-national differences are the result of
Irrespective of the precise percentage of fail- cultural differences? Although it may seem
ures, all agree that the human and financial costs obvious, it is critical to pay careful attention to
associated with personnel who do not succeed initially developing a rigorous research design
are great. GLOBE researchers speculate that that minimizes the potential for alternative
expatriate failure rates should be associated explanations (Brett, 1997, p. 373). Careful
with the cultural distance between the home and selection of cross-national samples is crucial to
host country. It is reasonable to expect that fail- ruling out explanations alternative to culture. As
ure rates should be less for managers moving Campbell (1986) and van de Vijver and Leung
within a particular regional cluster and greater for (1997) argue, rival explanations to cross-cultural
managers moving across cultures. As a second, findings can be greatly reduced as the number of
10-House.qxd 3/19/2004 5:30 PM Page 204
regional differences rule. This would have this approach a step further by a priori developing
enormous implications for developing more societal clusters, validating the grouping of cultures
parsimonious leadership and organizational within the cluster, and then empirically examining
behavior theories. differences among clusters.
As a practical matter, it becomes increas-
Cross-cultural replication using qualitative ingly difficult to make sense of the cross-
research. Reliability is a problem inherent in all cultural data and communicate findings if the
research and may be particularly important in number of countries exceeds more than a hand-
qualitative research. For example, qualitative ful. As an example, consider the difficulty of
research methods frequently use in-depth inter- not only identifying which cultural dimensions
views with relatively small sample sizes. An account for differences, but also describing
important issue concerns the reliability, and country-by-country differences on these dimen-
hence validity, of findings in qualitative studies sions. For instance, assume that you have a sam-
using interpretive methods with small sample ple of four countries, two from Anglo cultures
sizes. Indigenous culture-specific results from (the United States and Australia) and two from
qualitative research should be similar to results Confucian Asia cultures (Japan and China). We
from societies within the same societal cluster can expect that each group of two similar
in contrast to societies from more distant clus- countries varies from the other group on cultural
ters. That is, increased credibility of findings dimensions such as collectivism and power
can be inferred if similar results are apparent in distance. Now increase the sample to 20 or 40
cultures that belong to a particular societal clus- countries. Although we may find a relationship
ter, but not in dissimilar clusters. Replication between a specific cultural dimension and the
within societal clusters provides a triangulation phenomenon of interest (e.g., participative
check that may be used for different research leadership style), conveying the rank order of
paradigms, but might be particularly useful for 40 specific countries on this dimension
qualitative research. becomes a terribly difficult task. It is much
easier to discuss how one particular country
Insights, parsimony, and ease of communication: (e.g., South Korea) is in the Confucian Asia
Our experience has been that it is easier to gain cluster, which scored higher on Power Distance
insights to macro cultural influences on and Collectivism than did the Anglo cluster. We
organizational behavior by viewing the “big can then predict that this country, and other
picture” rather than presenting data for more countries within this cluster, is likely to place
than 60 countries. For instance, common lan- low importance on participative leadership.
guage, geography, and the level of development
(Chemers, 1997) stood out as primary factors
that determined the Ronen and Shenkar (1985) PART B: CLIMATIC
grouping of countries. From a practical per- CLUSTERING OF SOCIETAL CULTURES
spective, by grouping societal cultures into a
set of “like-minded” clusters, we can demon- A second method to classify societal cultures is to
strate differences among societal clusters on sev- use physical climates as a basis for clustering.
eral cultural dimensions that have managerial Hofstede (1980) identified physical climate as
implications for the entire set of countries the primary force influencing societal cultures.
comprising this cluster. Hofstede (1980, p. 48) Unfortunately, due to a variety of methodological
accomplished something like this in empirically problems, there have been very limited tests of
grouping countries based on high–low scores on the strength and the direction of the relationship
his culture dimensions. The resulting figures between physical climates and societal cultures.
from Hofstede’s research that compares countries Managers may benefit considerably from
on two culture dimensions simultaneously are understanding how cultural regions may be
reproduced in numerous textbooks and journal grouped according to the climatic feature shared
articles. They capture the essence of country dif- by them. Common climatic groupings can be use-
ferences in an easy to understand format. We take ful in organizing joint research and development
10-House.qxd 3/19/2004 5:30 PM Page 206
Tropical Humid:
Costa Rica, Ecuador, Columbia, Philippines, Singapore, Indonesia, Malaysia, India
Tropical Wet and Dry (Savanna):
Guatemala, El Salvador, Venezuela, Thailand, Nigeria, Zambia, Zimbabwe
Desert:
Egypt, Israel, Kazakhstan, Kuwait, Namibia, Turkey, Qatar, South Africa (Black Sample), South Africa
(White Sample), Mexico, Iran
Subtropical Humid:
Bolivia, Brazil, Argentina, Hong Kong, Taiwan
Subtropical Wet and Dry (Mediterranean):
Albania, Greece, Italy, Portugal, Slovenia, Spain, Morocco
Marine West Coast (Maritime):
Denmark, France, Germany (former East), Germany (former West), Ireland, New Zealand, Netherlands,
Switzerland, Switzerland (French Speaking), United Kingdom
Continental:
Austria, Finland, Hungary, Japan, Korea, Sweden, Poland, Canada, Russia, China, Georgia, United
States, Australia
NOTE: The above classification is tentative. Most societies cut across several climatic areas. One could, for instance,
alternatively classify Bolivia into the Tropical Humid climate, Brazil into the Savanna climate, Iran into the Continental
climate, and Georgia into the Mediterranean climate. A more sophisticated approach to climate classification is beyond the
scope of this chapter.
a 10-year mean of 5 centimeters might have 6. Marine West Coast climate, also known as the
received 50 centimeters in one year as a result maritime climate, generally lies on the western
of an unusual intrusion of moist air, followed by sides of the continents between 35° and 50°
9 years with no measurable precipitation. Daily latitude north and south. There is significant pre-
temperature variations are usually extreme. cipitation throughout the year, though monthly
amounts vary as a function of the storm systems.
4. Subtropical humid climate is typically found The temperature tends to have few extremes.
on the eastern sides of the continents between Winters are mild and summers are cool because
20° and 35° north and south latitude. It tends to of the moderating influences of the ocean.
have a uniform distribution of precipitation
throughout the year, with summers somewhat 7. Continental climate is typically found
wetter than winters. between 30° and 60° latitude, which has large
landmasses. Summers tend to be hot and winters
5. Subtropical wet and dry climate, also known cold, with below freezing temperature means
as the Mediterranean climate, is characteristic from one to several months, because of the lack
of the western sides of the continents, which are of any moderating influences of the ocean.
between 30° and 45° latitude north and south. It Winter precipitation frequently occurs in the
tends to show the unusual combination of hot, form of snow. There are both significant yearly
dry summers and cool, wet winters. Annual and daily variations in weather.
temperature ranges are generally smaller than in
humid tropical climates, and the atmosphere Table 10.6 lists the GLOBE societies in each
tends to be drier. of the above seven climatic regions.
10-House.qxd 3/19/2004 5:30 PM Page 208
Of the several attributes distinguishing centuries, who invoked the climatic logic to
between the above climatic regions, seven stand posit and rationalize a divine determinism of the
out for their prominence (Oliver, 1996; Parker, superiority of Christian Europe (see Herder,
1995). These include: (a) square of absolute lat- 1791/1868; Ritter, 1865).
itude, as a measure of solar radiation; (b) mean In the early 20th century, Huntington (1915)
temperature (celsius); (c) rain (mm/month), as a became a strong proponent of the so-called cli-
measure of the amount of precipitation; (d) rain mate hypothesis, positing that most of the vari-
days (per month), as a measure of the fre- ances in social and economic behaviors can be
quency of precipitation; (e) atmospheric pressure explained in terms of the differences in physical
(millibars/square cm) in January; (f) atmos- climates. He concluded suboptimal climate con-
pheric pressure (millibars/square cm) in July; tributed to the low level of development in the
(g) humidity (percentage). tropical world in terms of physical, mental, and
Table 10.6 shows the average climatic statis- socioeconomic attributes of people in Asia. In
tics for the seven clusters. his view, the relationship between climate and
Consistent with the predictions of Greek schol- efficiency depended on climate both directly
ars and Koeppen, seven climatic clusters and indirectly “through race, diet, parasitic dis-
accounted for about 84% of the intersociety vari- eases, hygiene, sanitation and social and political
ation in absolute latitude, and more than 50% customs” (Huntington, William, & Valkenburg,
variation in temperature, rainfall, and rain days. 1933, p. 124).
Thus, the climatic clustering of 61 GLOBE dis- Subsequent scholars particularly emphasized
criminates among societies by the amount of solar the thesis that climate influences labor productiv-
radiation as well as by the mean temperature and ity levels. Specifically, an equivalent amount of
the amount and frequency of rain precipitation. physical effort is more productive in temperate
and polar regions than it is in equatorial regions,
where greater heat demands higher amount of
Relationship Between Physical
effort for a given level of performance (Parker,
Climate and Societal Culture Dimensions 1995). Year-round heat in tropical regions
Many scholars have suggested that physical encourages the proliferation of insects and para-
climate is an important factor influencing sites, which results in faster transmission of dis-
culture (Diamond, 1997; Hausmann, 2001; eases (Landes, 1998). In contrast, winter is less
Hofstede, 1980; Huntington, 1915; Landes, supportive of insect and parasite population and
1998; Montesquieu, 1748/1989; Sachs, 2001). enables the people in colder regions to enjoy
Indeed, physical climate hypothesis has been healthier lives. Using data from the World Health
nearly as popular for explaining societal behav- Organization, Landes (1998) observed that in
iors as the competing hypotheses of religion Sub-Saharan Africa many people harbor not one
(Weber, 1904/1930) and history-shaped institu- parasite, but several, and hence are too sick to
tions (North, 1990). work. However, others show that tropical diseases
In one of the early studies, Montesquieu (1748/ cause about a fourth of human deaths in compari-
1989) articulated how physical climate shapes a son to respiratory diseases, most of which are not
range of sociocultural behaviors, including work as important in humid tropics as in cooler regions
ethic, religious beliefs, aggression, combativeness, (Porter & Sheppard, 1998). One may speculate
mortality, fertility, obesity, sexism, and indus- that the climatic influences on labor productivity
trial development (agriculture, trade, and com- and health may be relevant for societal culture
merce). Several of these sociocultural behaviors dimensions such as power distance. In conditions
can be seen as unobtrusive manifestations of the where the labor productivity is low, people may
societal culture dimensions, such as Performance be more willing to allow some control of labor for
Orientation (work ethic), Assertiveness (aggres- its effective organization, and thus societies could
sion, combativeness), and Gender Egalitarianism evidence higher levels of power distance.
(sexism). The climatic explanation for cultural More recently, Diamond (1997) also asserted
behaviors gained much popularity among the that environment molds history. He proposed that
European intellectuals of the 18th and 19th all of the important differences between human
10-House.qxd
3/19/2004
5:30 PM
Table 10.7 Mean Attributes of Climatic Regions, Using GLOBE Society Sample
Atmospheric Atmospheric
Page 209
Climatic Regions Latitude Temperature Rainfall Humidity Rain Days Pressure July Pressure January N
209
10-House.qxd 3/19/2004 5:30 PM Page 210
societies, and all of the differences that led some plow, systems of crop rotation, domesticated
societies to prosper and progress and others to lag animals, and high productivity European
behind, were due to the nature of each society’s crops such as wheat and rice, had comparatively
local environment and its geographical location. limited influence in the tropical areas. One
In his view, culture (termed “proximate factor”) may speculate that the advanced technologies,
itself is determined by the climate (termed “ulti- such as the heavy plow and high productivity
mate factor”), and has a relatively minor role in crops, would be relevant for societal culture
understanding societies’ historical practices. In dimensions such as in-group collectivism.
particular, he emphasized the significance of the Availability of advanced technologies can reduce
East–West geographical layout of Eurasia, and the need for relying on the family and group
the North–South layout in Africa and the effort and help inculcate more independent
Americas. He posited that because climate values and practices.
changes little with longitude but quite rapidly Hausmann (2001) summarized several devel-
with latitude, the Eurasian nations enjoyed fairly opmental liabilities for the societies that are tropi-
uniform climatic conditions. Hence, agricultural cal, landlocked and far from the coast. First, the
innovations developed in one region could diffuse average income of the tropical nations currently is
and be shared by several societies, generating a about a third of the income of temperate-zone
large pool of common plant and animal varieties societies. Except for Singapore and Brunei, none
available throughout the region. In contrast, the of the tropical nations is in the top 30 richest
diffusion of the new varieties in the Americas or nations of the world. Over the past few decades,
in Africa was limited because of significant vari- the economic growth rates in the tropical nations
ation in climates, and so the technological growth have been one and a half percentage points lower
was local and economic growth slow. One may than that in the temperate nations (Hausmann,
speculate that the greater feasibility of cross- 2001). Second, coastal nations with their entire
societal exchange and higher returns on technol- population within 100 kilometers of the sea tend to
ogy development in Eurasia would be associated grow 0.6 percent faster per year than the nations
with several societal culture dimensions, such as whose population is farther than 100 kilometers
future orientation and uncertainty avoidance. from the sea. The cost of shipping goods a kilo-
Sachs (2001), on the other hand, asserted that meter over the sea is a seventh of the cost of ship-
the temperate climates supported greater popu- ping over the land (Hausmann, 2001). The lack of
lation over the past centuries and thus facilitated access to cheap shipping impedes the ability of
more rapid advances in temperate climate tech- these nations’ economies to take-off because the
nologies than in tropical climate technologies. societies in early stages of development tend to
Further, “Since technologies in the critical areas have an advantage in bulky, low-value-added
of agriculture, health, and related areas could goods for which cost of shipping is important for
diffuse within ecological zones, but not across gaining economies of scale. Third, because land-
ecological zones, economic development spread locked nations’ transported goods need to cross
through the temperate zones but not through the other nations’ borders, they have particularly high
tropical regions” (Sachs, 2001, p. 12). Acemoglu, transportation costs. The median landlocked
Johnson, and Robinson (2002) elaborated that country pays up to 50% more in transportation
there has been a gradual drift over the course costs than the median coastal nation (Hausmann,
of history to development of more advanced 2001). The cost of shipping a standard container
technologies in the temperate climates. More from the United States to a landlocked African
specifically, the tropical areas offered a favor- nation can be up to four times the cost of shipping
able environment for early civilizations, reflected the same to a coastal African nation. The coordi-
in terms of the tropical origin of humans and nation of infrastructure expenditures among the
lower-required calorie intake in the warmer landlocked nations can be a big impediment. The
areas. However, appropriate technologies devel- landlocked Paraguay, for instance, suffered from
oped more rapidly in the temperate areas and low agricultural productivity until the Mercosur
made temperate societies more productive. agreement in the mid-1990s enabled barge
These technologies, which included the heavy transportation through Brazil and Argentina
10-House.qxd 3/19/2004 5:30 PM Page 211
(Hausmann, 2001). The coastal versus landlocked of technology, and such democracy inevitably
factor, which is an important variable in climate meant a culture of questioning authority—or
determination, could potentially be significant for lack of power distance. In contrast, the stable
several societal culture dimensions such as institu- conditions in the tropical zone facilitated agri-
tional collectivism and humane orientation. In culture in rural communities. There was no need
noncoastal, landlocked regions, a more collective for education of the lower classes: they could
spirit could emerge, and also there may be more farm and live a self-sustained life. Accordingly,
concern about the humane condition because the military or autocratic rulers gained power, and
attention of people is focused primarily on the enforced a culture of dependency and lack of
internal society and proximate neighbors, as questioning of authority.
opposed to trading exchange with the larger world. Taking into account these contrasting
perspectives, GLOBE proposes that it is mis-
leading to attribute a zero-sum relationship
Climate–Culture Relationship Hypothesis
to the climatic variables. Under extreme condi-
In summary, physical climate may shape sev- tions, the physical climate variables may have
eral socioeconomic behaviors, which we expect a significant influence on the socioeconomic
will be associated with the nine GLOBE societal behaviors and GLOBE societal culture dimen-
culture dimensions. Interestingly, there are two sions. On the other hand, under moderate con-
different and contrasting views on the relation- ditions, the physical climate variables may not
ship between physical climate and societal have any significant role in socioeconomic
culture dimensions. behaviors, and religion and historically developed
According to Landes (1998), positive physical institutions could instead be more pertinent shapers
climate shapes development-supporting societal of societal culture dimensions.
culture dimensions. He notes that better climate There exists substantial prior research evidence
is the first and foremost factor accounting for supporting the above extreme climate hypothe-
the wealth of Europe (and the West) in compar- sis. Those interested in exploring this evidence
ison with the poverty of the other nations. In his in greater depth should refer to Parker (1995).
view, such climatic conditions as better soils, Parker lists 3,000-odd references on research
better topography, better natural vegetation, that has evaluated the link between climate and
better health, and better nutrition, have played physiological effects (acclimation and physical
a critical role in technological and economic comfort, allergies, diet and nutrition, disease
progress in Europe. and illness, hormones, metabolism and ther-
In contrast, Hofstede (1980) holds that the moregulation), psychological effects (affective
adverse physical climate shapes development- disorders, aggression, cognition, mood and per-
supporting societal culture dimensions. Hofstede sonality, mental illnesses, and stress), sociolog-
used geographical latitude of capital cities as a ical effects (accidents, culture, crime, birth and
proxy for climate and found that 43 percent of death rates, migration), and economic effects
the variation in the cultural dimension of power (food, housing, clothing, health care, agricul-
distance is accounted by climatic factors. In his ture, manufacturing, tourism, and savings). Of
research, tropical countries, with the exception the seven climatic attributes, extremity of ther-
of Israel, tended to be higher on power distance, moclimate (temperature) has a particularly
compared with the higher latitude countries. He important influence on the societal behaviors.
proposed that in the colder zone, the adverse On the one hand, extreme cold in the polar
climate created survival challenges for the tra- regions has an important influence on human
ditional societies. They needed to develop behaviors. Here, people show more frequent
appropriate technology for surviving the tough symptoms of seasonal affective disorders, depres-
cold conditions. Such technology could not be sion, and suicidal intent and occurrences (Lamberg,
developed and applied without urbanization, 1994; Lewy, Wehr, Goodwin, Newsome, &
mass literacy, social mobility, and emergence Markey, 1980). In these regions, the mood disord-
of a middle class. Further, a representative ers rise during the fall and winter months, and
government was needed to encourage the use peak toward the end of winter. Specifically, human
10-House.qxd 3/19/2004 5:30 PM Page 212
body compensates for external temperatures lower climate-related products, including housing,
than 37 degrees Celsius, the normal body temper- clothing, food, and household energy; the figure
ature, by inducing overeating and weight gain for goes up to 90% in societies with lower levels of
enhanced hormonal secretion (Parker, 1995). A socioeconomic development (Parker, 2000).
small part of the brain, the pineal body, monitors On the other hand, in moderate climates,
the amount and quality of light received by the societal development would be expected to be
human eyes. With reduced light, the pineal body rather independent of the constraining influ-
secretes extra hormonal chemical melatonin, ences of the climatic attributes such as thermo-
which controls sleep and puts the body into a climate. Van de Vliert and colleagues (2002),
hibernating mode for the winter months, thereby for instance, studied 33,584 voluntary workers’
generating mood disorders and aversion to work reasons for helping in a sample of 33 nations.
(Lewy et. al, 1980). In the United States, the pre- They found that an individual volunteer’s
valence of winter affective disorder rises dramati- self-serving and altruistic motivations were un-
cally from Florida to Maine and from southern related with comfortable climates (i.e., moderate-
California to Oregon, in tandem with declining temperature climates), but were related with
hours of daylight (Lamberg, 1994). uncomfortably colder or hotter climates. On the
Similarly, extreme heat also has an important whole, physical climate may be expected to
influence on societal behaviors. In Southern Asia, shape distinctive patterns of cultural behaviors
for instance, rural communities have adopted in each physical climate cluster, depending on
behaviors that enable them to adapt to the hot and the specific climatic variables that are salient in
humid tropical climate efficiently. They tend to that cluster. Yet, the effect of physical climate is
rest during the peak hot hours and work during the not likely to be determinative; people may adapt
cooler mornings and evenings. Srinanda (1997) to the effects of physical climate and may enact
notes that the Southern Asian farmers cultural behaviors that lack any relationship
with the physical climate.
adopted numerous time-tested techniques in soil In summary, GLOBE proposes that
management, crop combinations, irrigation and Hypothesis 1: The physical climate accounts
drainage and general land use practices to mini- for small variance in societal culture and leaves
mize the adverse effects of excesses of heat and overwhelming flexibility for societies to shape
moisture or seasonal moisture deficits, the impact their societal cultures.
of high rainfall intensities or strong winds. (p. 226)
Table 10.8 Chi Square Test for the Dominant Relationship Between Climatic Regions and Cultural
Clusters, Using GLOBE Sample
** p < 0.5.
** p < .01.
and expected frequencies/agregated expected overlap in the cultures of Nordic and Germanic
frequencies. societies, and of Sub-Saharan Africa and
For a valid test of independence, the expected Middle East societies.
value in each cell should be at least 1. Therefore, The chi square for the relationship between
we collapsed tropical humid and tropical wet and cultural and climate clusters was significant:
dry regions into a composite tropical group; and 79.30 (p < .01, df = 28). We computed the max-
subtropical humid and subtropical wet and dry imum chi square component for each of the
regions into a composite subtropical group. In cultural clusters, as shown in Table 10.8. The
using latitude as a basis for developing a com- physical climate region associated with the
mon tropical grouping and a common subtropi- maximum chi square component of a cultural
cal grouping, we were guided by the literature cluster indicates the type of climate typically
that latitude is the most important factor in deter- associated with a given type of culture.
mining physical climates and also our finding As shown in Table 10.8, the Latin American
that latitude accounts for greatest variation in the and Southern Asian societies tend to be in trop-
seven physical climate regions. ical climates, Sub-Saharan Africa and Middle
We also collapsed the cultural clusters of East societies in desert climates, Latin European
Nordic Europe and Germanic Europe into a societies in subtropical climates, Anglo and
composite Northern Europe group; and the cul- Nordic Europe and Germanic Europe societies
tural clusters of Sub-Saharan Africa and Middle in maritime climates, and Eastern European and
East into a composite Greater Africa group. In Confucian Asian societies in continental cli-
choosing the cultural clusters to combine, we were mates. These eight components account for
guided by the literature that Northern Europe 44.71 of the total 79.30 (p < .01). After taking
and Greater Africa have distinctive cultures, and into account these strongest culture–climate
the finding of Part A that there is considerable components, there is much weaker, though still
10-House.qxd 3/19/2004 5:30 PM Page 214
Table 10.9 General Linear Multivariate Analysis: Physical Climate and Societal Culture
General
Linear Model Wilks’ Lambda F Hypothesis df Error df Significance Eta Squared
significant, evidence (p < .05) for the relation- Middle Eastern societies have desert climate.
ship between culture and climate. Anglo and Germanic societies have maritime
To analyze this residual relationship, we climate. Nordic, Eastern Europe, and Confucian
reexamined our assumption that the Sub-Saharan Asian societies generally have continental climate.
Africa and Middle East clusters are characterized Put differently, different climatic conditions
by one dominant climate. We conducted a chi are indeed associated with alternative cultural
square analysis using only 10 Greater Africa clusters. Now we investigate if the climatic
societies, of which 3 were from tropical climate conditions, as evident in the GLOBE sample of
(all in Sub-Saharan Africa culture), 6 were from societies, overwhelm the ability of the societies
desert climate, and 1 was from Mediterranean to shape their societal cultures. The results of
climate. The societies in Sub-Saharan Africa the Multivariate General Linear Analysis are
cluster were associated with tropical climate, shown in Table 10.9.
whereas those in Middle East cluster had a As shown in Table 10.9, the physical
dominant association with desert climate. climatic region accounts for 31% of the variation
Similarly, we reexamined the assumption in cultural practices and 35% of the variation in
that the Nordic and Germanic clusters are cultural values. On the other hand, cultural clus-
characterized by one dominant climate. We ter accounts for 47% of the variation in climatic
conducted a chi square analysis using only 8 attributes. Thus, only about a third of the varia-
Northern European societies, of which 5 were tion in cultural practices and values may be
from maritime climate (only 1 in Nordic), and attributed to the differences in physical climates.
3 were from continental climate (only 1 in Further, there is about as much variation in
Germanic). The findings indicated that the physical climate attributes within each cultural
societies in the Germanic cluster are associated cluster, as there is between different cultural
with maritime climate, and those in the Nordic clusters. On the whole, on the basis of the analy-
cluster are associated with continental climate. sis of the GLOBE sample, these findings confirm
Thus, in summary, tropical climate is pro- Hypothesis 1 that the societies enjoy consider-
minent in Latin American, Southern Asian, and able flexibility in shaping their cultures. Physical
Sub-Saharan African societies. Latin European climates do not preclude or dominate the influence
societies tend to have subtropical climate, and of other factors, such as religion and institutional
10-House.qxd 3/19/2004 5:30 PM Page 215
history, as manifested in and captured by the effects of historical forces and climatic forces on
cultural clusters. societal behavior.
cultural patterns. In D. Landis & R. Bhagat Einola, K., & Turgeon, N. (2000). Scandinavian
(Eds.), Handbook of intercultural training (2nd television markets: A regional perspective
ed., pp. 327–348). Thousand Oaks, CA: Sage. (Working Paper No. 2000–05). Montreal:
Alisov, B. P. (1956). Climate of the USSR. Moscow: CETAI-HEC.
Moscow State University publications. Furnham, A., Kirkcaldy, B. D., & Lynn, R. (1994).
Anderson, L. (1995). Peace and democracy in the National attitudes, competitive, money and
Middle East: The constraints of soft budgets. work: First, second and third world differences.
Journal of International Affairs, 49(1), 25–44. Human Relations, 47, 1, 119–132.
Aristotle. (1969). Meteorologica (H. D. Lee, Trans.). Gupta, V., Surie, G., & Chhokar, J. (2002). Cultural
Cambridge, MA: Harvard University Press. worldviews and their foundations: The case of
(Original work dates to 334 B.C.) Southern Asia. In B. Pattanayak & V. Gupta
Berry, J. W., Poortinga, Y. H., Segall, M. H., & (Eds.), Creating performing organizations:
Dasen, P. R. (1992). Cross-cultural psychology: International perspectives for Indian manage-
Research and applications. NY: Cambridge ment (pp. 440-465). New Delhi: Sage.
University Press. Guttman, L. (1968). A general normative technique for
Bond, M. H., & Smith, P. B. (1996). Cross-cultural finding the smallest coordinate space for a con-
social and organizational psychology. Annual figuration of points. Psychometrika, 33, 469–506.
Review of Psychology, 47, 205–235. Haire, M., Ghiselli, E., & Porter, L. (1966). Managerial
Brett, C. (1997). The importance of learning in the thinking: An international study. New York: Wiley.
Ancient East: As illuminated by Confucian Hausmann, R. (2001). Prisoners of geography.
aphorisms. In T. J. Sienkewicz & J. E. Betts Foreign Policy, 122, 44–53.
(Eds.), Festschrift in honor of Charles Speel Herder, J. G. (1868). Reflections on the philosophy of
(Part IV, chap. 1). Retrieved from http:// the history of mankind. Chicago: University of
department.onm.edu/classics/Speel_Festschrift/ Chicago Press. (Original work published 1791)
brett.htm Hofstede, G. (1976). Nationality and espoused values
Brodbeck, F. C., Frese, M., Ackerblom, S., Audia, G., of managers. Journal of Applied Psychology,
Bakacsi, G., Bendova, H., et al. (2000). Cultural 61(2), 148–155.
variation of leadership prototypes across 22 Hofstede, G. (1980). Culture’s consequences:
European countries. Journal of Occupational International differences in work-related values.
and Organizational Psychology, 73,1–29. Beverly Hills, CA: Sage.
Campbell, J. D. (1986). Similarity and uniqueness: Huntington, E. (1915). Civilization and climate. New
The effects of attribute type, relevance, and indi- Haven: Yale University Press.
vidual differences in self-esteem and depression. Huntington, E., William, F. E., & Valkenburg, S.
Journal of Personality and Social Psychology, (1933). Economic and social geography.
50, 281–294. New York: John Wiley.
Cattell, R. (1950). The principal culture patterns Huxley, A. (1997). Golden yoke, silken text. The Yale
discoverable in the syntax dimensions of exist- Law Journal, 6(106), 1885–1951.
ing nations. Journal of Social Psychology, 32, Hyden, G. (1983). No shortcuts to progress: African
215–253. development management in perspective.
Chemers, M. M. (1997). An integrative theory of Berkeley: University of California Press.
leadership. Mahwah, NJ: Lawrence Erlbaum. Inglehart, R., & Baker, W. E. (2000). Modernization,
Danowitz, A. K., Nassef, Y., & Goodman, S. E. (1995). cultural change, and the persistence of traditional
Cyberspace across the Sahara: Computing in values. American Sociological Review, 65, 19–51.
North Africa. Communications of the ACM, Inglehart, R., & Carballo, M. (1997). Does Latin
38(12), 23–28. America exist? (And is there a Confucian cul-
Diamond, J. M. (1997). Guns, germs, and steel: ture?): A global analysis of cross-cultural differ-
The fates of human societies, New York: ences. Political Science & Politics, 30(1), 34–46.
W. W. Norton. Inglehart, R., Petterson, T., & Puranen, B. (1994).
Dunlap, J. F., Harbison, F. H., Kerr, H. C., & Global coordination of the World Values Surveys:
Myers, C. (1975). Industrialism and industrial A proposal to the ban of Sweden Tercentenary
man reconsidered. Princeton, NJ: Princeton Foundation. Available from http://wvs.isr.umich.
University Press. edu/papers/global.htm
10-House.qxd 3/19/2004 5:30 PM Page 217
Karsh, E., & Karsh, I. (1999). Empires of the sand: The Osland, J. S., De Franco, S., & Osland, A. (1999).
struggle for mastery in the Middle East, 1789–1923. Organizational implications of Latin American
Boston, MA: Harvard University Press. culture: Lessons for the expatriate manager.
Kerrigan, J. E., & Luke, J. (1987). Management Journal of Management Inquiry, 8(2), 219–234.
training strategies for developing countries. Ozment, S. E. (1986). The age of reform 1250–1550:
Boulder, CO: Lynne Rienner. An intellectual and religious history of late
Koeppen, W. (1936). Handbuch der Klimatologie medieval and reformation Europe. New Haven:
[Handbook of climatology]. Berlin: Verlagsbu- Yale University Press.
chhandlung. Parker, P. M. (1995). Climatic effects on individual,
Lamberg, L. (1994). Body rhythms: Chronobiology and social and economic behavior: A physioeco-
peak performance. New York: William Morrow. nomic review of research across disciplines.
Landes, D. S. (1998). The wealth and poverty of Westport, CT: Greenwood.
nations: Why some are so rich and some so poor. Parker, P. M. (2000). Physioeconomics: The basis for
New York: W.W. Norton. long-run economic growth. Cambridge: MIT Press.
Latourette, K. S. (1965). Christianity through the Pascal, B. (1995). Pensees 1623–62 (A. J.
ages. New York: Harper & Row. Krailsheimer, Trans.). New York: Penguin Books.
Levi-Strauss, C. (1951). Foreword to documents on Pliny the Elder. (1991). Natural history: A selection
South Asia. International Social Science (J. F. Healy, Trans.). New York: Penguin Books.
Bulletin, III(4). Porter, P., & Sheppard, E. (1998). A world of differ-
Lewy, A. J., Wehr, T. A., Goodwin, F. K., Newsome, ence: Society, nature, development. New York:
D. A., & Markey, S. P. (1980). Light suppresses Guilford Press.
melatonin secretion in humans. Science, 210 Portes, A., & Zhou, M. (1994). Should immigrants
(4475), 1267–1269. assimilate? Public Interest, 116, 18–33.
Lord, R. G. (2000). Leadership. In A. E. Kazdin (Ed.), Ritter, K. (1865). Comparative geography (W. Gage,
Encyclopedia of psychology (Vol. 3, pp. 775– Trans.). Philadelphia: Lippincott.
786). Washington, DC: American Psychological Ronen, S., & Kraut, A. (1977). Similarities among
Association. countries based on employee work values and
Lowe, S. (1998). Culture and network institutions in attitudes. Columbia Journal of World Business,
Hong Kong: A hierarchy of perspectives. A 12, 89–96.
response to Wilkinson. ‘Culture institutions and Ronen, S., & Shenkar, O. (1985). Clustering
business in East Asia.’ Organization Studies, countries on attitudinal dimensions: A review
19(2), 321–343. and synthesis. Academy of Management Review,
Mangaliso, M. P. (2001). Building competitive 10(3), 435–454.
advantage from Ubuntu: Management lessons Rosenn, K. S. (1988). A comparison of Latin
from South Africa. Academy of Management American and North American legal traditions. In
Executive, 15(3), 23–33. L. A. Tavis (Ed.), Multinational managers and
Mbigi, L., & Maree, J. (1995). Ubuntu: The spirit of host government interactions (pp. 127–152).
African transformational management. Randburg, South Bend, IN: University of Notre Dame Press.
South Africa: Knowledge Resources (Pty) Ltd. Rugman, A. (2001). The end of globalization: Why
McNeill, W. H. (1964). Europe’s steppe frontier, global strategy is a myth & how to profit from
1500–1800. Chicago: University of Chicago Press. the realities of regional markets. New York:
Montesquieu, C. (1989). The spirit of laws (A. M. AMACOM.
Cohler, B. C. Miller, & H. S. Stone, Trans.). Sachs, J. D. (2001). Tropical underdevelopment
New York: Cambridge University Press. (Original (NBER Working Paper No. 8119). Cambridge,
work published 1748) MA: NBER.
North, D. C. (1990). Institutions, institutional change Saloner, G., Shepard, A., & Podolny, J. (2001). Stra-
and economic performance. New York: Cambridge tegic management. New York: John Wiley.
University Press. Schake, L. M. (1998). The Schakes of La Charette:
Oliver, J. E. (1996). Climatic zones. In S. H. 1855–1996. Available from http://www.rootsweb.
Schneider (Ed.), Encyclopedia of climate and com/~mowarren/schake/part1b.html
weather (Vol. 1., pp. 141–145). New York: Schwartz, S. H. (1999). A theory of cultural values
Oxford University Press. and some implications for work. Applied
10-House.qxd 3/19/2004 5:30 PM Page 218
Psychology: An International Review, 48(1), From the third millennium to the seventh century
23–47. BC (pp. 246–265). Geneva: UNESCO.
Sirota, D., & Greenwood, M. (1971). Understand your Tixier, M. (1994). Management and communication
overseas work force. Harvard Business Review, styles in Europe: Can they be compared and
19, 53–60. matched? Employee Relations, 16(1), 8–26.
Smiley, X. (1999, January 23). Survey: The Nordic Toynbee, A. (1947). A study of history. New York:
countries: Happy family? The Economist, Oxford University Press.
350(8103), N3–N6. Trepo, G. (1973, Autumn). Management style a la
Smith, P. B. (1997). Leadership in Europe: Euro- Francaise. European Business, 71–79.
management or the footprint of history? Triandis, H. C. (1993). Collectivism and individual-
European Journal of Work and Organizational ism as cultural syndromes. Cross-Cultural
Psychology, 6, 375–386. Research, 27, 155–180.
Smith, P. B., & Bond, M. (1994). Social psychology Trompenaars, F. (1993). Riding the waves of culture:
across cultures. Boston: Allyn & Bacon. Understanding cultural diversity in business.
Smith, P. B., Dugan, S., & Trompenaars, F. (1996). London: Economist Books.
National culture and the values of organizational Vallibhotama, S., & Saraya, D. (1996). South-east
employees. Journal of Cross-Cultural Psychology, Asia from AD 300 to 700. In J. Herrmann &
27, 231–264. E. Zurcher (Eds.), History of humanity: Vol. III.
Smith, P. B. & Peterson, M. F. (1994). Leadership as From the seventh century BC to the seventh
event-management: A cross-cultural survey century AD. Geneva: UNESCO.
based upon middle managers from 25 nations. In van de Vijver, F., & Leung, K. (1997). Methods and
Cross-cultural studies of event management. data analysis for cross-cultural research.
Symposium conducted at the 23rd International Thousand Oaks, CA: Sage.
Congress of Applied Psychology, Madrid, Spain. Van de Vliert, E., Huang, X., & Levine, R. V.
Sowards, S. W. (1996). Twenty-five lectures on (2002). National wealth and thermal climate
modern Balkan history (The Balkans in The Age as predictors of motives for volunteer work.
of Nationalism). Available from http://www.lib. Unpublished paper, University of Groningen,
msu.edu/sowards/balkan/index.htm The Netherlands.
Srinanda, K. U. (1997). Climate and societies in de Vries, M. K. (2000). Spotlight on Manfred Kets
Southeast Asia. In M. Yoshino, M. Domroes, A. de Vries. In Emerald Now, available from
Douguedroit, J. Paszynski, & L. Nkemdirim http://www.emeraldinsight.com/now/archive/
(Eds.), Climate and societies: A climatological june2000/spotlight.htm
perspective (pp. 203–233). The Netherlands: Weber, M. (1930) The protestant ethic and the spirit
Kluwer Academic. of capitalism. New York: Harper Collins
Stead, G. B. (1996). Career development of black Academic. (Original work published 1904)
South African adolescents: A developmental- Wei-ming, T. (ed.) (1996). Confucian traditions in
contextual perspective. Journal of Counseling East Asian modernity. Cambridge, MA: Harvard
and Development, 74(3), 270-275. University Press.
Stenton, F. M. (1971). Anglo-Saxon England. The Williamson, R. (1999). Who were the Anglo Saxons?
Oxford history of England (3rd ed.). Oxford: Available from http://www.regia.org/Saxon1.htm.
Clarendon Press. Woliver, R., & Cattell, R. (1981). Reoccurring
Strabo. (1917-1932). The geography of Strabo: Vols. national patterns from 30 years of multivariate
1-8 (H. L. Jones. Trans.). Cambridge: Harvard cross-cultural studies. International Journal of
University Press. (Original work dates to 9–5 B.C.) Psychology, 14, 171–198.
Swanton, M. (Trans.). (2000). The Anglo-Saxon Zander, L. (1997). The licence to lead: An 18 country
chronicles. London: Phoenix Press. study of the relationship between employees’pref-
Thapar, B. K., & Moghul, M. R. (1996). The Indus erences regarding interpersonal leadership and
Valley (3000–1500 B.C.). In A. H. Dani & national culture. Sweden: Stockholm School of
J. P. Mohen (Eds.), History of humanity: Vol. II. Economics.
11-House.qxd 3/19/2004 4:15 PM Page 219
11
RATIONALE FOR GLOBE
STATISTICAL ANALYSES
Societal Rankings and Test of Hypotheses
PAUL J. HANGES
MARCUS W. DICKSON
MINA T. SIPE
n this chapter, we discuss the rationale aid the reader’s understanding and interpretation
AUTHORS’ NOTE: We would like to thank Paul Bleise, Katherine Klein, and Robert Ployhart for their
very helpful comments on an earlier draft of this material. We would also like to thank David Hofmann for his
helpful comments on hierarchical linear modeling.
219
11-House.qxd 3/19/2004 4:15 PM Page 220
to generate these rankings. Thus, all differences scale and rxx represents the reliability of this
in the relative position of societies are inter- scale. For our ranking of societies, we used
preted as meaningful even though the magni- generalizability theory as our reliability esti-
tude of the societal difference with regard to mate to estimate our bands because generaliz-
the actual scale values is sometimes trivial. ability theory combines the information about
In an attempt to address this problem, we used the scale’s internal consistency and interrater
a technique discussed in the psychometric and reliability (Hanges & Dickson, Chapter 8). The
personnel selection literature known as test SEDs for our societal cultural and CLT scales
banding (Cascio, Outtz, Zedeck, & Goldstein, are shown in Table 11.1.
1991; Guion, 1998; Hanges, Grojean, & Smith, The actual size of the test band is determined
2000; Zedeck, Cascio, Goldstein, & Outtz, 1996). by multiplying the SED by some standardized
This procedure groups test scores into bands in normal distance (C):
which the scores within a particular band are
considered as being not meaningfully different. Bandwidth = C × SED (11.2)
The rationale for such banding lies with the
concept of measurement unreliability. In other For a 95% confidence around the test band, C
words, the premise for using test banding in is set to 1.96 (Hanges et al., 2000).
the present context is that there is an imperfect Once the width of a test band for a particular
relationship between the obtained scale score scale is determined, we applied the test band
for a society on a particular cultural dimension concept in the following manner. For a particu-
and that society’s population value (i.e., the lar cultural dimension, the GLOBE societies
scale score obtained if all society members were placed in rank order. One bandwidth was
complete the survey) on that cultural dimension. then subtracted from the scale score of the
The primary factor contributing to this imper- society holding the top position in the rankings.
fect relationship is the presence of random error. All societies that fall within the top band are
Random error can result in two or more societies considered to be equivalent even though their
appearing different on the obtained cultural scale relative positions in the rankings differ. The
scores (and thus being ranked differently) even remaining bands are derived by taking the scale
though these societies are in actuality essen- score for the highest ranked society that did not
tially equivalent on the construct of interest. fall within the first band and subtracting one
Test banding was thus developed to identify a bandwidth from that score, and so on.
range of scores that cannot be distinguished For example, as shown in Sully de Luque
from the top score in a band. It should be noted, and Javidan’s chapter on uncertainty avoidance
however, that using this banding procedure for (Chapter 19), Thailand is the top-ranked society
the GLOBE scales assumes that the influence of in terms of societal Uncertainty Avoidance val-
systematic biases such as construct misspecifi- ues whereas Greece is in the 17th position on
cation, biased sampling of construct domain, and this dimension (see Table 19.8). The differences
question-construction problems are minimal. in the rank position of these two societies may
The width of a test band is a function of the give the impression that they meaningfully dif-
confidence level that the researcher requires fer on the degree to which Uncertainty Avoi-
before declaring that two societies truly differ dance values are espoused. However, as shown
and by the standard error of the difference (SED) in Table 19.8, both of these countries fall within
of some measure (Cascio et al., 1991; Guion, the same band. In other words, in contrast to the
1998; Zedeck et al., 1996). The SED is calculated conclusion one would reach by just examining
by the following formula (Gulliksen, 1950): the rankings, Greece cannot be reliably differ-
entiated from Thailand in terms of societal
SED = Sx 2 (1 − rxx´) (11.1) Uncertainty Avoidance values. As this example
illustrates, the test banding procedure provides
useful additional information to help the reader
In this equation, Sx represents the standard appropriately interpret differences among the
deviation of the GLOBE societies on some culture GLOBE societies.1
11-House.qxd 3/19/2004 4:15 PM Page 221
SED
meso (House, Rousseau, & Thomas, 1995), or able to handle constructs operating at multiple
mixed-determinant (Klein, Dansereau, & Hall, levels in a database with a nested structure.
1994) models or theories in the scientific Historically, researchers have used one of
literature. Consistent with the multilevel model two statistical approaches to testing conceptual
driving the GLOBE project, we developed models that have variables at different levels
scales that measured these constructs at the of analysis or data that has a nested structure
appropriate level of analysis (Hanges & (Bryk & Raudenbush, 1987; Hofmann, 1997;
Dickson, Chapter 8). Clearly, the statistical Hofmann et al., 2000). The first traditional
technique used to test the GLOBE conceptual statistical approach, which we will refer to as
model and hypotheses requires that it is able to the aggregated approach, involved averaging
handle relationships operating at multiple levels both the independent and dependent variables to
of analysis. some higher level of analysis (e.g., scale values
Beyond the multilevel nature of the con- from individuals from the same organization
ceptual model, scales, and hypotheses, the are averaged to obtain an organizational mean
structure of the GLOBE database produced score). Both of these variables would be aggre-
several constraints that need to be considered gated even if only one truly operated at the
when choosing a statistical methodology. To group level. Statistical analyses were then per-
be more specific, the responses in the GLOBE formed on the aggregated data and the degrees
database came from middle managers emplo- of freedom associated with these analyses
yed in one of three industries (i.e., the financial, were a function of the number of groups sam-
food service, and telecommunications indus- pled (Hofmann, 1997). The second traditional
tries) throughout the world. The primary sam- approach, which we will refer to as the disag-
pling strategy used by the GLOBE country gregated approach, required assigning the aver-
co-investigators was to gain access to multiple age group score on the higher-level construct
organizations from at least two of these three (e.g., organizational culture) to each individual
industries in each society and to distribute the group member. Statistical analyses were then
GLOBE questionnaires to as many middle man- performed using the entire database (i.e., all
agers in each organization as possible. middle managers) (James & Williams, 2000;
This sampling strategy produced what is Kreft & Leeuw, 1998).
called a nested structure in the GLOBE data- Both of these approaches have limitations
base. In other words, the middle managers are for testing the GLOBE multilevel conceptual
nested within organizations because multiple model. For example, the aggregated approach
middle managers were sampled from each cannot assess models in which more than one
organization. Further, organizations are nested level of analysis is operating (Hofmann, 1997),
within societies because multiple organizations whereas the disaggregated approach has been
were sampled from each society. The conse- criticized for ignoring the nested structure of
quence of having this nested structure is that data (Goldstein, 1995; Hox, 1995) and is most
it produces covariation among the responses appropriately used if the dependent variable in
of middle managers. This covariation among the analysis operates at the individual level of
respondents is not surprising given that we were analysis (James & Williams, 2000). Because a
asking middle managers from the same organi- conceptual model that has multiple independent
zation and society questions about their organi- and dependent variables that operate at different
zation’s and society’s culture. Covariation levels of analysis is being tested in Project
among respondents creates problems for tradi- GLOBE, neither of these two traditional
tional statistical analyses, such as ordinary least approaches are optimal approaches for testing
squares regression analyses, because indepen- our hypotheses.
dence of observations is a critical assumption One statistical technique that has been dis-
for these procedures (Bryk & Raudenbush, cussed in the organizational literature as an effec-
1987; Hofmann, Griffin, & Gavin, 2000). The tive tool for analyzing multilevel conceptual
statistical technique used in Project GLOBE to models and nested data is hierarchical linear
analyze the conceptual model thus had to be modeling (HLM) (Hofmann, 1997; Hofmann
11-House.qxd 3/19/2004 4:15 PM Page 223
et al., 2000). Although HLM is somewhat new is the unadjusted mean participative leadership
to the organizational literature, it actually has rating for organization j in society k and eijk
been used in other fields (such as agricultural represents random level variation in res-
and physical sciences) for quite some time, ponses among the middle managers. As
although it is typically not referred to as HLM in shown in this equation, there are no predictors
these fields. HLM has been referred to as multi- of individual-level variation on participative
level linear models in the sociological research leadership ratings.
(Goldstein, 1995), mixed effects and random The second step of HLM considers organiza-
effects models in the biometrics literature, random tional-level variation in the dependent variable.
coefficient regression models in the economet- In contrast to the individual level, our hypothe-
rics literature, and as covariance components sis specified two predictors of organizational-
models in the statistical literature (Bryk & level variation. So, we used the intercepts (π0jk)
Raudenbush, 1987). obtained for each organization from the prior
HLM can be thought of as a multistep process step (i.e., unadjusted mean participative leader-
designed to test relationships between indepen- ship ratings for a particular organization) as the
dent and dependent variables at multiple levels. dependent variable and predicted this variable
We will use an example to help explain how we by using the aggregated organizational-level
used this analytic procedure in GLOBE. A pos- Humane Orientation practices (As Is) and values
sible hypothesis that we might want to test is (Should Be) scales. More specifically:
whether middle managers in high humane orien-
tation organizations and societies share the π0jk = β00k + β01k (ORG_AS_IS)jk
perception that being “participative” is an effec-
+ β02k (ORG_SHOULD_BE)jk + r0jk
tive attribute for leaders. Using this example, it
should be noted that there are three levels of (11.4)
variation implicitly specified in this hypothesis.
Differences among multiple middle managers In this equation, π 0jk represents the intercept
from the same organization contribute to indi- obtained from the prior HLM step (Equation
vidual-level variation in the dependent variable. 11.3), β00k represents the unadjusted mean
Differences among the average response across participative leadership rating for society k,
the various organizations represent organizational- “ORG_AS_IS” represents the group-mean
level variation, and finally, differences in the centered2 (organizational score is subtracted
average response across the various societies from the societal average) Humane Orientation
represent societal-level variation. We will con- practices for organization j and “ORG_
sider all three of these levels in our analysis of SHOULD_BE” represents the group-mean-
our data. centered Humane Orientation values score for
Although there is probably some individual- organization j. Further, β01k and β02k represent
level variation in perceptions of middle managers the organizational-level HLM coefficients for
concerning the effectiveness of participative these two predictors in society k, respectively.
leadership, it should be noted that our hypo- If one or both of these HLM coefficients are
thesis did not specify any particular individual- significant, there is support for the hypothesis
level predictors of this variation. Thus, in the that organizational-level culture is predictive of
first step of the HLM analysis, we use the fol- participative leadership. Finally, r0jk represents
lowing equation: organizational-level error.
The third step of the HLM analysis considers
Participativeijk = π0jk + eijk (11.3) societal-level variation in the dependent vari-
able. As with the step 2 of HLM, our hypothesis
In this equation, “Participativeijk” represents specified two predictors of this societal-level
a particular middle manager’s rating on the Par- variation. Thus, we used the societal intercepts
ticipative CLT dimension. This middle manager (β00k) from the prior step (i.e., aggregated societal
comes from organization j, which is embedded participative leadership ratings) as the dependent
in society k. The intercept (π0jk) in this equation variable and predicted societal-level differences
11-House.qxd 3/19/2004 4:15 PM Page 224
the other hand, if the predictor does not have a thought of as performing a hierarchical analysis
meaningful zero point (as is true with most mea- in which the significance of some higher group-
sures used in the social sciences), the researcher level predictor (e.g., societal culture) reflects
may want to transform the predictor in some the unique contribution of this predictor over
way to make the intercept more meaningful. In the contribution of all lower-level predictors
HLM, researchers affect the meaning of the (e.g., organizational culture) (Hofmann, 1997;
intercept by centering their predictors. Hofmann et al., 2000). Thus, the particular center-
There are several options available for cen- ing method used in the HLM analysis is very
tering data in HLM (i.e., grand-mean centering, important.
raw-metric scaling, group-mean centering) In Project GLOBE, we used group-mean
(Bryk & Raudenbush, 1987; Kreft & Leeuw, centering in our HLM analyses. We decided
1998), and each of these options have different to use this centering approach for two reasons.
implications for the nature of the hypotheses First, we were interested in the effects of
being tested by HLM (Bryk & Raudenbush, culture on CLT leadership dimensions at both
1987; Hofmann et al., 2000). The option that the organizational and societal levels of analy-
we eventually selected for GLOBE analysis is sis. Thus, we developed specific hypotheses
to group-mean center the predictors in the concerning the relationship between organiza-
analysis. If predictors are group-mean centered, tional culture and CLT leadership dimensions
the intercepts of the equation in the three levels and societal culture and CLT leadership dimen-
of our HLM example are interpreted as the sions. As indicated above, group-mean centering
unadjusted mean of the dependent variable for allows us to partition the dependent variable’s
organizations, societies, or the entire sample variance to these different levels of analyses.
(Hofmann & Gavin, 1998). Thus, a random Second, and more importantly, our conceptual
intercept HLM analysis with group-mean model indicated that in addition to affecting CLT
centering can be thought of as conceptually leadership dimensions, societal culture also
related to the traditional aggregated level analy- affected organizational culture. We wanted to
sis because group-level variables are used to follow the traditional logic used in data analytic
predict dependent variable group means (i.e., procedures in which the significance of exoge-
intercepts). In addition, group-mean centering nous variables (i.e., variables whose causes are
divides the dependent variable’s variance into not established in the model being tested) are
three separate portions (i.e., that due to individ- first established and then the influence of the
ual-, organizational-, and societal-level varia- exogenous variables are controlled for before
tion) (Hofmann et al., 2000). Thus, group-mean testing the significance of any endogenous vari-
centering enables researchers to a priori specify ables (i.e., variables whose causes are outlined in
the level of analysis at which a specific predic- the model being tested) (Pedhazur & Schmelkin,
tor should influence the dependent variable (e.g., 1991). Although group-mean centering does not
organizational culture predictors should account perform this statistical control, we decided that
for organizational variation in leadership scales) it would be inappropriate to use grand-mean
(Hofmann & Gavin, 1998). centering because use of that centering approach
In addition to group-mean centered, pre- would control for an endogenous variable in
dictors can also be grand-mean centered. With the GLOBE model (i.e., organizational culture)
grand-mean centering, the average score on some before testing the influence of an exogenous
predictor for the entire database is removed variable (i.e., societal culture). Considering all
from all scores on that predictor. With grand- factors together we therefore used group-mean
mean centering, the value of the intercept for centering in our analyses.
a particular group is interpreted as the mean for
that group after the influence of the predictors Issue 2: Assessment
included in the equation has been removed (i.e.,
of Predictor Importance
the intercept is now interpreted as the adjusted
group mean). In addition, with grand-mean After determining that a predictor exhibits
centering, the HLM analyses can be basically a significant relationship with the dependent
11-House.qxd 3/19/2004 4:15 PM Page 226
R2
variable, researchers typically want to assess the interpretation of R2 is slightly more complex.
strength of the connection between these two Figure 11.2 presents a Venn diagram illustrating
variables. In regression analysis, the usual way of how to interpret R2 when a HLM analysis is
assessing the contribution of a predictor to an conducted. The biggest circle in this figure
equation is to calculate R2 or the “amount of vari- represents the dependent variable’s variance.
ance explained” or shared in the dependent vari- Continuing with the example used earlier in this
able by a particular predictor or set of predictors. chapter, the dependent variable is the Participa-
Figure 11.1 shows a Venn diagram illustrating R2 tive leadership CLT scale and, as shown in
as it is typically conceptualized in ordinary least Figure 11.2, the variance of this variable cuts
squares regression with one predictor (also across three levels of analyses (i.e., individual,
known as simple regression). The circle on the organizational, and societal). In Figure 11.2, it
left of this figure represents the variance of a pre- appears that approximately 30% of the depen-
dictor whereas the circle on the right represents dent variable’s variance is at the individual level,
the variance of the dependent variable. The over- 50% is at the organizational level, and 20% is at
lapping portion of these two circles represents the societal level.
the relationship between these two variables. As As outlined in Equation 11.3 and discussed
the magnitude of the relationship increases, the earlier in this chapter, the GLOBE model does
amount of overlap of these two circles increases. not specify any individual-level predictors of the
The percentage of the dependent variable vari- CLT leadership dimensions. This is because we
ance contained in the overlapping region, labeled were primarily interested in the shared aspects of
R2, is interpreted as the percentage of dependent CLT profiles in Project GLOBE and not differ-
variable variance explained by the predictor. ences among individuals within organizations
Similar to simple regression analysis, R2 and societies. Indeed, as discussed by Hanges
can also be estimated when performing HLM and Dickson in Chapter 8 the constructs under-
analyses. However, because HLM assesses the lying the CLT leadership scales were conceptu-
predictability of the dependent variable across alized as convergent–emergent constructs. Thus,
different levels of analysis, the calculation and the individual-level portion of the dependent
11-House.qxd 3/19/2004 4:15 PM Page 227
Variance of
Dependent
Variance of Societal Level
Variable
Humane Cultural Practices
Societal
R2Societal
Variance of Organizational
Level Humane Cultural
Level of Analysis
Organizational
Practices
R2Organizational
Individual
variable’s variance can be thought of as error organizational Participative CLT ratings (see
variance caused by random as well as system- Equation 11.4). The large oval in Figure 11.2 rep-
atic differences among respondents. Though resents the variance associated with this organiza-
individual-level predictors could have been tional-level predictor, and the region of this circle
identified (see the discussion of gender differ- that overlaps with the dependent variable is the R2
ences by Emrich, Denmark, and Den Hartog in for this organizational-level predictor. It is impor-
Chapter 14), the GLOBE hypotheses outlined tant to note that the variance of this predictor only
by House and Javidan in Chapter 2 were not exists at the organizational level of analysis. Thus,
concerned with individual-level variation. the overlapping portion of the predictor and
Whereas this level had to be included in the dependent variable (labeled R2 Organizational)
HLM analyses so that unbiased significance only occurs at the organizational level of analysis.
tests can be obtained for the GLOBE hypothe- This indicates that R2-organizational is a level-
ses regarding the organizational and societal specific measure in that it only indicates the
predictors (e.g., Bliese & Hanges, 2002; Kenny percentage of organizational-level variance in
& Judd, 1986), the individual level of analysis the Participative CLT leadership rating explained
will not be included in any estimates of by the Humane Orientation cultural practices
explained variance in the present book. scale. In Figure 11.2, the circle for Humane
At the organizational level, we hypothesized Orientation organizational cultural practices over-
that Humane Orientation organizational cultural laps with approximately 60 percent of the vari-
practices would predict differences in average ance of the dependent variable that exists at the
11-House.qxd 3/19/2004 4:15 PM Page 228
these products across all levels of analysis. leadership dimensions. The six higher-order CLT
More concisely, dimensions were used as dependent variables
and the nine cultural practices and nine cultural
values measures were used as independent vari-
2
RTotal = 2
Rlevel (ICClevel) (11.7)
ables in these analyses. We tested these hypo-
theses by performing two kinds of analyses.
Continuing with the example in Figure 11.2, First, we ran a dimension-specific HLM analy-
Humane Orientation cultural practices accounts sis in which the culture scales for only one
for 46% of the total variance in Participative dimension were used to predict each CLT lead-
CLT ratings (i.e., R2-societal (.80)* ICCsocietal ership dimension (e.g., organizational and soci-
(.20) + R2-organizational (.60)* ICCorganizational etal Humane Orientation cultural practices and
(.50) = .46). In this example, Humane Orienta- values scales predicting Participative CLT
tion cultural practices are clearly a nontrivial dimension ratings). This type of analysis pro-
predictor of Participative CLT ratings. In sum- vides useful information about the connection
mary, examining both the amount of explained between each cultural dimension and the six
variance at a particular level of analysis as well CLT dimensions. The results of these analyses
as the amount of total explained variance is a are discussed in the chapters focusing on the
useful strategy to minimize the chances of mis- cultural dimensions in Part IV.
leading oneself about the importance of the Second, we also tested our hypotheses using
HLM results (Klein & Kozlowski, 2000; Lance a more competitive approach that included all
& James, 1999). The HLMs reported in Part IV hypothesized organizational and societal cul-
primarily report the R2 total for each analysis. tural predictors in our HLM analyses to assess
In summary, we have discussed HLM as which cultural predictors, if any, uniquely pre-
well as our rationale for using group-mean cen- dicted each CLT dimension. It is possible that
tering. We also reviewed how to interpret the some of the findings obtained in the dimension-
importance of predictors in random intercept specific analyses will not be supported in the
models. In the next section of this chapter, we competitive analyses because of the interrela-
specify the exact HLM models that were used tionships among the GLOBE cultural dimen-
to test the GLOBE conceptual model. There are sions. Thus, the competitive analyses can be
two separate aspects of the conceptual model thought of as producing a parsimonious model
that were tested via HLM. First, HLM was used connecting the various cultural dimensions with
to test the hypotheses connecting the nine cul- the CLT scales. The results of the competitive
ture dimensions with the six CLT leadership analyses are discussed in Chapter 21.
dimensions. Second, HLM was used to test the Regardless of whether the GLOBE hypo-
hypotheses concerning the effect of the society theses were being tested in a dimension-specific
and industry variables on organizational cul- or competitive manner, all of these HLM analy-
ture. As noted at the beginning of this chapter, ses can be technically described as random-
we do not discuss the results of any analyses in intercept models in which three levels (i.e.,
this chapter. Rather, the purpose of this discus- individual, organizational, and societal) of
sion is to provide a framework to help the analysis were specified and the cultural pre-
reader interpret the empirical results discussed dictors were group-mean centered. As specified
in Part IV. by House and Hanges in Chapter 6, different
middle managers completed the two GLOBE
questionnaires. We used Form Alpha responses
Specification of Random Intercept to obtain the organizational culture and CLT
data for our analyses.
HLM models in Project GLOBE:
It should be noted that we only included
Test of Culture-CLT Hypotheses data from organizations that had 7 or more
We conducted a series of HLM analyses to respondents. This cutoff was chosen so that we
test the hypotheses concerning the effect of had an average of 45 middle managers respond-
organizational and societal culture on the CLT ing to the organizational culture scales per
11-House.qxd 3/19/2004 4:15 PM Page 230
as a standard that we compared to our next two tests. Two-tailed statistical tests are useful for
models. testing hypothesized as well as detecting unex-
In the second model, we tested whether the pected statistical relationships while maintain-
effect of industry on organizational cultural prac- ing a desired error-rate level for these statistical
tices interacted with some organizational charac- tests. We used two-tailed tests because of the
teristic. We tested for this interaction by treating relative novelty of the CLT leadership construct
the coefficients for the two orthogonally coded and the fact that Project GLOBE is the first
industry variables as random factors at the orga- large-scale, cross-cultural empirical study explor-
nizational level of analysis. In other words, we ing organizational and societal correlates of
allowed the industry coefficients to differ across culturally endorsed beliefs about leadership.
organizations. If the fit of this second model was The only time that a one-tailed test was used
significantly better than the fit of the baseline was when we verified the HLM analyses by
model, then there is evidence that the effect of using the response-bias corrected scales.
industry on organizational cultural practices was With the background in statistical methodol-
moderated by some organizational-level charac- ogy and rationale presented in this chapter, we
teristic. Additional tests would have to be run to trust that the reader will be well prepared to
identify the specific organizational characteristic appropriately and accurately interpret the results
causing the interaction.7 presented by our colleagues in the dimension-
The third and final model that we ran specific chapters that follow.
assessed whether the effect of industry on orga-
nizational cultural practices was moderated by
societal-level factors (e.g., societal culture val- ENDNOTES
ues). This was tested by treating the coefficients
for the two orthogonally coded industry vari- 1. This approach to banding is what Cascio and
ables as random factors at the society level of colleagues (1991) refer to as the “fixed banding”
analysis. If the fit of this third model was approach. There are other approaches (Zedeck et al.,
significantly better than the fit of the baseline 1996) and the reader could even use the SED infor-
model, then there is evidence that the effect of mation provided in Table 11.1 to obtain more detailed
industry on organizational cultural practices banding information than we can provide by using
was moderated by some societal-level charac- the fixed band methodology. For example, the reader
teristic. Additional tests would have to be run to could develop bands around each of the GLOBE
identify the specific societal characteristic caus- societies (i.e., society-specific bands) to gain a more
ing the interaction. Results of these analyses refined understanding of how the various societies
are reported in Chapter 20 by Brodbeck and differ from one another. By using the SED reported
colleagues. for each scale (e.g., SED of 0.30 for the Societal
Uncertainty Avoidance cultural values scale) and
choosing a particular confidence level (e.g., 95%
SUMMARY AND CONCLUSION confidence level is associated with a C of 1.96), the
scale’s bandwidth for each dimension can be deter-
In this chapter, we have described the procedure mined (i.e., societal Uncertainty Avoidance cultural
used to rank order the societies participating in values is approximately 0.59). Adding this bandwidth
the GLOBE Project on the nine cultural practices to Greece’s scale score yields the top value for this
and values dimensions. To facilitate interpreta- society’s band (i.e., 5.68). Subtracting this bandwidth
tion of which societies truly differ from each from the average scale score for Greece yields the
other on a given dimension, we incorporated test bottom value for this society’s band (i.e., 4.50).
banding. We then presented our rationale for Comparing this band (i.e., 4.50–5.68) with the other
using hierarchical linear modeling to test the societies in Sully de Luque et al’s (this volume) Table
main GLOBE hypotheses, and described our 19.8 reveals that the value placed on Uncertainty
specific application of HLM to the GLOBE data. Avoidance in Greece is significantly greater than the
Almost all of the statistical analyses reported value placed on Uncertainty Avoidance in either Italy
in this book were performed using two-tailed or Portugal or in any of the societies listed in the
11-House.qxd 3/19/2004 4:15 PM Page 232
in organizations: Foundations, extensions, and Kreft, I., & Leeuw, J. D. (1998). Introducing
new directions (pp. 382–424). San Francisco: multilevel modeling. Thousand Oaks, CA: Sage.
Jossey-Bass. Lance, C. E., & James, L. R. (1999). V 2: A pro-
Kenny, D. A., & Judd, C. M. (1986). Consequences portional variance accounted for index for
of violating the independence assumption in some cross-level and person-situation research
analysis of variance. Psychological Bulletin, designs. Organizational Research Methods,
99, 422–431. 2, 395–418.
Klein, K. J., Dansereau, F., & Hall, R. J. (1994). Pedhazur, E. J., & Schmelkin, L. P. (1991).
Levels issues in theory development, data col- Measurement, design, and analysis: An integrated
lection, and analysis. Academy of Management approach. Hillsdale, NJ: Lawrence Erlbaum.
Review, 19, 195–229. Rousseau, D. M. (1985). Issues of level in organiza-
Klein, K. J., & Kozlowski, S. W. J. (2000). From tional research: Multilevel and cross-level per-
micro to meso: Critical steps in conceptualizing spectives. In L. L. Cummings & B. Staw (Eds.),
and conducting multilevel research. Organiza- Research in organizational behavior (pp. 1-37).
tional Research Methods, 3, 211–236. Greenwich, CT: JAI.
Kozlowski, S. W. J., & Klein, K. J. (2000). A multi- Zedeck, S., Cascio, W. F., Goldstein, I. L., & Outtz, J.
level approach to theory and research in organi- (1996). Sliding bands: An alternative to top-down
zations: Contextual, temporal, and emergent selection. In R. S. Barrett (Ed.), Fair employ-
processes. In K. J. Klein & S. W. J. Kozlowski ment strategies in human resource management
(Eds.), Multilevel theory, research, and methods (pp. 222–234). Westport, CT: Quorum Books.
in organizations (pp. 3–90). San Francisco, CA:
Jossey-Bass.
11-House.qxd 3/19/2004 4:15 PM Page 234
12-House.qxd 3/19/2004 4:27 PM Page 235
PART IV
EMPIRICAL FINDINGS
MANSOUR JAVIDAN
he chapters in this part of the book The cultural findings reported in these
measured a societal culture for several reasons: In the GLOBE project, we performed this
First, it is impossible to identify a truly random statistical correction procedure and correlated
sample in many societies. Second, each culture our original scales with these “response-bias-
consists of a variety of subcultures that may corrected” scales to ascertain the extent to
differ on some aspects and be similar on others. which response bias was a problem with our
Third, strictly speaking, to make generalizable scales. Further, we developed an extension
statements about a phenomenon like culture, of this procedure to help interpret the rank
we need more than one study and one sample. ordering of the societies on the basis of this
Given the paucity of such studies, any cross- correction. This extension of the traditional
cultural researcher needs to be cautious in response bias correction procedure and its
making such claims. ability to identify specific societies that exhibit
The above limitations and constraints make substantial response bias is discussed in
any well-designed cross-cultural research Appendix B. This appendix also includes
program a truly daunting task, but one that is society rankings using adjusted scores corrected
critical. Cross-cultural researchers need to be for cultural response bias.
extremely diligent in their research design and
interpretation of their findings. We at GLOBE
have taken unprecedented steps in terms of our BANDS
research design and, as will be apparent in the
following chapters, have taken a strict approach In addition to the concern about response bias,
in interpreting our findings. The reader will we also need to caution the reader about not
find the wording in these chapters to be delib- over-interpreting differences in the rank order
erate, meticulous, and accurate. We believe of societies. There is a tendency to concentrate
this is warranted to reflect the care in the on the relative position of the societies and
research program and to communicate our ignore the scale values used to generate these
findings clearly. rankings. Thus, all differences in the relative
position of societies are interpreted as meaning-
ful even though the magnitude of the societal
RESPONSE BIAS difference with regard to the actual scale values
is sometimes trivial. In an attempt to address
Each dimension chapter provides the rank order this problem, Hanges, Dickson, and Sipe in
of the 62 GLOBE societies in terms of their Chapter 11 used a technique discussed in the
societal cultural practices and values scale psychometric and personnel selection literature
scores. The rank ordering of the 62 societies on known as test banding. This procedure groups
the basis of the six CLT dimension scores is pre- societal scores into bands in which the scores
sented in Chapter 21 by Dorfman, Hanges, and within a particular band are considered as being
Brodbeck. When interpreting the rank order of not meaningfully different. Thus, this banding
societies, it is important to note that people from methodology identifies a range of scores that
different cultures sometimes exhibit different cannot be distinguished from the top score in a
response patterns when completing question- band.
naires. The presence of these culturally based
response patterns is believed to bias subsequent
cross-cultural comparisons based on self-report QUARTETS
data because these response patterns are not
a function of the intended construct of interest. Separate culture items were written for the nine
A statistical standardization correction proce- core GLOBE dimensions at both the societal
dure has been developed in the cross-cultural and the organizational levels. In addition to
literature that is believed to remove cultural- these two levels of analysis, we wrote separate
response biases from the original “uncorrected” culture items to reflect cultural practices (As Is
questionnaire responses (van de Vijver & response format) versus values (Should Be
Leung, 1997). response format). The items were written as
12-House.qxd 3/19/2004 4:27 PM Page 237
“quartets” having isomorphic structures across intervals associated with these correlations.
the two levels of analysis (societal and organi- The confidence intervals provide the best
zational) and across the two culture manifesta- estimate of the population correlation between
tions (practices and values). The basic structure two variables. To help give the reader a feel for
of the items comprising quartets is identical, but the magnitude of our estimation accuracy,
the frame of reference is varied according to the Appendix D gives confidence intervals for
particular cultural manifestation and levels of various correlations.
analysis being assessed. The specific way that
these are measured is explained in the dimen-
sion chapters. REFERENCES
12
PERFORMANCE ORIENTATION
MANSOUR JAVIDAN
239
12-House.qxd 3/19/2004 4:27 PM Page 240
every day make some progress, however small. and hard work led to the evolution of the spirit
Therefore, let us not cease to strive, that we may of modern and rational capitalism. McClelland
be incessantly advancing in the way of the Lord, proposed a more detailed process in which
nor let us despair on account of smallness of our the Protestant values led to greater emphasis
success; for however our success may not corre- among parents on teaching their children to
spond to our wishes, yet our labor is not lost, have high standards and to value independence.
when this day surpasses the preceding one; pro- Such emphasis in turn resulted in higher need
vided that with sincere simplicity we keep our end for achievement among sons, who then became
in view, and press forward to the goal, not prac- active entrepreneurs and generated higher rates
ticing self-adulation, nor indulging our own evil of economic growth. Winterbottom (1953)
propensities, but perpetually exerting our endeav- showed that mothers of 8- to 11-year-old boys
ors after increasing degrees of amelioration, till with high nAch tended to set high performance
we shall have arrived at a perfection of goodness, expectations for them at an earlier age. Rosen
which indeed, we seek and pursue as long as and D’Andrade (1959) also showed that parents
we live. . . . (Calvin, I, pp. 775–776, excerpt in of sons with high nAch tended to set higher
McClelland, 1961, p. 49) standards for them than those of the sons with
lower need for achievement. Other data show
Weber thus chronicled a historical evolution that Protestant parents tend to put more empha-
of religious and social views toward work, from sis on the importance of knowledge, indepen-
the medieval notion that work in this world is a dence, self-reliance, planning, and achieving
distraction from the godly life to the Protestant (McClelland, Rindlisbacher, & deCharms,
ethic, which saw work as a fundamental require- 1955; Rosen, 1959; Veroff, Atkinson, Feld, &
ment for being convinced of being among the Gruin, 1960).
elect and for satisfying God’s requirements. The Several studies have compared the motives for
most fundamental consequence of this change having children in a variety of countries (Bulatao,
is the increased emphasis on improving the 1979; Darroch, Meyer, & Singarimbun, 1981;
social and human lot, and greater attention to Fawcett, 1983; Hoffman, 1987; Kagitçibasi,
performance improvement as a religious calling. 1982). They have identified two sets of values
The societies that practiced Calvinist doctrine attributed to having children: economic and psy-
viewed their religious mandate as one of con- chological. In many less developed countries,
stantly striving to improve their economic and children are seen as the source of support for
social conditions. To them, performance orien- parents in old age. They are expected to be obe-
tation was a gift and a duty from God. dient and to take care of their parents when they
Several decades later, David C. McClelland, are in need. Iranians refer to their children as the
the noted American psychologist, and his col- “cane for the hand in the old age.” In contrast,
leagues (1955, 1958, 1961) introduced the con- American and German parents dismiss such a
cept of need for achievement, which is defined notion.
as the need to do better all the time (1987, In societies in which support for the parents
p. 228). He argued that individuals with high is a paramount virtue, obedience rather than
need for achievement (nAch) tend to achieve autonomy is highly valued. Family-interest
pleasure from progressive improvement, like rather than self-interest of the child is the pri-
to work on tasks with moderate probabilities ority. Close parental supervision is a sign of
of success because they represent a challenge, love and affection rather than unwelcome
take personal responsibility for their actions, intrusion. Such family surroundings and prac-
seek frequent feedback, search for information tices are not conducive for development of the
on how to do things better, and are generally child’s sense of self-identity and self-confi-
innovative. dence. Children’s identity is part of that of the
McClelland (1961) set out to propose a psy- family. They are socialized to be followers and
chological perspective on Weber’s chronicle of to accept the wishes of those in positions of
the Protestant Reformation. Weber’s hypothesis authority, such as parents and older siblings.
was that the Protestant values of self-reliance Self-reliance, self-confidence, innovativeness,
12-House.qxd 3/19/2004 4:27 PM Page 241
and the will to achieve new things and break to those of Weber’s (1904/1930, 1904/1998)
tradition are frowned on. in the sense that cultural values of perfor-
To McClelland, the critical outcome of the mance orientation exist more strongly among
Protestant doctrine and the spirit of reform was some peoples than others, and are a critical
its impact on individuals’ need for achievement. force in shaping and influencing their social
Parents became supportive of the children’s and economic behavior.
self-confidence and self-reliance and encour- Further evidence for a broader interpretation
aged them to set higher standards for their suc- of the Protestant ethic was found by Fyans,
cess and to work hard toward achieving their Salili, Maehr, and Desai (1983), who explored
goals. Social and economic success then pro- the possibility that there may be a universal def-
vided the feedback necessary to institutionalize inition of achievement. On the basis of Duda’s
higher levels of nAch at the family and societal (1980, 1981) argument for a cross-cultural per-
level: spective in defining achievement, they used
Maehr’s (1974) theoretical definition, which
If people had to set standards for their behavior consists of personal responsibility, standards of
based on their personal revelations rather than on excellence, and challenge. They factor analyzed
church authority, they tended to be more achieve- the data collected by Osgood, Miron, and May
ment oriented and more interested in personally (1975) from 15- to 18-year-old male students
figuring out what was best for them to do. from 30 different language communities.
(McClelland, 1987, p. 259) The authors identified a universal factor of
achievement consisting of such concepts as
The Protestant ethic is not the only religion knowledge, progress, masculinity, success,
to foster hard work and worldly performance. In work, freedom, and courage. They concluded
a study of cultural values in many countries, that there is a universally accepted definition of
Hofstede and Bond (1988) supported Kahn’s the concept of achievement. Their results led
(1979) hypothesis that one of the main reasons them to the conclusion, similar to McClelland’s
for the impressive economic growth and pros- (1961), that the Protestant ethic is a special
perity of Southeast Asian countries during the case of a universal achievement ethic, focusing
period 1965–1985 was the Confucian or neo- on individual responsibility, hard work, knowl-
Confucian cultural roots of the region. edge, and challenge. But they also pointed
Kong Fu Zu, who was renamed Confucius by out that despite such universality in definition,
the Jesuit missionaries, was an influential civil cultures are quite different in terms of the
servant and social philosopher in China around way they actually manifest the concept. They
500 B.C. He preached many rules and procedures showed a wide range in the 30 different
for daily life in his country. Among the key language communities’ scores on this factor.
principles of the Confucian teachings were his They found that high-scoring cultures tend
emphasis on hard work, acquisition of new to focus on the future, achievement, taking
skills, patience and perseverance, and thrift initiative, and independent competence. The
(Hofstede and Bond, 1988). These values are low-scoring cultures tend to focus on tradition,
very similar to the Puritan principles enunciated family, affiliation, and social ties. Other
by Benjamin Franklin, which heavily influenced authors have shown that some societies have
Max Weber’s writings. very strong paternalistic values, which may be
Kahn (1979) and Hofstede and Bond in conflict with performance orientation
(1988) argued that Confucian principles of (James, Chen, & Cropanzano, 1996; Kanungo
perseverance, working hard, and learning & Aycan, 1997). In such societies as Turkey,
new skills have been instrumental in helping Pakistan, and Taiwan, people in positions
shape and drive the economic progress in the of authority are expected to act like parents and
Southeast Asian region. So, although they to take care of their employees and their
explored a different part of the world, a dif- families. They are expected to have a holistic
ferent time period, and a different religious view of employees rather than a narrow, task-
environment, they reached conclusions similar based view.
12-House.qxd 3/19/2004 4:27 PM Page 242
have people report to bosses who are younger There is a common trend among the vast majority
than they are. An extreme and unfortunate case of societies, both traditional and modern, as to
of such a culture was excerpted from the the distribution of gender roles apart from procre-
Economist by Schneider and Barsoux: ation: Men must be more concerned with eco-
nomic and other achievements and women must
. . . an American oil company set up a drilling be more concerned with taking care of people in
operation on a Pacific island and hired local general and children in particular. (2001, p. 280)
labor. Within a week, all the foremen were found
lined up on the floor, their throats cut. Only after- Hofstede (2001) interpreted his masculinity
wards did they understand that hiring younger dimension as embodying such attributes as
men as foremen to boss older workers was not challenge and job recognition, advancement,
acceptable in a society where age indicates status. importance of money, importance of students’
(1997, p. 9) performance at school, stress on equity, and
the ideal value of performance. However, his
The two types of culture also differ in terms interpretation of the dimension also included
of their perspective on feedback and evaluation. two other distinct attributes, namely the extent
Those in achievement cultures tend to desire of assertiveness and gender egalitarianism. As
and value feedback because it helps them find a result, Hofstede’s notion of performance
out how well they are doing (McClelland, 1961; or achievement orientation is intermingled
Schneider & Barsoux, 1997). On the other hand, with other cultural attributes and is not directly
those in ascribing cultures tend to frown measured.
on the notion of evaluation because they see it
as an evaluation of who the person is rather than
of how he or she is doing. As one French man- COMPARISON OF HIGH VERSUS LOW
ager explained, “The French get offended by PERFORMANCE-ORIENTED SOCIETIES
positive or negative feedback. If you question
my job, you are questioning my honor, my Culture is a set of basic and shared practices and
value, and my very being” (Schneider & values that help human communities find solu-
Barsoux, 1997, p. 141). tions to problems of external adaptation—how
Trompenaars and Hampden-Turner (1998) to survive—and internal integration—how to
produced empirical support for the existence stay together (Schein, 1992). Performance ori-
of the two types of cultures. In a study of entation is an important dimension of a commu-
30,000 respondents, mostly managers, from nity’s culture. It relates to the issues of both
47 countries, they showed that people from external adaptation and internal integration. It
countries like the United States, Norway, and is an internally consistent set of practices and
Australia believe in getting the job done even at values that have an impact on the way a society
the expense of individual freedom, and that defines success in adapting to external chal-
respect does not depend on family background. lenges, and the way the society manages inter-
In contrast, the respondents from Oman and relationships among its people.
Argentina prefer individual freedom even at The medieval Catholic view toward external
the expense of getting the job done, and believe adaptation was one of total submission. Nature,
in family background as the key determinant of represented by God’s emissaries and monas-
social respect. teries, was in full control. The masses could
As mentioned earlier in this chapter, the not survive in a religious sense without the
seminal cross-cultural study by Hofstede (1980, overpowering presence of the Church. The
2001) did not directly assess performance orien- Protestant Reformation ushered in an era of
tation, but one of his cultural dimensions inter- self-confidence and power among the masses.
sects with our GLOBE construct. He viewed Not only were they believed to have the power
performance or achievement orientation as a to survive and succeed, but they were also
part of a broader cultural dimension that he required to stand on their own and to do their
called masculinity/femininity. As he described it, best.
12-House.qxd 3/19/2004 4:27 PM Page 244
Performance Orientation
Trompenaars and Hampden-Turner (1998)
and Internal Integration
used Rotter’s (1966) scale to identify the extent
to which societies varied in terms of their inter- Human communities over time tend to
nal or external locus of control. In their survey develop a series of assumptions, practices, and
of 30,000 respondents, they found that people in values that influence the nature of interrelation-
several Arab countries see no value in attempt- ships within the community (Schein, 1992).
ing to control natural forces. They also found These assumptions help determine desirable and
that most of the people from Norway, Israel, and undesirable behavior among the members and
the United States believe that what happens to thus help sustain the community over time.
12-House.qxd 3/19/2004 4:27 PM Page 245
Table 12.1 Higher Performance Orientation Societies Versus Lower Performance Orientation Societies
strong culture of performance orientation and a that constitute a second As Is measure (see
society with a weak performance orientation. Table 12.4 for a sample item). The questions
One needs to be cognizant, however, that soci- relate to the extent to which organizations facili-
etal culture is far too complex to be presented tate and reward improved performance, and the
in black and white. Extreme cases are presented extent to which individuals set challenging goals
here to help explain the concept, but it should for themselves.
also be clear that most cultures do not neatly fit All questionnaire items, including those
into the extremes in any typology stereotype. presented as sample items in Tables 12.2
Although the table shows cultural attributes through 12.5, met exacting statistical stan-
that tend to cluster together, it does not rule dards as described by Hanges and Dickson
out the fact that not all these attributes go in Chapter 8. The construct validity of these
together at all times. Societies usually have scales was confirmed across all GLOBE
differing mixes of the extreme cases presented societies.
in the table. The previously described scales assessed
societal and organizational practices (As Is);
two additional scales were developed to assess
GLOBE MEASURES each respondent’s values (Should Be). Tables
OF PERFORMANCE ORIENTATION 12.3 and 12.5 show questionnaire items that
measure participants’ expressed views on the
Four GLOBE Performance Orientation scales way things “Should Be” in their societies and
were developed to assess this construct. Two organizations. From here on we refer to Should
scales reflect Performance Orientation prac- Be measures as values in contrast to As Is mea-
tices and two scales reflect Performance sures, which indicate practices.
Orientation values. As with all culture dimen- To sum up, the GLOBE scales measure the
sion scales in the GLOBE project, practices extent to which a society or an organization
and values are measured for societies and orga- encourages the practice of rewarding perfor-
nizations within societies. The participants’ mance improvement and setting challenging
assessment of the extent to which a society goals, and the extent to which the respondents
engages in Performance Orientation practices value these practices.
is one of our As Is measures. This scale is a It may be useful at this point to review the
composite of three items that assess societal conceptualization and measurement of perfor-
cultural practices (see Table 12.2 for a sample mance orientation by Trompenaars (1993). In a
item). The questions relate to the society’s cur- survey of 15,000 managers in 53 countries, he
rent practices regarding innovation, improve- conceptualized performance orientation as a
ment, and reward systems. In essence, they contrast to social relations. He asked each
measure the extent to which a society is per- respondent to select one of the following two
ceived to encourage and reward performance statements:
improvement.
Similarly, four questionnaire items were used 1. A company is a system designed to perform
to measure existing practices in organizations functions and tasks in an efficient way.
In this society, students are encouraged to strive for continuously improved performance. (reverse
scored)
Strongly Neither agree nor Strongly
agree disagree disagree
1 2 3 4 5 6 7
12-House.qxd 3/19/2004 4:27 PM Page 247
I believe that teen-aged students should be encouraged to strive for continuously improved performance.
(reverse scored)
Strongly Neither agree nor Strongly
agree disagree disagree
1 2 3 4 5 6 7
In this organization, employees are encouraged to strive for continuously improved performance.
(reverse scored)
Strongly Neither agree nor Strongly
agree disagree disagree
1 2 3 4 5 6 7
In this organization, employees should be encouraged to strive for continuously improved performance.
(reverse scored)
Strongly Neither agree nor Strongly
agree disagree disagree
1 2 3 4 5 6 7
People are hired to fulfill these functions with really are, even if you do not get things done.
the help of machines and other equipment. (1 = strongly agree; 5 = strongly disagree)
They are paid for the tasks they perform.
2. The respect a person gets is highly dependent
(Performance)
on his or her family background. (1 = strongly
2. A company is a large group of people working agree; 5 = strongly disagree)
together. The people have social relations with
other people and with the organization. The The first item reflects the two extreme points
functioning is dependent upon these relations. of a societal value that the appropriate behavior is
(Social relations) one that truly represents the individual’s freedom
versus the one that produces results. The second
Trompenaars and Hampden-Turner (1998) question deals with a particular way that a society
measured the extent to which each society ascribes value to its members, namely their
uses performance or other criteria for granting family background. To the extent that the respon-
status by the percentage of respondents who dis- dent disagrees with either of these questions, he
agreed with the following two statements: or she holds a performance-oriented perspective.
GLOBE’s measures are different from these
1. The most important thing in life is to think in the sense that they focus on a community’s
and act in the ways that best suit the way you values and practices on whether or not innovation
12-House.qxd 3/19/2004 4:27 PM Page 248
Standard
Variable Mean Deviation Minimum Maximum Valid N
and improvement should be rewarded. The one has received little attention from other cross-
GLOBE item that is similar to Trompenaars and cultural authors.
Hampden-Turner’s (1998) is an item that mea- Table 12.6 shows that there is a substantial
sures the extent to which rewards should be and significant difference between the average
based on performance or other factors. scores for the current perceived level of societal
To summarize, the GLOBE scales of Performance Orientation and the desired level.
Performance Orientation seem to be the only Respondents’ aspirations about how much their
direct measure of this concept in terms of societies should focus on performance are far
practices and values. They clearly and explicitly beyond their perceptions of the level of their
contrast societies and organizations on the societies’ current practices. They tend to be
extent to which they facilitate and reward their rather harsh in their assessment of how much
people for wanting to meet higher standards and Performance Orientation is practiced in their
achieve higher goals. countries as compared to how much Perfor-
mance Orientation should be practiced. For
instance, the highest score on practices is for
GLOBE FINDINGS Switzerland at 4.94 out of a range of 1 to 7,
ON PERFORMANCE ORIENTATION which is not very high. In contrast, the highest
score for values is 6.58, for El Salvador.
Table 12.6 shows the grand means of society Two possible explanations are plausible.
practices and society values scales for Perfor- First, there might be a human need to belong to
mance Orientation across all GLOBE societies. a high performance-oriented and successful
Compared to the other cultural dimensions society. It may be satisfying for humans, no
measured, the society practices score has a matter what culture they are from, to be associ-
midlevel average rating of 4.10 with a range of ated with success and achievement. This is
3.20 to 4.94. reflected in the fact that societies with the high-
The society values average score for est Performance Orientation values scores are
Performance Orientation, at 5.94 with a range from all parts of the world. They represent Asia,
of 4.92 to 6.58 is the highest of all GLOBE Europe, South America, North America, and
scales. Compared with the other GLOBE Africa. It may be a basic human need to excel
cultural dimensions of Uncertainty Avoidance, and to succeed, and to have high performance
Future Orientation, Power Distance, Humane standards (McClelland, 1961). Another possible
Orientation, Collectivism, and Assertiveness, explanation is the issue of social desirability in
Performance Orientation is the most cherished. responding to these questions. This point has
The middle manager responses indicate that been addressed in Chapters 8 and 9.
people from all over the world are seeking a
society that strongly encourages and rewards
Overall Scores and Industry Scores
innovation, challenging goals, and improve-
ment. This is a particularly important finding In each society, data were collected from
because despite such importance, the concept at least one of the following industries: Food
12-House.qxd 3/19/2004 4:27 PM Page 249
* Sample sizes are different because not all three industries were selected in each country.
** p < .01 (2-tailed).
processing, finance, and telecommunications. practices and values rather than those driven by
Table 12.7 shows the correlation between the the forces in a specific industry.
overall societal practices score for each society
and the scores for the participating industries. The Correlation Between
As seen in this table, all correlations are
Overall Practices and Values Scores
significant.
The values scores are also highly correlated, The society scores on values scales have a
as shown in Table 12.8. These findings increase modest, but significant −0.28 ( p < 0.05 N = 61)
our confidence in the generalizability of the correlation with the scores on practices scales. In
results in the sense that they are not industry other words, the two constructs are minimally
specific. They represent three very different related such that people’s aspirations are not
industries with varied histories and dynamics. strongly related to their current assessments.
What they do share is their national heritage. The lack of a strong relationship is borne
Our findings seem to represent general societal out by Tables 12.9 and 12.10, which show the
12-House.qxd 3/19/2004 4:27 PM Page 250
Band
A B C
NOTES: Our response bias correction procedure identified * Higher scores indicate greater performance orientation.
response bias in some countries for this scale (see endnotes).
a South Africa (Black sample)
b Canada (English-speaking)
c Switzerland (French-speaking)
d Germany (West): Former FRG
e South Africa (White sample)
f Germany (East): Former GDR
individual country scores on each scale (a with high practices scores, the Philippines has a
detailed explanation of the scores and the bands high values score, Taiwan has a low values
is provided in Chapter 8 by Hanges and score, and New Zealand has an average values
Dickson).1 For example, among the societies score. Canada and the United States have
12-House.qxd 3/19/2004 4:27 PM Page 251
Band
A B C D E
Country Score Country Score Country Score Country Score Country Score
El Salvador 6.58 South Africaa 6.23 Switzerland 5.82 Russia 5.54 Japan 5.17
Zimbabwe 6.45 Mexico 6.16 Greece 5.81 Netherlands 5.49 South Africaf 4.92
Colombia 6.42 Canadab 6.15 Spain 5.80 Kazakhstan 5.41
Slovenia 6.41 Guatemala 6.14 Sweden 5.80 Turkey 5.39
Namibia 6.40 USA 6.14 Morocco 5.76 South Korea 5.25
Portugal 6.40 Brazil 6.13 Israel 5.75
Venezuela 6.35 Poland 6.12 Thailand 5.74
Argentina 6.35 Finland 6.11 Taiwan 5.74
Ecuador 6.32 Austria 6.10 Indonesia 5.73
Philippines 6.31 Germanyc 6.09 Singapore 5.72
Nigeria 6.27 Iran 6.08 Georgia 5.69
Zambia 6.24 Italy 6.07 China 5.67
Bolivia 6.05 France 5.65
India 6.05 Hong Kong 5.64
Malaysia 6.04 Albania 5.63
Kuwait 6.03 Denmark 5.61
Germanyd 6.01
Switzerlande 5.98
Ireland 5.98
Qatar 5.96
Hungary 5.96
New Zealand 5.90
Costa Rica 5.90
Egypt 5.90
England 5.90
Australia 5.89
NOTES: Our response bias correction procedure identified * Higher scores indicate greater performance orientation.
response bias in some countries for this scale (see endnotes).
a South Africa (White sample)
b Canada (English-speaking)
c Germany (East): Former GDR
d Germany (West): Former FRG
e Switzerland (French-speaking)
f South Africa (Black sample)
midlevel values scores but represent high prac- The weak correlation between practices and
tices scores. South Africa (Black sample) has a values measures is exhibited across all three
relatively low values score, but a high practices industries. As can be seen from Table 12.11, no
score. This finding further supports the argu- discernable relationship exists between the par-
ment, made earlier, that it is a fundamental ticipants’ assessment of their society’s current
human attribute to desire a highly performance- level of Performance Orientation and their
oriented society independent of the current level views on its value across industries. However,
of societal practices. there is a small negative relationship between
12-House.qxd 3/19/2004 6:23 PM Page 252
Table 12.11 Performance Orientation Industry Correlation Matrix-Society Values and Society Practices
Food Finance
Overall Industry Industry Telecommunication
Score— Score— Score— Industry Score—
Industry Score Practices Practices Practices Practices
overall practices and values scores (r = −.308, more performance-oriented societies enjoy
p < .05). healthier economies? Will societies with higher
aspirations for Performance Orientation exhibit
higher economic prosperity? You may recall
that in the first section of Chapter 7, we pro-
GLOBE PERFORMANCE posed a series of hypotheses. The following are
ORIENTATION AND OTHER those relating to Performance Orientation.
ECONOMIC AND SOCIAL INDICATORS Societies that score high on performance orien-
tation practices tend to:
In this section, we compare and contrast the
GLOBE findings on Performance Orientation a. Be more economically prosperous and compet-
with those of other major cross-cultural and itively successful
comparative studies focusing on different
countries’ social and economic performance. b. Have stronger societal support for competiti-
As explained in Chapter 7 by Javidan and veness
Hauser, four major ongoing reports produce c. Enjoy higher levels of human development
the relevant data: The IMD’s Global Competi-
tiveness Ranking, the United Nations’ Human Table 12.12 provides the answers to these
Development Report, and the World Values questions. It shows the correlation coefficients
Survey. The findings of these studies are between the two GLOBE dimensions of
grouped into four categories: Economic health, Performance Orientation practices and values,
success in science and technology, human con- and the various elements of economic health. As
dition, and societal values. explained in Chapter 7, these are economic
prosperity, which refers to consumption and
growth; economic productivity, which refers to a
Performance Orientation
supportive labor environment and growth in
and Economic Health
productivity; government support for prosper-
What is the relationship between Perform- ity, reflecting the extent to which the govern-
ance Orientation and economic health? Do the ment and the political body are supportive of
12-House.qxd 3/19/2004 6:23 PM Page 253
economic progress; and finally, societal support considering the many sources of error that could
for competitiveness, which is a measure of the affect our measures. A concern for achievement as
general social attitude toward and support for expressed in imaginative literature—folk tales and
business competitiveness. stories for children—is associated in modern times
In addition to these measures, we also exam- with a more rapid rate of economic development.
ined the relationship between GLOBE findings The generalization is confirmed not only for
and the World Competitiveness Index produced Western, free-enterprise democracies like England
by IMD. and the United States but also for Communist
The table shows that respondents’ aspira- countries like Russia, Bulgaria, or Hungary, or
tions in terms of how much a society should primitive tribes that are just beginning to make
value performance is not positively related to contact with modern technological society. It holds
the society’s economic performance. in the main whether the society is developed or
underdeveloped, poor or rich, industrial or agricul-
Economic Productivity, tural, free or totalitarian. In other words, there is a
National Competitiveness, Economic strong suggestion here that men with high achieve-
Prosperity, and GLOBE Findings on ment motives will find a way to economic achieve-
Performance Orientation Practices ment given fairly wide variations in opportunity
and social structure. What people want, they some-
As shown in Table 12.12 above, respondents’
how manage to get, in the main and on the average,
views on their societies’ current level of Perfor-
though . . . other factors can modify the speed with
mance Orientation are significantly related to
which they get it. (p. 105)
their countries’ level of economic prosperity and
national competitiveness. In other words, Hypo- This view is not, however, universally
thesis a is supported—performance-oriented shared. Trompenaars and Hampden-Turner
societies are more economically prosperous. (1998) compared societies in terms of their ori-
These findings are consistent with the notion entation toward ascription or achievement. They
suggested by Weber (1904/1930, 1904/1998), argued that “there is no evidence that either ori-
Rostow (1952), and Parsons (1951): Societal entation belongs to a ‘higher’ level of develop-
practices that encourage achievement orientation ment, as modernization theorists used to claim”
are key to economic and business success. As (p. 110). They also criticized the notion that
McClelland (1961) eloquently stated in his book, Protestant countries are more economically suc-
The Achieving Society: cessful than Catholic countries:
In conclusion, if we look back over the diverse A second glance at the scores shows that there
findings reported in this chapter, they confirm are growing difficulties with the thesis that an
our general hypothesis to a surprising extent, achievement orientation is the key to economic
12-House.qxd 3/19/2004 4:27 PM Page 254
success. In the first place, Protestant cultures A few other researchers have also shown
are no longer growing faster than Catholic or results that do not support the culture-economic
Buddhist ones. Catholic Belgium, for example, performance linkage. In a study of 41 countries,
has a slightly higher GDP per capita than the more Lynn (1991) found no relationship between
Protestant Netherlands. Catholic France and Italy work ethic, achievement motivation, or impro-
have been growing faster than the UK or parts of vement orientation and economic growth.
the Protestant Scandinavia. (p. 110) Furnham, Kirkcaldy, and Lynn (1996) studied a
sample of 42,000 students in 42 countries and
Two points need to be made in response to showed that work ethic, achievement motiva-
the authors’ criticism. First, they do not provide tion, and improvement orientation were nega-
any evidence on the validity of their research tively correlated to gross domestic product and
instruments and there are serious questions as had no significant relationship with economic
to whether specific items match the construct growth. An important point about their study
of interest. For example, their measure of a is that their measures seem to be a combination
society’s achievement orientation is the percent- of values, societal practices, and individual
age of respondents who disagree with the state- practices and, as such, may be confounded.
ment: “The most important thing in life is to think Combining values and practices tends to com-
and act in the ways that best suit the way you plicate the interpretation of findings due to the
really are, even if you do not get things done.” fact that, as we have shown here, they are not
It is not clear at all that, in the respondents’ necessarily related. Furthermore, earlier in this
minds, this item refers to achievement orienta- section we showed that values were not related
tion. It could easily be interpreted to mean indi- to most economic variables.
vidualism and individual freedom. Having said In summary, the economic prosperity and
that, we correlated the authors’ results on this competitiveness of a society is a complex and
item with measures of economic health used by multidimensional concept. It evolves over time as
GLOBE and found strong and significant corre- a result of many forces and drivers. It is strongly
lation with three of the four scales: economic dependent on the nation’s economic productivity
prosperity (correlation = .573, p < .001), public because a paramount goal in any society is to pro-
support for economic prosperity (correlation duce rising standards of living for its citizenry
= .547, p < .001), and societal support for eco- (Porter, 1990). Our findings point to the conclu-
nomic prosperity (correlation = .367, p < .05). sion that cultural practices are an important corre-
So, contrary to their claim about lack of a rela- late of economic productivity and prosperity.
tionship, their purported measure of achievement Along with other variables (Porter, 1990), cul-
orientation is in fact related to economic success. tures with high Performance Orientation practices
The second point related to their criticism scores foster leaders who value and reward hard
is in regard to the authors’ comment about work, ambition, high standards, and performance
the distinction between Catholic and Protestant improvement. They lead to the formation of cor-
countries. As we have stated earlier in this porations that are ambitious, hard working, com-
chapter, the Protestant Reformation may simply petitive, and successful. The success of the private
be a specific case of change in societal values sector, in turn, drives the overall prosperity of the
leading to self-confidence and higher achieve- economy and the population.
ment motivation and performance orientation, but
is certainly not the only one. Hofstede and Bond
Government and Societal Support
(1988) showed that Confucian Dynamism was
for Economic Prosperity and GLOBE
strongly correlated with economic growth among
Findings on Performance Orientation
22 nations between 1965 and 1985. Their descrip-
tion of Confucian Dynamism has much in com- As shown in Table 12.12, Performance
mon with McClelland’s need for achievement and Orientation practices scores are significantly
GLOBE’s Performance Orientation, lending sup- related to the level of societal support for
port for the argument that no particular religion or economic success, confirming Hypothesis b—
philosophy has a monopoly on achievement. Performance oriented societies have stronger
12-House.qxd 3/19/2004 4:27 PM Page 255
social support for competitiveness. The technology and the increasing mobility of the
relationship is significant even after controlling workforce, the role of the state is to facilitate
for gross national product (GNP) per capita. and encourage flexibility and reinvigoration by
Societies that are reported to be more promoting innovation, small business, techno-
performance oriented are associated with more logical development, and competition. Onis
advanced political systems, more flexible (1995) suggested that such state intervention
labor regulations, and higher general levels of policies as subsidized loans, price ceilings, and
economic literacy among the population. Their protectionism are dysfunctional due to the
competition laws are more supportive of busi- state’s inability to manage influence processes
ness competitiveness and their labor-relations and the uncontrollable power of bureaucrats
environments are less hostile. Their public and politicians. Finally, Porter (1990), in his
service is less intrusive and less influenced by study of 10 major trading countries, concluded
political interference. that the political and legal environment in a
For much of the 20th century, the world society plays a critical role in its competitive-
of politics was the battleground for two ness among other nations. He suggested that
fundamentally different political ideologies the government’s proper role is
(Macpherson, 1966): The liberal ideology,
which was based on rationality, self-interest, to encourage—or even push—companies to raise
and property rights; and the socialist ideology, their aspirations to higher levels of competitive
which was most concerned with the treatment performance. . . . Government policies that suc-
of labor and distribution of wealth. The founda- ceed are those that create an environment in which
tion for the liberal ideology is the equity norm, companies can gain competitive advantage. . . .
which focuses on equal opportunity for all to Government has critical responsibilities for funda-
contribute and to gain, and to be rewarded mentals like the primary and secondary education
according to their contributions (Deutsch, systems, basic national infrastructure, and
1975). The foundation for the socialist ideol- research in areas of broad national concern.
ogy is the protection of human well-being. The (pp. 87–88)
two value systems result in two different
approaches to the role of government and the Our findings provide large-scale empirical
nature of distribution of wealth and resources in support for Porter’s and other authors’ argu-
the society. With the collapse of the communist ments, although it is important to remember that
ideology, liberalism has emerged as the win- we showed a significant negative correlation
ning value system for most of the world. between scores of Performance Orientation val-
The picture emerging from our findings is ues and practices at the society level (and no
consistent with the writings of many authors significant relationships at the industry level). In
who have explored the role of government from other words, the societies with low Performance
different perspectives (Hirsch, 1991; Hofstede & Orientation practices scores tend to aspire to
Bond, 1988; Jessop, 1991; McClelland, 1987; higher levels than those that report high prac-
Onis, 1995; Porter, 1990; Swanson, 1967). tices. This is easily understandable and probably
Swanson showed that the Protestant Reformation reflects a universal finding for all GLOBE cul-
was more successful in those European commu- tural dimensions that have socially desirable
nities where governments tended to be more characteristics: the less one has, the more one
democratic and less centralized. McClelland desires.
(1987) suggested that “an open, competitive Our findings also show that governments do
structure in which self-reliance is rewarded, as in not act in a vacuum. Governments operate in a
the democratically ruled cities at the time of the milieu of national cultural values and practices.
Reformation, seems to be an important source of Government policies and actions are derived
encouragement for developing achievement from and are reflections of societal values.
motivation” (p. 461). They, in turn, act to reinforce and sustain those
Jessop (1991) and Hirsch (1991) argued values. The fact that Canadian tax rates are
that due to changes in customer demand and higher than those in the United States is a
12-House.qxd 3/19/2004 6:26 PM Page 256
Human
Performance Societal Human Life General Psychological Development
Orientation Health Health Expectancy Satisfaction Health Index (HDI)
other four remaining dimensions related to economic conditions. Our analysis shows that
human condition. These findings strongly the HDI values are significantly related to eco-
suggest that people in societies whose prac- nomic prosperity, public support for economic
tices stress innovation, accomplishment, and prosperity, and competitiveness rankings. The
improvement tend to enjoy a more civil society. findings can be interpreted in three ways. One
They benefit from the rule of the law and private is to conclude that this relationship is caused
property and a higher quality of life. They only by the fact that gross domestic product per
generally feel happier, healthier, and more satis- capita numbers are included in both variables. A
fied with life. They do not, however, neces- second interpretation is that those countries that
sarily enjoy longer lives. They also do not score improve their literacy rates and life expectancy
any higher on the United Nations Human tend to reap the benefits in terms of a more com-
Development Index. In other words, Hypothesis petitive and more prosperous society. An alter-
c (Performance oriented societies enjoy higher native interpretation is that those societies that
levels of human development) is not confirmed. are competitive and prosperous tend to do a
These findings are consistent with those by better job of taking care of their people.
Furnham et al. (1996), whose study of 42,000 Perhaps a surprising finding is that there is
students in 42 countries showed that work ethic no relationship between economic productivity
and achievement motivation were not signifi- and HDI values. In other words, better living
cantly related to the Human Development Index conditions are not associated with a more pro-
(HDI). ductive workforce. Societal support for produc-
One possible way of understanding these tivity and economic productivity do not seem
results is that cultures with high Performance to be connected to what the society does to help
Orientation practices scores have demanding improve the living conditions of its people.
requirements and expectations of their members.
As such, they are associated with a variety of
Political Ideology
behavioral manifestations and cultural artifacts.
What these findings show is that higher levels of Table 12.14 shows no relationship between
economic and societal health, and a more posi- Performance Orientation practices and the
tive state of mind, are the more visible manifes- extent of socialist thinking in political ideology.
tations of such cultural practices. On the other In contrast, those societies with higher Perfor-
hand, societies that are reported to being perfor- mance Orientation practices scores tend to show
mance oriented are not necessarily associated a more positive view toward democracy and
with longer lives, greater psychological health, prefer a less active role for the government.
or a more humanistic atmosphere, as measured They believe that democracies run well and are
by HDI. It seems that the emphasis on hard work good at maintaining order. They also prefer a
and continuous improvement, although generat- strong role for the private ownership of business
ing a positive state of mind, may in fact take a and a greater reliance on individual rather than
toll in terms of psychological well-being. On the government responsibility for individual well-
other hand, the societies that are reported to be being. It is difficult to reach any causal direc-
less performance oriented and less economically tional conclusions. One possible interpretation
successful may indeed enjoy a more stress-free is that the popularity of democratic systems has
lifestyle that would help improve psychological enabled societies to become performance ori-
well-being and life expectancy. ented. On the other hand, it could be argued that
An interesting issue is the relationship high performance oriented societies have been
between HDI and economic health. The HDI economically successful and have built middle
was created by the United Nations in 1990 on classes that have nurtured democratic views.
the premise that economic prosperity is not an
end in itself, but a means to a more humane
Religion
society. It was designed as a measure of the
extent to which societies have been able to Our analysis of the relationship between
improve life expectancy, adult literacy, and reported Performance Orientation practices and
12-House.qxd 3/19/2004 4:27 PM Page 258
religion showed no significant results. Such lack eternal world does not take away the desire for
of relationship between religiousness and success and achievement. Although medieval
Performance Orientation practices warrants Catholicism may have prevented people from
attention. As we described earlier in this paying attention to worldly affairs, today’s
chapter, religion, in the form of the Protestant Catholic countries may not see worldly success
ethic, may have played an important role in cre- and eternal peace as being at odds or mutually
ating greater interest in worldly success. With exclusive. A possible explanation is that cross-
the passage of time, the connection between cultural communication and contact has created
religiousness and Performance Orientation a new set of external adaptation problems: How
may have diminished. Confucian Dynamism to succeed in the global village. It is possible
(Hofstede & Bond, 1988) or the Protestant ethic that nonreligious societies have learned from
may have had an early impact as a trigger the religious ones to work hard and to set high
toward hard work and accomplishment, but standards of performance just to be able to keep
these values may now be in place without their up in the global arena. So, over time, the more
religious undertone. The explanation for this religious societies and the less religious
may lie in the nature of how cultures evolve. societies may have been converging and becom-
Cultural values come to being in response to ing more similar in their views toward perfor-
particular external adaptation or internal inte- mance orientation, although for very different
gration challenges because they help the human reasons.
community solve such challenges. The commu- To summarize, the following are the key
nity’s success with the cultural values provides findings from the comparison of the GLOBE
the feedback necessary for institutionalization Performance Orientation cultural dimension and
of the cultural values. But over time, the con- those in other published comparative reports.
nection between the initial reason for the cre- Societies that score higher on Performance
ation of the value and the practicing of the value Orientation practices
tends to diminish and fade and the cultural value
becomes “mental programming of the mind” • are economically more successful and globally
(Hofstede, 1980). Further, as GLOBE’s unob- more competitive
trusive measures show, high Performance • enjoy a more positive attitude toward life and
Orientation practices scores tend to be related live in a more civil society
to lack of dominance of a single religion. The • prefer a strong role for private ownership of
societies with stronger reported Performance business
Orientation practices tend to draw from a • prefer individual accountability for their own
diversity of religious ideologies, and are not well-being
dogmatic about religions.
On the other hand, the data also show that In contrast, societies that score higher on
being highly dedicated to religion and the Performance Orientation practices do not
12-House.qxd 3/19/2004 4:27 PM Page 259
• enjoy healthier or longer lives men of genius in the North Sea countries would be
• score higher on the United Nations’ Human more energetic than those of other regions because
Development Index they would enjoy better health. . . . They would be
• have stronger religious practices continually stimulated by their cool, bracing cli-
mate, and would feel like working hard all the year,
Societies that score higher on Performance whereas their southern and eastern colleagues in
Orientation values either hot weather or cold would be subject to peri-
ods of depression which are a regular feature of the
• experience weak economic prosperity less favored parts of Europe. (p. 233)
• are less competitive
• have lower life expectancy Woytinsky and Woytinsky (1953) summa-
rized Huntington’s ideal climate as “mean tem-
In contrast, societies that score higher on perature of approximately 40°F in winter and
Performance Orientation values do not 64°F in summer, relative humidity of about
60% at noon and high enough at night that dew
• enjoy government or societal support for is precipitated, and variability of weather with
economic success frequent but not extreme changes” (p. 29).
• benefit from a civil society with high levels of McClelland (1961) found that climate did in
health and life expectancy fact have an impact on the need for achievement.
• score higher on the United Nations’ Human He found that cultures with high nAch scores
Development Index generally live in relatively dry and moderate cli-
• represent any particular political ideology mates and those with low scores tended to be in
tropical climates. Toynbee’s (1947) explanation
As is clear from the above list, reported was that nations’ success depends on the level of
Performance Orientation values do not show challenge they face. Those societies in which the
strong associations with such other measures challenge from the physical environment is “just
as economic productivity or societal health. This right” tend to rise and succeed. Those that face
is consistent with their lack of strong relationship too small (tropical) or too great (arid) a chal-
with reported Performance Orientation practices. lenge tend to underperform.
The GLOBE clustering of climates was
explained in Chapter 10. We conducted analy-
PERFORMANCE ORIENTATION ses of variance (ANOVAs) tests to determine if
AND SOCIETY DEMOGRAPHICS there exist climatic differences in the reported
practices and values of societal Performance
To explore the impact of societal demographics, Orientation. We found weak evidence of
we examined the relationship between Perfor- climatic differences in the societal practices
mance Orientation and two variables: Physical (F[6, 54 = 2.10], p < .07) and societal values
climate and geographic region. We categorized (F[6, 54 = 2.02], p < .08) concerning Perfor-
the GLOBE societies in terms of their climate. mance Orientation. Further analysis (not shown)
As discussed in Chapter 10 by Gupta and found that climatic differences account for
Hanges, we created seven categories of climate: 25% of the variation in societal Performance
Tropical humid, tropical wet and dry (savanna), Orientation practices scores, whereas 75%
desert, subtropical humid, subtropical wet and represent idiosyncratic societal differences.
dry (Mediterranean), marine west coast, and With respect to societal Performance Orientation
continental. values, 13% of the variation was explained by
Huntington (1924) was one of the first schol- climatic differences, whereas 87% repre-
ars to focus on the relationship between physical sented unique societal differences. In other
climate and achievement: words, climatic differences have a modest
but nontrivial significant effect on societal
Climate influences health and energy, and these in Performance Orientation culture. Table 12.15
turn influence civilization. . . . On an average the shows the scores for each climate cluster.
12-House.qxd 3/19/2004 4:27 PM Page 260
Standard Standard
Climate Mean N Deviation Mean N Deviation
Standard Standard
Climate Mean N Deviation Mean N Deviation
Standard Standard
Geographic Region Mean N Deviation Mean N Deviation
Standard Standard
Geographic Region Mean N Deviation Mean N Deviation
societies differently in terms of how important world want a society that is more interested in
performance is, and in their aspirations about getting better results and improving its perfor-
how important it should be. It is noteworthy that mance, those in different regions and climates
the Confucian Asia cluster is among the highest do have differing aspirations.
scoring groups, lending support to Hofstede and Second, geography and climate also have a
Bond’s (1988) argument for the role of cultural role to play in how people assess their societies’
undertones in building strong performance- current Performance Orientation practices.
oriented societies. People in different geographic and climate
To summarize, our findings lead to three clusters have differing assessments of their
important conclusions: First, respondents’ aspi- societies’ current practices. These differences
rations in terms of Performance Orientation are seem to be associated with the dominant
somewhat dependent on geography and climate. religion, language, and cultural heritage of the
It appears that although people all over the cluster. It appears that geographic proximity
12-House.qxd 3/19/2004 4:27 PM Page 264
Performance Orientation
McClelland Need for
Achievement Society Practices Society Values
Society scores −.12 .04
(N = 27) (N = 27)
The first item reflects the two extreme points From the inconsistent results described
of a societal value that the appropriate behavior above, it appears that the construct of perfor-
is one that truly represents the individual’s pref- mance orientation at a cross-cultural level is
erences versus the one that produces results. neither well defined nor well developed. It has
The second question deals with a particular way been conceptualized by various authors in dif-
that a society ascribes value to its members, ferent ways. It seems that the previous work has
namely their family background. To the extent focused on the concept in contrast to other
that the respondent disagrees with either of things, whereas GLOBE is taking a different
these questions, he or she is purported to hold a approach. GLOBE’s conceptualization of per-
performance-oriented perspective. formance orientation is different from those
We correlated the GLOBE Performance of other researchers in the sense that its mea-
Orientation societal practice scale scores with sures focus on drive for performance improve-
the individual items and the average of the two ment and challenging goals. It examines the
items in Trompenaars and Hampden-Turner’s responses on a continuum of low to high perfor-
study for the 37 countries that were in com- mance orientation. Other researchers have taken
mon. There are no significant correlations with an indirect approach to performance orientation.
GLOBE results. We also correlated the res- Trompenaars and Hampden-Turner (1998)
ponses to the aggregated GLOBE Performance explored it in contrast to being oneself or the
Orientation items with the scores reported by person’s background and family connection.
Trompenaars, Trompenaars and Hampden- Trompenaars (1993) examined it in contrast
Turner, and McClelland. Except for three items, to maintenance of social relations. McClelland,
all the other correlations were insignificant. The (1958, 1961) measured it in terms of projections
three significant correlations are: based on children’s stories. Although none of
these approaches are necessarily superior or
• Trompenaars 3 (The most important thing inferior, the GLOBE measurement approach has
in life is to think and act in the ways that the advantage of relying on informed reports of
best suit the way you really are, even if you a society’s performance orientation or lack of it.
do not get things done) and the GLOBE item, Furthermore, these measures are independently
“In this society, major rewards are based validated by unobtrusive measures.
on only performance effectiveness,” or are
based on “Factors other than performance
effectiveness.” Pearson correlation: 0.32 PERFORMANCE ORIENTATION
( p < .05). AS A DIMENSION OF
• Trompenaars 3 (The most important thing ORGANIZATIONAL CULTURE
in life is to think and act in the ways that
best suit the way you really are, even if you An organization is a group of individuals who
do not get things done) and the GLOBE item, are expected to work together as a community.
“I believe that people should set challenging As such, they need a set of values to help man-
goals for themselves.” Pearson correlation: .39 age the interactions among the members and to
(p < .05). help succeed in dealing with the challenges they
• There is a significant relationship between face (Schwartz & Bilsky, 1990; Schein, 1992).
the GLOBE item (“I believe that people Where do these values come from? Among the
should set challenging goals for themselves”) several potential sources, the societal culture
and Trompenaars’s percentage selecting per- plays an important role for two reasons. First, the
formance (A company is a system designed success of the organization in external adapta-
to perform functions and tasks in an efficient tion requires closeness to the contextual culture.
way. People are hired to fulfill these functions To succeed, the organization needs to assimilate,
with the help of machines and other equip- or at a minimum, respect and appreciate its
ment. They are paid for the tasks they broader environment. Second, the employees
perform). The Pearson correlation is −.534 working in the organization are members of the
( p < .01). society and have been socialized into its values
12-House.qxd 3/19/2004 4:27 PM Page 266
and assumptions. It is only natural to expect (p < .01). Both analyses support a principal
them to bring those same values into their deal- proposition in the GLOBE theoretical model
ings within the organization. (i.e., Proposition 3, Figure 2.1, Chapter 2, by
Kreder and Zeller (1988) compared 27 House & Javidan): organizational cultural values
German and 13 U.S. companies in terms of the and practices are significantly related to societal
extent to which their control mechanisms were values and practices.2 Further, the finding that
task or employee oriented. They found that the societal values are associated with and likely
U.S. companies were more task-oriented, using predictors of organizational values supports the
formalized planning systems and meticulous argument that performance orientation is a
monitoring mechanisms. The German compa- deeply held value related to a fundamental
nies, however, were more employee-oriented, human need for achievement (McClelland,
with greater participation and concern for 1961) and for enhancing one’s self-concept
employees’ personal problems. (Shamir, 1991).
Hofstede, Neuijen, Ohayv, and Sanders
(1990) factor-analyzed 54 items relating to
organizational practices in 20 organizational PERFORMANCE ORIENTED
units of 10 organizations in Denmark and the AS A LEADERSHIP CHARACTERISTIC
Netherlands. They identified six factors, two of
which related to performance orientation. They As explained in Chapter 8 (Hanges & Dickson),
called them results orientation versus process one of the confirmed first-order factors for lead-
orientation, which opposes concern for the ends ership is performance oriented. It refers to the
with that of the means, and results orientation extent a leader stresses continuous improvement
versus employee orientation, which opposes a and has high standards of performance. Table
concern for people to a concern for getting the 12.18 shows the three items that were confirmed
job done. as the components of the first-order leadership
factor called performance orientation.
The construct of performance orientation
THE RELATIONSHIP BETWEEN SOCIETAL has not been directly studied in the leadership
AND ORGANIZATIONAL PRACTICES AND literature. The earliest work related to this
SOCIETAL AND ORGANIZATIONAL VALUES concept was the leadership studies conducted
at Ohio State University in the United States,
As indicated in Chapter 2 by House and Javidan, where a leader’s role was classified as consid-
the GLOBE theoretical model postulates that eration versus initiating structure (Fleishman,
societal practices and values affect organiza- 1953; Halpin & Winer, 1957; Hempill &
tional practices and values. Two hierarchical lin- Coons, 1957). Similarly, a major research pro-
ear models (HLMs) were conducted to test these gram was conducted at the University of
hypotheses for organizational Performance Michigan that characterized leader behavior in
Orientation practices and values. We tested the terms of task-oriented behavior, relations-
GLOBE hypothesis regarding the effect of soci- oriented behavior, and participative leadership
etal culture on organizational culture by con- (Katz & Kahn, 1952; Katz, Maccoby, & Morse,
ducting HLM analyses in which organizational 1950).
GLOBE Performance Orientation was predicted The consideration or relations-oriented
by GLOBE societal Performance Orientation. dimension relates to those behaviors that are
These analyses supported our hypotheses that primarily concerned with building and sus-
societal Performance Orientation practices have taining relationships with people, increasing
a significant and strong positive relationship teamwork, and building identification with the
with organizational Performance Orientation organization (Yukl, 1998). Examples are friend-
practices (p < .01). We found a similar signifi- liness, showing concern, and being interested in
cant and even stronger relationship between the well-being of subordinates.
societal Performance Orientation values and Initiating structure or task-oriented behavior
organizational Performance Orientation values reflects those behaviors that are important in
12-House.qxd 3/19/2004 4:27 PM Page 267
Table 12.18 Performance Orientation as a Leadership Characteristic: Leader Attribute Item Components
1 7
Greatly Contributes
Inhibits 2 3 4 5 6 Greatly to
Leadership Somewhat Slightly No Contributes Contributes Leadership
Effectiveness Inhibits Inhibits Impact Slightly Somewhat Effectiveness
getting the task completed, improving efficiency, dimensions are consistently more effective in
and increasing reliability. Examples are setting Japan.
goals, defining subordinates’ roles, and explain- A more recent genre of leadership theory is
ing the way the work is done. the neo-charismatic school of thought (House &
Although the two dimensions were originally Aditya, 1997), which postulates that effective
thought to be independent, later research has leaders not only combine task-oriented and
shown it to be otherwise. In a meta-analysis employees-oriented behaviors, but they also
of the research conducted after 1968, Fisher show specific skills and mind-set relating to,
and Edwards (1988) showed a clear relation- among other things, setting ambitious goals and
ship between considerate leadership and communicating high expectations of their sub-
employee productivity. They showed that the ordinates (House, 1977). They achieve ambi-
correlation between considerate leadership tious goals by building their subordinates’
and job performance ranged from .27 to .45. self-confidence and by intellectually challeng-
Schriesheim, House, and Kerr (1976) showed ing them (Bass, 1985).
that the correlation between consideration and GLOBE’s empirically derived construct of
initiating structure had a mean of .52, casting leadership Performance Orientation directly
further doubt on the expected independence of relates to this notion of ambition and challenge.
the two concepts. It reflects the leader’s excellence orientation
The interdependence between the two con- and constant pursuit of improvement. It is not
cepts was addressed by Blake and Mouton concerned with task orientation or employee
(1964), who developed the managerial grid and orientation per se. It goes beyond these cate-
proposed that effective leaders are high on both gories and focuses on the extent to which the
concern for people and concern for production. attribute of excellence orientation is seen to lead
Although their model has received little to effectiveness.
research support, Misumi and Peterson (1985), Table 12.19 shows the overall and industry
in a 30-year research program in Japan, identi- averages across all GLOBE societies. On a
fied two types of leadership behavior, which 7-point scale, the average is around 6.00, which
they called the performance-maintenance (PM) is very high, reflecting the participants’ belief
theory of leadership. As stated earlier, they sug- that performance orientation is a highly effec-
gested that the two dimensions are not indepen- tive characteristic for leaders. The universal
dent, but interdependent. The Maintenance view seems to be that those who set high stan-
function facilitates the Performance function. dards and pursue continuous excellence are
They showed that leaders who are high on both effective leaders.
12-House.qxd 3/19/2004 4:27 PM Page 268
Standard
Variable Mean Deviation Minimum Maximum Valid N
The reason for such high demand for not as strong as those in the A band (a detailed
excellence-oriented leaders may be related to explanation of scores and bands is provided in
human motivation and self-concept (Shamir, Chapter 11).
1991). Self-concept is one’s sense of one’s own As shown in Table 12.19, the mean scores
competence, power, achievement, and abilities. on this dimension were almost identical across
Shamir, House, and Arthur (1993) suggested the three industries participating in the study.
that people are intrinsically motivated to Table 12.21 shows the correlation coefficients
enhance their self-concept. Performance- for the scores in the three industries. The scores
oriented leaders elevate their subordinates’ self- are significantly correlated across all three
concept by acting as a role models, convincing industries.
them that the standards are indeed very high
but achievable, showing confidence in them,
supporting and rewarding their accomplish-
ments, giving them autonomy, and intellectually SOCIETY DEMOGRAPHICS AND
challenging them. This argument is supported by LEADERSHIP PERFORMANCE ORIENTATION
Haire et al. (1966), who showed that among
managers in 12 countries, autonomy and self- To explore the impact of society demographics,
actualization are the most-important and yet we examined the relationship between the
least-satisfied need categories. Javidan and GLOBE Performance Oriented leadership
Carl (1998), in a study of a large group of scores and two variables: Physical climate and
Canadian executives, showed that leaders geographic region. We categorized the GLOBE
who set high targets and ambitious goals tend countries in terms of their climate and found no
to elevate their subordinates’ self-confidence significant difference in leadership scores
and self-actualization. across the different climate types (p = .25).
Although there is universal agreement Geographic region showed a significant
across countries and industries on the value of (p < .01) relationship. Clusters explained 44.3%
leader performance orientation, there is still, as of the societal level differences in performance-
shown in Table 12.20 below, a range of views oriented leadership. Table 12.22 shows that
on its importance. Countries like Ecuador rate performance-oriented leadership is considered
this leadership dimension very high with an most effective in the Anglo cluster at 6.33, and
average of 6.64. The lowest score belongs to least effective in the Middle East culture at 5.48.
Qatar at 4.51, which is still positive although Chapter 21 (Dorfman, Hanges, & Brodbeck)
(Text continues on page 272)
12-House.qxd 3/19/2004 4:27 PM Page 269
Band
A B C D
a Canada (English-speaking)
b Germany (East): Former GDR
c Germany (West): Former FRG
d South Africa (White sample)
e Switzerland (French-speaking)
f South Africa (Black sample)
12-House.qxd 3/19/2004 4:27 PM Page 270
Standard
Geographic Region Mean N Deviation
Standard
Geographic Region Mean N Deviation
Morocco
Qatar
Turkey
shows significant relationships between societal please see the next section and Chapter 8 by
clusters and the Charismatic/Value-Based Hanges & Dickson).
culturally endorsed leadership (CLT), in which To summarize our findings, it is clear that
one essential aspect of this CLT is Performance the GLOBE respondents throughout the world
Orientation. see leaders who have high standards and are
determined to seek continuous improvement as
highly effective. This desire is so strong and
THE RELATIONSHIP BETWEEN universal that it is independent of the societies’
Performance Orientation practices. We explain
PERFORMANCE ORIENTATION
such high expectations in terms of the construct
AS A SOCIETAL DIMENSION AND of self-concept (Shamir et al., 1993), which is
PERFORMANCE ORIENTED the mental image that an individual has of
AS A LEADERSHIP CHARACTERISTIC him- or herself. It is one’s sense of one’s own
competence, power, achievement, and abilities.
Table 12.23 shows the relationship between Leaders who are demanding and excellence
the overall society practices and society values oriented elevate their followers’ self-concept.
scores and performance oriented leadership, The experience of working for such leaders
as well as the scores for each industry. The helps build the followers’ self-confidence and
values scores are significantly and positively self-worth because being associated with such
correlated with performance oriented leader- leaders provides constant feedback of real or
ship. The practices scores are not related to perceived success and improvement, thus
performance oriented leadership. We should making the subordinates feel they are capable
remind the reader that performance oriented of achieving better results. To demand high
leadership is the first-order factor for the performance orientation from leaders seems to
second-order factor labeled Charismatic/Value- satisfy one’s need for a positive and enhanced
Based leadership. (For more information, self-concept.
Table 12.23 Performance Orientation as a Societal Dimension and Performance Oriented as a Leadership
Characteristic-Correlation Matrix
Charismatic/Value-Based (S,O)
Cultural Values
Team Oriented (O)
+ Participative (S,O)
• Improving
performance Humane Oriented (S,O)
• Rewarding
performance
• Being innovative
Self-Protective (S)
• Challenging goals
−
Figure 12.1 Performance Orientation Cultural Values as Drivers for CLT Leadership Dimensions
* Only statistically significant relationships are shown (p < .05; see Table 21.10, Chapter 21 by Dorfman et al.). The most
important leadership CLT relationships are in bold (i.e., relationship is significant at both society and organization levels of
analyses or highest HLM coefficient within each level of analysis).
O = Organizational level
S = Societal level
significantly related to the Team-Oriented associated with forces operating at the societal
leadership dimension and explained a total level of analysis.
of 10.4% of the organizational and societal
variance. All of the explained variance was The organizational Performance Orientation
associated with forces operating at the organi- cultural values scores were positively related
zational level of analysis. The organizational (p < .01) to the Participative leadership dimen-
Performance Orientation cultural values scores sion. Participative leadership is more likely to
were positively related (p < .01) to the Team- be a part of a shared leadership belief system
Oriented leadership dimension. Team-Oriented in organizations with high Performance Orien-
leadership is more likely to be a part of a shared tation values scores. The societal Performance
leadership belief system in organizations with Orientation cultural practices scores were posi-
high Performance Orientation values scores. tively related (p < .05) to the Participative lead-
ership dimension. Participative leadership is
• Participative leadership. Performance
more likely to be a part of the shared leadership
Orientation cultural values scores were sig-
belief system in societies with high Performance
nificantly related to the Participative leader-
Orientation values scores.
ship dimension and explained a total of 8.3%
of the organizational and societal variance. • Humane-Oriented leadership. Performance
Approximately 30.2% of this variance was Orientation cultural practices and values scores
associated with forces operating at the organi- were significantly related to the Humane-
zational level of analysis. The remaining por- Oriented leadership dimension and explained
tion of the explained variance (69.8%) was a total of 13.6% of the organizational and
12-House.qxd 3/19/2004 4:27 PM Page 275
societal variance. Approximately 46.5% of this In summary, our results show that Performance
variance was associated with forces operating Orientation cultural practices and values are
at the organizational level of analysis. The significantly related to all six of our CLT dimen-
remaining portion of the explained variance sions. However, some of these relationships were
(53.5%) was associated with forces operating stronger than others. In particular, Charismatic/
at the societal level of analysis. Value-Based leadership, Humane-Oriented lead-
ership, and Team-Oriented leadership exhibited
The organizational Performance Orientation stronger relationships with Performance Orien-
cultural values scores were positively related tation culture than did Autonomous leadership or
(p < .01) to the Humane-Oriented leadership Self-Protective leadership.
dimension. Humane-Oriented leadership is
more likely to be a part of a shared leadership What Does It All Mean?
belief system in organizations with high
Performance Orientation values scores. The Overall, consistent with Project GLOBE’s
societal Performance Orientation cultural prac- conceptual model, our analyses indicate that the
tices scores were positively related (p < .01) to kind of leadership viewed as effective is a
the Humane-Oriented leadership dimension. reflection of the extent to which an organization
Humane-Oriented leadership is more likely to and society practice and value Performance
be a part of the shared leadership belief system Orientation. When organizations were viewed
in societies with high Performance Orientation as valuing Performance Orientation, the CLT
practices scores. endorsed at the organizational level was more
likely to be composed of Charismatic/Value-
• Autonomous leadership. Performance Based, Participative, Humane-Oriented, Team-
Orientation cultural values scores were signifi- Oriented, and Autonomous leadership styles.
cantly related to the Autonomous leadership However, Self-Protective leadership was less
dimension but explained a total of just 1.2% likely to be a component of the CLTs of these
of the organizational and societal variance. All organizations.
of the explained variance was associated with When societies were viewed as valuing
forces operating at the organizational level of Performance Orientation, the CLT endorsed at
analysis. The organizational Performance the societal level was more likely to be composed
Orientation cultural values scores were posi- of Charismatic/Value-Based and Participative
tively related (p < .01) to the Autonomous lead- leadership styles. We also found that societal
ership dimension. Autonomous leadership is Performance Orientation practices were associ-
more likely to be a part of a shared leadership ated with the content of the CLTs. In particular,
belief system in organizations with high those societies that were viewed as having
Performance Orientation values scores. Performance Orientation practices were more
likely to have Humane-Oriented leadership as a
• Self-Protective leadership. Performance component of their CLT.
Orientation cultural practice and values scores By far, there were more significant relation-
were significantly related to the Self-Protective ships with Performance Orientation values and
leadership dimension but explained a total of the CLTs than there were with Performance
just 0.3% of the organizational and societal Orientation practices at both the organizational
variance. All of the explained variance was asso- and societal levels of analysis. This is an impor-
ciated with forces operating at the organizational tant finding. When individuals think about
level of analysis. The organizational Perfor- effective leader behaviors, they are more influ-
mance Orientation cultural values scores were enced by the value they place on the desired
negatively related (p < .01) to the Self-Protective future than their perception of current realities.
leadership dimension. Self-Protective leadership The concept of leadership has an idealistic
is less likely to be a part of a shared leadership undertone based on the leader’s impact on their
belief system in organizations with high self-concept and ambitions (Shamir et al.,
Performance Orientation values scores. 1993). Our results, therefore, suggest that
12-House.qxd 3/19/2004 4:27 PM Page 276
leaders are seen as the society’s instruments for empower individuals to help achieve it; leaders
change. They are seen as the embodiment of the who are willing to trust others and allow them
ideal state of affairs. Our findings also suggest the opportunity to think and decide for them-
how big the challenge facing the leaders is. selves and to influence the leader’s decisions
They have to work under conditions and cul- (Yukl, 1998).
tural practices that are not very conducive or
supportive, but they are assessed to a significant
extent on their ability to fulfill the expectations SUMMARY AND CONCLUSIONS
of their followers.
Charismatic/Value-Based leadership has Performance Orientation is an important cul-
significant associations with Performance tural dimension that has not been sufficiently
Orientation at both organizational and societal examined in past theoretical or empirical
levels of analysis. It is the reported most- research. The few researchers who have dis-
effective attribute of leaders in societies that cussed it have usually explored it in contrast
value performance. Performance-oriented to other societal values such as ascription
societies are in pursuit of excellence. They desire (Parsons & Shils, 1951). GLOBE focused on
innovation, challenge, and ambition. Their Performance Orientation as a distinct cultural
reported effective leader behavior is one that is dimension. We defined it as the extent to which
Charismatic/Value-Based: A visionary, inspira- a human community encourages and rewards
tional, decisive, and performance-oriented setting challenging goals, innovation, and per-
leadership with high integrity and willingness formance improvement. Its importance is due to
to accept self-sacrifice to achieve their vision. its impact on the way the community addresses
House (1977), Burns (1978), and Bass (1985) challenges of external adaptation and internal
have suggested that charismatic leaders tend to integration.
articulate ideological goals that are deeply Societies whose respondents report strong
rooted in values and aspirations of the group Performance Orientation tend to have specific
(Yukl, 1998). They communicate high expecta- distinguishing characteristics. They value
tions and express confidence in their subordi- education and learning, emphasize results, set
nates. Because almost all societies in our study high performance targets, value taking initia-
scored very high on society Performance tive, and prefer explicit and direct communica-
Orientation values, it appears that charismatic tions. In contrast, societies whose respondents
leadership is a universally endorsed instrument report low Performance Orientation are reported
for satisfaction of human ideals. to value social and family relations, loyalty,
Participative leadership also was posi- tradition, and seniority, and use subtle and
tively associated with Performance Orientation indirect language.
cultural values at both the organizational and Societal culture influences and regulates
societal levels. The existing literature on the human behavior in the society. Cultural values
benefits of participative leadership is not and practices help identify socially acceptable
conclusive: “In summary, after 40 years of and unacceptable behavior. Two specific arenas
research on participation, we are left with the for the impact of these values are organizational
conclusion that participative leadership some- culture and effective leadership attributes.
times results in higher satisfaction, effort, and Organizations are a micro version of the society
performance, and at other times it does not” in which they operate. They are populated by
(Yukl, 1998, p. 126). individuals who have grown up in the host
From our results, it appears that despite the culture, and their success in external adapta-
lack of conclusive evidence for its effectiveness, tion and internal integration depends on their
Participative leadership is consistently reported ability to assimilate their broader environment.
to be effective by organizations and societies In this chapter we have demonstrated that
reported to value performance excellence organizations are reported to reflect the culture
highly. Respondents report as effective leaders (practices and values) in the society in which
those who not only provide a vision, but also they are embedded.
12-House.qxd 3/19/2004 4:27 PM Page 277
Not only are organizations reflective of societal ethic, strong self-confidence and self-reliance,
values and practices, societal cultures also influ- and a strong desire for improvement in their
ence the nature of leadership. Leaders have to children. Families have been a fertile ground for
deal with followers who are part of the broader socializing children in the society’s values and
societal and the narrower organizational cultures. practices.
The followers’ criteria for assessing their leaders’ GLOBE’s measures of the construct at the
performance are in all likelihood influenced by societal and organizational levels are different
their societal and organizational practices. from the existing scales in the literature. They
Therefore, although there may be other mediat- do not compare the extent to which a society
ing variables, we hypothesized that Performance values performance as against other variables
Orientation, as a specific dimension of societal such as family connections. Rather, they directly
culture, has an impact on organizational practices assess the extent to which a community is
and values and the society’s effective leadership reported to encourage and reward setting chal-
attributes. We found this to be true. lenging goals, innovation, and performance
At the organizational level, Performance improvement. Not surprisingly, we did not find
Orientation relates to the extent to which the strong relations between GLOBE’s measures and
organization is focused on ambitious and chal- the other measures of performance orientation.
lenging goals and results and is driven by com- Our results show that reports of societal
petition and winning through innovation and values of Performance Orientation are only
performance improvement. At the leadership weakly related to reports of societal practices.
level, Performance Orientation reflects the For the most part, regardless of the extent of
leader’s excellence orientation and constant their current practices, respondents in all
pursuit of improvement. It relates to the extent societies studied report a desire for high stan-
to which leaders set ambitious goals, communi- dards of performance. At the same time, we
cate high expectations for their subordinates, showed that reports of societal values have little
build their subordinates’ self-confidence, and relation with other available data on societies.
intellectually challenge them. GLOBE’s con- There was little connection to the societies’
ceptualization of performance oriented leader- economic and social health. Societies in which
ship is different from the traditional taxonomy Performance Orientation is reported to be
of task orientation versus consideration. How- highly valued do not necessarily enjoy a greater
ever, it is related to leadership theories that economic prosperity, or a public attitude or gov-
stress goal setting and performance improve- ernment sector that encourages more competi-
ment such as those found in the more recent tiveness. They also do not enjoy a higher level
theories of charismatic and transformational of societal, human, or psychological health, and
leadership. their life expectancy is not any higher.
The chapter examined the historical and reli- On the other hand, reported societal
gious roots of the concept of performance orien- Performance Orientation practices are highly
tation. Although the most well-known writing correlated with important measures of econo-
on the topic is Weber’s (1904/1930, 1904/1998) mic health and tend to enjoy higher levels of
analysis of the evolution of the Protestant ethic economic prosperity. They are supported by a
and its distinction from Catholicism, others public attitude that encourages competitiveness,
have shown that the Protestant religion is not the and a public sector that facilitates prosperity.
only one that encourages performance orienta- They are also more competitive in global mar-
tion. Various authors seem to point to the con- kets. However, we also showed that strong
clusion that several religious schools in different Performance Orientation practices are not associ-
parts of the world have emphasized the impor- ated with better human or psychological health,
tance of hard work and achievement. Some have higher life expectancy, or higher ratings on the
even viewed it as the path to salvation. United Nations Human Development Index.
Over time, religious practices have led Geography was shown to have a particularly
parents in many communities to instill high strong impact on the society’s level of reported
standards of performance, a high level of work Performance Orientation practices, but a weaker
12-House.qxd 3/19/2004 4:27 PM Page 278
Bulatao, R. A. (1979). On the nature of the transition Hall, E. T. (1959). The silent language. New York:
in the value of children (Report No. 60-AP). Anchor.
Honolulu: East-West Center Publishers. Hall, E. T. (1960, May/June). The silent language of
Burns, J. M. (1978). Leadership. New York: Harper overseas business. Harvard Business Review,
& Row. 38(3), 87–95.
Campbell, J. (1988). The power of myth. New York: Halpin, A. W., & Winer, B. J. (1957). A factorial
Doubleday. study of the leader behavior descriptions. In
Darroch, R. K., Meyer, P. A., & Singarimbun, M. R. M. Stogdill & A. E. Coons (Eds.), Leader
(1981). Two are not enough: The value of behavior: Its description and measurement
children to Javanese and Sudanese parents (pp. 39–51). Columbus, OH: Ohio State
(Report No. 60-D). Honolulu: East-West Center University, Bureau of Business Research.
Publishers. Hempill, J. K., & Coons, A. E. (1957).
Deutsch, M. (1975). Equity, equality, and need: What Development of the leader behavior descrip-
determines which values will be used as the tion questionnaire. In R. M. Stogdill & A. E.
basis of distributive justice? Journal of Social Coons (Eds.), Leader behavior: Its description
Issues, 31, 137–149. and measurement (pp. 6–33). Columbus, OH:
Duda, J. L. (1980). Achievement motivation among Bureau of Business Research, Ohio State
Navajo students: A conceptual analysis with University.
preliminary data. Ethos, 8, 316–337. Hirsch, J. (1991). From the Fordist to the post Fordist
Duda, J. L. (1981). A cross-cultural analysis of state. In B. Jessop (Ed.), The politics of flexibil-
achievement motivation in sports and the ity: Restructuring state and industry in Britain,
classroom. Unpublished doctoral dissertation, Germany, and Scandinavia (pp. 67–81). Hants,
University of Illinois at Urbana-Champaign. UK: Edward Elgar.
Fawcett, J. T. (1983). Perceptions of the value Hoffman, L. W. (1987). The value of children
of children: Satisfactions and costs. In R. A. to parents and child rearing patterns. In
Bulatao, R. D. Lee, P. E. Hollerbach, & C. Kagitçibasi (Ed.), Growth and progress in
J. Bongarrts (Eds.), Determinants of fertility in cross-cultural psychology (pp. 159–170). Lisse:
developing countries: Vol. I. Washington, DC: Swetz and Zeitlinger.
National Assembly Press. Hofstede, G. (1980). Culture’s consequences:
Fisher, B. M., & Edwards, J. E. (1988). International differences in work-related values.
Consideration, initiating structure and their Beverly Hills, CA: Sage.
relationships with leader effectiveness: A Hofstede, G. (1998). Masculinity and femininity: The
meta-analysis. In Best paper proceedings taboo dimensions of national cultures. Thousand
(pp. 201–205). Anaheim, CA: Academy of Oaks, CA: Sage.
Management. Hofstede, G. (2001). Culture’s consequences: Compa-
Fleishman, E. A. (1953). The description of supervi- ring values, behaviors, institutions, and organiza-
sory behavior. Personnel Psychology, 37, 1–6. tions across nations (2nd ed.). Thousand Oaks,
Furnham, A. (1990). The Protestant work ethic. CA: Sage.
London: Routledge. Hofstede, G., & Bond, M. H. (1988). The Confucian
Furnham, A., Kirkcaldy, B., & Lynn, R. (1996). connection: From cultural roots to economic
National attitudes to competitiveness, money growth. Organization Dynamics, 5–21.
and work amongst young people: First, second, Hofstede, G., Neuijen, B., Ohayv, D. D., & Sanders,
and third world differences. Human Relations, G. (1990). Measuring organizational cultures: A
47, 119–132. qualitative and quantitative study across twenty
Fyans, Jr., L. J., Salili, F., Maehr, M. L., & Desai, K. A. studies. Administrative Science Quarterly, 35,
(1983). A cross-cultural exploration into the 286–316.
meaning of achievement. Journal Of Personality House, R. J. (1977). A 1976 theory of charismatic
and Social Psychology, 44(5), 1000–1013. leadership. In J. G. Hunt & L. L. Larson
Haire, M., Ghiselli, E. E., & Porter, L. W. (1966). (Eds.), Leadership: The cutting edge
Managerial thinking: An international study. (pp. 189–209). Carbondale: Southern Illinois
New York: John Wiley. University Press.
12-House.qxd 3/19/2004 4:27 PM Page 280
House, R. J., & Aditya, R. N. (1997). The social Lance, C. E., & James, L. R. (1999). A propor-
scientific study of leadership: Quo vadis? Jour- tional variance accounted for some cross-level
nal of Management, 23(3), 409–473. and person-situation research designs. Organiza-
Huntington, E. (1924). The character of races. tional Research Methods, 2, 395–418.
New York: Scribners. Laurant, A. (1986). The cross-cultural puzzle of
Inglehart, R., Basanez, M., & Moreno, A. (1998). global human resource management. Human
Human values and beliefs: A cross-cultural Resource Management, 25(1), 91–102.
sourcebook. Ann Arbor: University of Michigan Lynn, R. (1991). The secret of the miracle economy.
Press. London: SAU.
James, K., Chen, D. L., & Cropanzano, R. (1996). Macpherson, C. B. (1966). The real world of democ-
Culture and leadership among Taiwanese and racy. New York: Oxford University Press.
U.S. workers: Do values influence leadership Maehr, M. L. (1974). Culture and achievement moti-
ideals? In M. N. Ruderman, M. W. Hughes- vation. American Psychologist, 29, 887–896.
James, & S. E. Jackson (Eds.), Selected research McClelland, D. C. (1955). Some social consequences
on work team diversity (pp. 33–52). Greensboro, of achievement motivation. In M. R. Jones
NC: Center for Creative Leadership. (Ed.), Nebraska Symposium on Motivation:
Javidan, M., & Carl, D. (1998). Motivational 1955 (pp. 41–64). Lincoln: University of
consequences of charismatic leadership: An Nebraska Press.
empirical investigation. Paper presented at the McClelland, D. C. (1958). Methods of measuring
meeting of the Decision Sciences Institute, Las human motivation. In J. W. Atkinson (Ed.),
Vegas, NV. Motives in fantasy, action, and society (pp. 7–42).
Jessop, B. (1991). The welfare state in the transition Princeton, NJ: Van Nostrand.
from Fordism to post-Fordism. In B. Jessop (Ed.), McClelland, D. C. (1961). The achieving society.
The politics of flexibility: Restructuring state and Princeton, NJ: Van Nostrand.
industry in Britain, Germany, and Scandinavia McClelland, D. C. (1987). Human motivation.
(pp. 82–105). Hants, UK: Edward Elgar. Cambridge, UK: Cambridge University Press.
Kagitçibasi, C. (1982). The changing value of McClelland, D. C., Rindlisbacher, A., & deCharms,
children in Turkey (Report No. 60-E). Honolulu: R. C. (1955). Religious and other sources of
East-West Center Publishers. parental attitudes toward independence training.
Kahn, H. (1979). World economic development: 1979 In D. C. McClelland (Ed.), Studies in motivation
and beyond. Boulder, CO: Groom Helm. (pp. 389–400) New York: Appleton-Century-
Kanungo, R. N., & Aycan, Z. (1997, June). Crofts.
Organizational culture and human resource Misumi, J., & Peterson, M. F. (1985). The
practices from a cross-cultural perspective. performance-maintenance (PM) theory of
Paper presented at the Canadian Psychological leadership: Review of a Japanese research
Association Annual Conference, Toronto. program. Administrative Science Quarterly, 30,
Katz, D., & Kahn, R. L. (1952). Some recent findings 198–223.
in human relations research. In E. Swanson, Omae, K. (1982). The mind of the strategist: The
T. Newcomb, & E. Hartley (Eds.), Readings in art of Japanese business. New York: McGraw-
social psychology. New York: Holt, Rinehart & Hill.
Winston. Onis, Z. (1995). The limits of neo liberalism: Toward
Katz, D., Maccoby, N., & Morse, N. C. (1950). a reformulation of development theory. Journal
Productivity, supervision, and morale in an of Economic Issues, 29(1), 97–119.
office situation. Ann Arbor: University of Osgood, C. E., Miron, M., & May, W. (1975).
Michigan, Institute for Social Research. Cross-cultural universals of affective meaning.
Kluckhohn, F. R., & Strodtbeck, F. L. (1961). Urbana: University of Illinois Press.
Variations in value orientations. New York: Parsons, T. (1951). The social system. Glencoe, IL:
Harper & Row. Free Press.
Kreder, M., & Zeller, M. (1988). Control in German Parsons, T., & Shils, E. A. (1951). Towards a general
and U.S. companies. Management International theory of action. Cambridge, MA: Harvard
Review, 28(3), 58–66. University Press.
12-House.qxd 3/19/2004 4:27 PM Page 281
Porter, M. (1990, March/April). The competitive Schwartz, S. H., & Bilsky, W. (1990). Toward a
advantage of nations. Harvard Business Review, theory of the universal content and structure
73–93. of values: Extensions and cross-cultural
Ronen, S., & Shenkar, O. (1985). Clustering replications. Journal of Personality and Social
countries on attitudinal dimensions: A review Psychology, 58(5), 878–891.
and synthesis. Academy of Management Review, Sirota, D., & Greenwood, M. J. (1971). Under-
10(3), 435–454. standing your overseas work force. Harvard
Rosen, B. C. (1959). Race, ethnicity and the achieve- Business Review, 49(1), 53–60.
ment syndrome. American Sociological Review, Swanson, G. E. (1967). Religion and regime. Ann
24, 47–60. Arbor: University of Michigan Press.
Rosen, B. C., & D’Andrade, R. G. (1959). The psy- Trompenaars, F. (1993). Riding the waves of culture,
chosocial origins of achievement motivation. London: Nicholas Brealey.
Sociometry, 22, 185–218. Trompenaars, F., & Hampden-Turner, C. (1998).
Rostow, W. W. (1952). The process of economic Riding the waves of culture (2nd ed.).
growth. New York: Norton. New York: McGraw-Hill.
Rotter, J. B. (1966). Generalized expectations for Toynbee, A. J. (1947). A study of history (Vols. 1–6,
internal versus external control of reinforce- abridged; D. C. Somervell, Ed.). New York:
ment. Psychological Monograph, 609, 1–28. Oxford University Press.
Shamir, B. (1991). Meaning, self, and motivation Veroff, J., Atkinson, J. W., Feld, S., & Gurin, G.
in organizations. Organization Studies, 12, (1960). The use of thematic apperception to
405–424. assess motivation in a nationwide interview
Shamir, B., House, R. J., & Arthur, M. B. (1993). The study. Psychological Monograph.
motivational effects of charismatic leadership: Weber, M. (1930). The Protestant ethic and the
A self-concept based theory. Organization spirit of capitalism (T. Parsons, Trans.).
Science, 4, 1–17. New York: Scribner. (Original work published
Schein, E. H. (1992). Organizational culture and lead- 1904)
ership (2nd ed.). San Francisco: Jossey-Bass. Weber, M. (1998). The Protestant ethic and the spirit
Schneider, S. C., & Barsoux, J. L. (1997). Managing of capitalism (2nd ed.). Los Angeles: Roxbury.
across cultures, London: Prentice Hall. (Original work published 1904)
Schriesheim, C. A., House, R. J., & Kerr, S. (1976). Winterbottom, M. R. (1953). The relation of child-
Leader initiating structure: A reconciliation of hood training in independence to achievement
discrepant research results and some empirical motivation. Unpublished doctoral dissertation,
tests. Organizational Behavior and Human Ann Arbor, University of Michigan.
Performance, 15, 297–321. Woytinsky, W. S., & Woytinsky, E. S. (1953). World
Schwartz, S. H., & Bilsky, W. (1987). Toward a population and production. New York: Twentieth
universal psychological structure of human Century Fund.
values. Journal of Personality and Social Yukl, G. (1998). Leadership in organizations. Upper
Psychology, 53(3), 550–562. Saddle River, NJ: Prentice Hall.
13-House.qxd 3/19/2004 5:20 PM Page 282
13
FUTURE ORIENTATION
NEAL ASHKANASY
VIPIN GUPTA
MELINDA S. MAYFIELD
EDWIN TREVOR-ROBERTS
AUTHORS’ NOTE: The authors would like to acknowledge the contributions to this chapter of Research
Assistants Louise Earnshaw and Alison Wallace.
282
13-House.qxd 3/19/2004 5:20 PM Page 283
section, we briefly review how the construct accomplished, though the goals varied by the
was represented in ancient civilizations. individual’s social class and aspirations and
Ancient societies tended to use concepts of potential. The student phase included childhood
time shaped by agriculture, such as “planting (spent with parents) and adolescence (spent with
time”—when it was time to plant the crop, or a teacher). As a student, an individual learned
“harvesting time”—when it was time to harvest the responsibilities of living in a community
the crop (Malinowski, 1926–1927). Time was from a teacher. After the end of 21 years, the
generally not used as a force to govern social individual was ready to leave the school and
and personal life. For instance, children were become a householder by marrying and accom-
suckled when they cried rather than at a regu- plishing the goal of acquiring material wealth
lar time. In the ancient Greek civilization, a and nurturing progeny, as well as learning to
Chronus concept of time evolved (Leach, 1961). exercise restraint and moderation. The house-
Chronus implied a significant event in time and holder life stage was informed by what the stu-
possibly derived from the word Cronus, which dent learned during the student phase. During
represented the god of the harvest. Greeks used the hermitage phase, the individual began with-
the concept of Chronus as a point of reference drawing from routine social activities, increas-
to place other less significant events. In modern ingly giving responsibility to the children and
times, the concept of Chronus has been offering support and assistances services.
extended beyond its original reference to the Finally, in the detachment phase, the individual
harvest time, and is evident in discussions that was to learn how to relate to the world in a spirit
refer to the birth of the Christ, two World Wars, of renunciation, thereby realizing the ultimate
the end of the Cold War, and the terrorist attack goal of spiritual communion.
on the United States on September 11, 2001. At In ancient China, orientation toward time and
the individual level, the Chronus concept is future was intertwined with a circular concept of
reflected in the date of one’s birth, graduation, time, which embraced past, present, and future
or marriage. In the Roman period, the week was generations and included a belief that the dead
used as a measure of time, though it referred not can influence succeeding generations (Teather
to a quarter of a lunar month, but to the time & Chow, 2000). Chinese philosophy viewed
between food markets (Sorokin & Merton, times as “either advantageous or disadvanta-
1990). Consequently, Roman society had sev- geous” (Yi, 1750/1998). Tools such as feng shui
eral weeks that did not consist of seven days. were used to identify selectively advantageous
The concept of time appeared to be more times for specified events for a particular indi-
sophisticated in Eastern civilizations. Both India vidual family or community, or for the nation as
and China relied on astrological and astronomi- a whole. Advantageous times, once gone, were
cal time, which was based on the shifts in sea- seen as repeating themselves in the future;
sons, phases of the moon, the position of the therefore, patience was deemed an important
sun, the length of shadows, and the direction virtue for success in life. In this sense, the
of winds. In ancient Vedic India, the life of Chinese concept of time appeared as cyclical,
individuals, at least in principle, was governed with its structure composed of the recurring
by the system of ashrama, based on different cycles of the seasons, of the phases of the moon
episodes or events in life (Ariarajah, 1986). and planets, and the associated rituals in one’s
Individual life was divided into four phases: life (Hu, 1995). Embedded within this cyclical
student, householder, hermitage, and detach- concept of time, however, was a deeper—trans-
ment, each of 21 years (25 years according to formational—worldview of time (Tu, 1991).
some accounts). This division was intended The transformational effects of time were seen
to help individuals achieve all the objectives as an interactive function of several human and
to lead a perfectly satisfied life and to achieve nonhuman factors; current human actions were
spiritual communion by systematically mov- seen to be only one of the factors influencing
ing from learning to working phase, then to realization of goals.
the servicing and meditation phases. In each A more sophisticated concept of time began
phase of life, specific goals were set to be emerging in the Western world only around the
13-House.qxd 3/19/2004 5:20 PM Page 284
17th century when public clocks were put in Bentham (1789/1970), observed a tendency of
most cities and market towns of Western individuals to undervalue future outcomes by
Europe. Though in the early years timekeep- spending time on dysfunctional activities such
ing was imprecise, the advent of the industrial as hedonistic pleasure, and suggested societal
revolution resulted in a new worldview that mechanisms for restraining individual dysfunc-
“required worker discipline if machine and tional action. However, hardcore behaviorists,
man were to be integrated” (Thrift, 1990, p. 114). including Watson (1913) and Skinner (1937),
The growth of railroads was a major turning rejected inclusion of a subjective concept of
point in forging a discipline for precision time- time in the scientific study of behavior. They
keeping and for pacing life according to time. instead focused on the influence of past-learned
Further, concepts of time and future orientation contingencies on current behavior, and showed
merged, as time became associated with the disinterest in possible feedback of imagined or
idea of progress, innovation, and an orientation expected future environments. On the other
toward achieving goals (Teather & Chow, hand, the social psychologists, particularly Kurt
2000). The past, on the other hand, was charac- Lewin (1942), showed their concern about char-
terized as “tradition-bound,” “backward,” and acterizing behavior just in terms of past experi-
“out of date.” The thrust of future orientation ences. They instead suggested classifying time
was to “save” time, which took the properties of in terms of three categories (past orientation,
a commodity and was deemed equivalent to present orientation, and future orientation), and
money (Becker, 1965). laid foundations for the subsequent psychologi-
With the advent of globalization a renewed cal, managerial, and cross-cultural research on
interest in understanding the concept of time future-time orientation.
and its relationship with orientation toward the Fresh philosophical interest in the concept of
future has emerged. Fundamental to this new future orientation as a distinguishing and essen-
perspective was a differentiation between phys- tial characteristic of cultures has emerged during
ical, or linear time, and social, or episodic and recent times. For illustrative purposes, one such
cyclical time. In terms of social time, several force is the concept of spiritual orientation, with
elements with a bearing on future orientation a thrust on spiritual fulfillment, self-expression,
have been identified, such as “enduring” time, and emotional well-being, which is a key element
which is slow and of long duration, or “erratic” of postmodern societies (Inglehart, 1997). Spiri-
time under situations of uncertainty (Gurvitch, tual orientation refers to a sense of purpose
1964). Gurvitch observed that in ancient and being, a sense of future, a sense of a higher
Western societies, enduring time guided the life power guiding and shaping existence (Cross,
of the peasant classes, and erratic time governed 2001). The industrialized, capitalist cultures
the life of the ruling classes. Similarly, in have tended to be short-term and materialistic,
modern societies, “just-in-time” has become which separated the material and the spiritual
particularly relevant for the stable environ- sphere without any integration (Hofstede, 2001).
ments, and “Internet time” for the high velocity However, the new and emerging service
environments. economy is showing a strong concern for virtue,
with a pragmatic integration of morals and
practice, as evidenced by the downfall of the
Philosophical View of Future Orientation
Enron Corporation, the accounting firm Arthur
The significance of time as a central process Andersen, and the WorldCom Corporation. The
in analyzing and understanding aspects of emphasis is more on long-term orientation, with
human nature has a long and controversial a greater integration of the material and the spir-
history among social and physical scientists. itual spheres. As opposed to the exclusive mate-
In Kant’s (1781/1965) philosophy, time con- rialism of the past, the emphasis is on a more
ception is an “innate ability” that colors the inclusive focus on spiritual fulfillment and self-
way people experience the world. Several early expression. The thrust of this new viewpoint is
philosophers, notably John Stuart Mill (1843/ that the spirit and soul are not limited by time
1986), David Hume (1739/1987), and Jeremy and space. Spiritual inclusiveness is fueled by
13-House.qxd 3/19/2004 5:20 PM Page 285
many forces including the aging of the strategies for meeting their future aspirations.
baby-boomer generation, which is beginning to However, they may lack a solid appreciation of
contemplate death, increasing fears of job secu- situational realities because of a neglect of their
rity, and the waning influence of institutions present personal and social relationships and
(such as churches and extended families) that interactions (Keough et al., 1999). In summary,
once provided spiritual support (Brandt, 1995). as a result, future-oriented individuals and cul-
tures have a capacity to enrich their lives and
maintain self-control, whereas present-oriented
Summary
individuals and cultures strive to simplify their
From historical and present, as well as emerg- lives and rely more on others.
ing standpoints, future orientation emerges as a Finally, the concept of past orientation
key factor guiding human behavior. Therefore, it shares considerable similarities with the concept
is important to explore the concept of future ori- of future orientation because individuals and
entation in more detail, and to analyze GLOBE collectives with high past orientation also show
findings related to this concept. a high capacity and willingness to correct their
current behavior. Past-oriented individuals
and collectives do so by recalling and show-
CONCEPT OF FUTURE ORIENTATION ing appreciation for prior learning, memories,
obligations, and traditions (Keough et al.,
Time perspective is a process of differentiating 1999). Therefore, they prefer to sustain their
personal and social experiences into temporal favorable past experiences and seek to plan for
frames, thereby giving order, coherence, and maintaining the status quo if favorable experi-
meaning to those experiences, events, and ences dominate in their past. Similarly, they
objects (Fraisse, 1963; Frank, 1939; Lewin, may prefer not to repeat their unfavorable past
1942). There are three broad categories of experiences, and thus make all-around efforts to
temporal frames: Past, present and future develop and realize a new vision and state of the
(Kluckhohn & Strodtbeck, 1961; Lewin, 1942). future, grounded in some ideals of the past, once
These temporal frames help in encoding, stor- the sequence of events disjoining the present
ing, and recalling not just perceived, but also from the past has transpired. In stable environ-
expected, targeted (goals), contingent, and ments, past orientation may result in repetition
imagined experiences and scenarios (Keough, of past behaviors in the present, as well as in the
Zimbardo, & Boyd, 1999). Some individuals future, and the schism with the concept of future
and collectives tend to use specific temporal orientation is empirically easier to identify.
frames extensively, whereas others use them More generally, past-oriented individuals and
very sparingly. Thus, cultures become differen- collectives believe that plans should be evalu-
tiated in their future orientation. ated in terms of their fit with the customs and
Cultures with low future orientation, or high traditions of society and that innovation and
present orientation, show the capability to change are legitimated only according to past
enjoy the moment and be spontaneous. They experience. By contrast, future-oriented individ-
are free of past worries or future anxieties, but uals and collectives seek to evaluate plans pri-
at the same time may seek hedonistic pleasures. marily in terms of anticipated future benefits,
They may show incapacity or unwillingness to going beyond their traditions.
plan a sequence to realize their desired goals, Based on the above, the GLOBE definition
and may not appreciate the warning signals that of Future Orientation is the extent to which
their current behavior negatively influences members of a society or an organization believe
realization of their goals in the future (Keough that their current actions will influence their
et al., 1999). future, focus on investment in their future,
In contrast, cultures with high future orienta- believe that they will have a future that matters,
tion have a strong capability and willingness to believe in planning for developing their future,
imagine future contingencies, formulate future and look far into the future for assessing the
goal states, and seek to achieve goals and develop effects of their current actions.
13-House.qxd 3/19/2004 5:20 PM Page 286
Next we review the construct of future in the degree of how strongly each prefered
orientation at three levels of analysis—societal, the present. The Spanish-Americans strongly
organizational, and individual—using cross- favored the present over the future, the Navaho
cultural, managerial, and psychological disci- and the Zuni moderately favored the present,
plinary perspectives. Then we will discuss how and Texans and Mormons were indifferent
Future Orientation is measured as part of Project between the present and the future.
GLOBE. The way societies deal with conceptions of
past, present, and future is one of their basic
Societal-Level Future Orientation: value orientations. Past-oriented societies use
the past to anticipate the future; present-oriented
The Cross-Cultural Perspective
societies resolve current problems without regard
Several cross-cultural studies have noted to long-term implications; and future-oriented
the importance of future orientation at the societies focus on the long-term implications
societal level. Here we describe four major of past and present actions (Kluckhohn &
cross-cultural studies on future orientation: Strodtbeck, 1961). Bluedorn (2000) describes
(a) Kluckhohn and Strodtbeck (1961), (b) this as the trichotomy of time (the past, present,
Hofstede and Bond (1988), (c) Trompenaars and future), but adds that some individuals and
and Hampden-Turner (1998), and (d) Hall societies have difficulty in determining where
(1960). Each of these studies uses a different the boundary of the present exists in relation to
operationalization of the construct of future the past and the future.
orientation, and thus highlight multifaceted
dimensions of this construct. Hofstede and Bond Study: Before describing
Hofstede and Bond’s study, it would be useful
Kluckhohn and Strodtbeck Study: In a pio- to understand the background for it. During the
neering cross-cultural study, Kluckhohn and 1970s and 1980s, scholars became increasingly
Strodtbeck (1961) demonstrated that time orien- interested in understanding societal differences
tation is an important value orientation that in future orientation (Agarwal, 1993; Matthews
differentiates cultures. In their landmark work, & Ornauer, 1976). A widespread interest in
Kluckhohn and Strodtbeck studied five com- future orientation emanated from the rise of
munities in the Southwestern United States: Japanese competitive dominance in the interna-
Spanish-American, Mormon, Texan, Zuni, and tional markets and the apparent shift in the locus
Rimrock Navaho. They found that the three of economic growth and monetary leadership to
dimensions generated from time orientation the Orient. Japanese development was followed
(past-present, past-future, and present-future) by the rapid growth of several East Asian
were effective in differentiating between the nations, particularly the four dragons—Hong
cultures. The past-present dimension divided Kong, Singapore, South Korea, and Taiwan.
the five communities into two distinct groups. During the 1980s, development extended
The first, made up of Navaho and Zuni, had no beyond Japan and the four dragons, to encom-
special preference for past or present orienta- pass Southeast Asian countries, especially
tion; the other (Spanish-American, Texan, and Malaysia, Thailand, and Indonesia. Though eth-
Mormon) strongly favored the present time nically a minority in the community, immigrant
position over the past. The past-future dimen- Chinese families controlled most of the business
sion showed four major divisions. The Zuni and economic operations in these Southeast
(who were moderately inclined toward the past), Asian countries. Scholars identified the mate-
the Navaho (who were indifferent between the rial success of overseas Chinese businesses
past and future), and the Spanish-Americans with their reliance on the Confucian ethic. The
(who moderately favored the future) formed Confucian ethic was worldly in its prescriptions,
three distinct groups. The fourth group contained focused on the quality of relationships that a
the Texans and the Mormons, who had a strong person should try to maintain, and the social and
preference for the future. Finally, the present- civic duties of a person for living a healthy
future dimension distinguished the groups only social life and creating a healthy civic society.
13-House.qxd 3/19/2004 5:20 PM Page 287
In the meantime, opening and liberalization saw (r = .11), and was correlated only with the
China emerging as the fastest growing nation on percentage of students identifying “persistence”
the international landscape. Scholars postulated as a real trait (r = 0.61; p < .05). Data on the stu-
that a high savings rate in East Asia and Japan dents came from Williams, Satterwhite, and
was actually a product of the Confucian ethic, Saiz’s (1998) study, and included 12 overlap-
which promoted a long-term orientation. ping nations.
Hofstede and Bond (1988) developed a mea- Nevertheless, Franke, Hofstede, and Bond
sure of Confucian Dynamism, and showed that (1991) postulated that although Confucian
their measure of Confucian Dynamism was Dynamism encompasses a wider set of values
positively correlated with the economic growth than future orientation alone, long-term orienta-
rate of nations. Predictably, East Asian and tion lies at its heart. Hofstede and Bond (1988)
Japanese societies scored high on the Confucian reported that Confucian Dynamism explained
Dynamism Index, whereas Anglo societies differences in national economic growth rates
scored low. These findings were subsequently between countries, whereas Takyi-Asiedu
replicated by Ralston, Gustafson, Elsass, (1993) emphasized that Confucian Dynamism
Cheung, and Terpstra (1992) in a study of prac- is linked to entrepreneurship. Hofstede and
ticing managers in the United States, Hong Bond (1988, p. 16) concluded that Confucian
Kong, and the People’s Republic of China. In Dynamism is, in essence, a “dynamic, future
Hofstede and Bond’s study some non- oriented mentality,” as opposed to a static or
Confucian countries, such as Brazil and India, tradition-oriented mind-set. More recently,
also scored fairly high on the Confucian Hofstede, Van Deusen, Mueller, Charles, and
Dynamism Index. The Business Goals Network (2002) reported
Hofstede (2001, p. 355) observed that the that business tycoons in Anglo, Latin American,
Confucian Dynamism Index captures only some and Latin European countries are perceived by
aspects of the Confucian ethic and excludes oth- MBA students from those nations as focused
ers, such as filial piety, and therefore needs to be more on “this year’s profits.” In contrast, busi-
relabeled. He emphasized that the core feature ness tycoons in India, China, and Hong Kong
of the Confucian Dynamism Index is its assess- tend to put greater priority on “profits ten years
ment of a society’s persistence with the value of from now.” The Confucian Dynamism Index
thrift; therefore, in his view, the index was showed a multiple correlation of 0.62 with “this
really a measure of long-term orientation, not of year’s profits” (negative) and “profits ten years
Confucian ethic. To support his assertion, from now” (positive).
Hofstede (2001) reported that the index is posi- On the basis of the available evidence,
tively correlated with the percentage of respon- Hofstede (2001) reinterpreted the meaning of his
dents in a society selecting “thrift” as a quality Confucian Work Dynamic construct and labeled
that children can be encouraged to learn at home it Long-Term Orientation. He reported that the
(r = .70, p < .01). On the other hand, the index Confucian Dynamism Index is positively corre-
was negatively correlated with the percentage of lated with Read’s (1993) measure of “marginal
respondents rating “leisure time” as important propensity to save,” defined as the change in real
in their life. The data for these two correlations per capita gross domestic saving from 1970 to
came from the World Values Survey (Inglehart, 1990, in percentages of the sum of the changes
Basanez, & Moreno, 1998). in private consumption plus domestic saving
However, only 13 nations overlapped between over the same period. Therefore, in Hofstede’s
the Confucian Dynamism Index scores sample view, a long-term orientation is critical for a high
and the World Values Survey sample; therefore, propensity to save, whereas a short-term orienta-
the correlations are not creditable by themselves. tion results in a propensity to spend and enjoy
Perhaps, not surprisingly, Hofstede (2001) also leisure in the present.
found and reported that the Confucian Dyna- However, although a high rate of savings has
mism Index did not show any correlation with the indeed been a distinguishing feature of the eco-
percentage of students in a society identifying nomic growth model of Confucian societies, it
“thrifty” as a real psychological trait of a person does not imply that the high savings rate in these
13-House.qxd 3/19/2004 5:20 PM Page 288
societies arises from long-term orientation. A food to survive. Thus, historically, neighbors
high savings rate may, for instance, be were distant, leading Americans to value inde-
accounted for by the comparatively high costs pendence and individuality. As a result, American
of consumption in most Confucian societies, organizations emphasized short-term employ-
along with high taxes on consumption and poor ment, rapid evaluation and promotion, and
availability of social security for the elderly, individual responsibility.
which forces people to save a higher proportion Also, some evidence suggests that Confucian
of their incomes. societies may not really be especially long-term
Further, although “persistence” and “thrift” oriented and that Anglo societies may not
positively loaded on the Confucian Dynamism really be that much more short-term oriented.
Index, the index also included other items that The Japanese, for instance, are skeptical of
have little association with the concept of long- anyone’s ability to forecast the future in spe-
term orientation. Other positively loaded items cific terms and believe that 5 years is too long
included “ordering relationships by status and a time for meaningful plans (Lauenstein, 1985).
observing this order” and “having a sense of The Japanese are acutely aware that plans
shame.” In addition, there were four negatively can go awry because of unanticipated events
loaded items: “Personal steadiness and stabil- (Lauenstein, 1985). Moreover, they reject the
ity,” “protecting your face,” “respect for tradi- possibility of developing “sustainable” compet-
tion,” and “reciprocation of greetings, favors, itive advantage, and believe that any advantage
and gifts.” Is it appropriate to identify “respect is necessarily “transient.”
for tradition” as inversely related with long-
term or future orientation? The answer is not Their approach is almost uniformly to identify a
necessarily affirmative. Larwood, Falbe, Kriger, promising area in which to invest, to find a way to
and Meising (1995), in fact, showed that a pos- achieve some temporary advantage on the basis of
itive relationship exists between past and future which to establish themselves in the market, and
orientation (see also Cottle, 1967; Trompenaars then to rely on unremitting effort to make further
& Hampden-Turner, 1998). Future orientation progress. (Lauenstein, 1985, p. 79)
involves taking a long-term perspective that
includes recognition of how history and tradi- Thus, Confucian Dynamism in Japan reflects
tion define opportunities and capabilities for a state of open-ended improvisational readiness,
the future. Low future orientation, on the other as opposed to a positive regard for the future
hand, induces societies to emphasize only their and planning for the future. More generally, the
immediate concerns. Confucian philosophy holds that what goes
In addition, postwar success of Japan and East up will eventually go down: This encourages
Asian societies may be an outcome of factors people to persist with what is working well until
other than a Confucian work ethic. Ouchi (1981), the times are good, and with what is not work-
for instance, emphasized the role of historical ing well until the times turn around. On the
context in shaping Japanese future orientation. other hand, although Anglo societies may lack
Because little of Japan is suited for agriculture, Confucian Dynamism, they still demonstrate
the planting and harvesting of rice can only be agility that is oriented toward the future, which
accomplished with the cooperation of 20 or more is backed by a strong endorsement of visionary
people. Thus, Ouchi asserted, the Japanese have leadership (House, Wright, & Aditya, 1997). In
had to develop the capacity to work together in Anglo societies, the leadership vision offers
harmony, and this accounts for the societal value not just long-term goals, but also the direction
that individual considerations are outweighed through which such goals may be accomplished.
by concerns for group welfare. This allowed With the economic meltdown of Japan and
Japanese organizations to emphasize collective several other Confucian societies over the
responsibility, lifetime employment, and slow 1990s, the issue of whether Anglo cultures
evaluation and promotion processes. are indeed short-term oriented and Confucian
In contrast, the United States is largely cultures are long-term oriented has been
arable: A single family can produce enough debated widely in the literature (Laverty,
13-House.qxd 3/19/2004 5:20 PM Page 289
1996). Recent scholarly work (see, for instance, the propensity of individuals and societies to be
Davis & Steil, 2001) emphasizes that Anglo involved in many tasks at once (also termed
societies tend to have more efficient markets. parallel processing) rather than sequentially
Under efficient markets, one does not expect (sometimes called synchronicity; see Trompenaars
organizations to maximize short-term profits at & Hampden-Turner, 1998). In sequential cul-
the cost of long-term benefits. Therefore, it is tures, time is seen as either a series of sequential
not appropriate to characterize Anglo societies events occurring at regular intervals or as cycli-
as short-term. On the other hand, lower effi- cal with repetitive patterns, both of which can
ciency of markets in most Confucian societies be worked out in advance (Trompenaars &
makes it difficult to validate the commonly Hampden-Turner, 1998). In polychronous cul-
held belief that those societies are more long- tures, time is seen as requiring several activities
term oriented. to be performed in parallel, which is often more
difficult to plan in advance.
Trompenaars and Hampden-Turner Study:
More recently, Trompenaars and Hampden- Hall’s Study: Hall (1960) highlighted how soci-
Turner (1998) presented data, based on a ques- etal orientation toward time influences standards
tionnaire item from Cottle (1967), to show that of temporal precision. The concept of time as
long- versus short-term time horizons differenti- linear and monochronic creates a sense of
ated among 42 national cultures. Time horizon urgency because time is seen as a commodity or
refers to the length of the planning horizon. a resource to be saved, spent, or wasted, as seen
Although their data showed Confucian cultures in North America and Northern Europe (Hall,
to be clustered at the long-term end of the scale 1960). On the other hand, if time is conceived
and Anglo cultures to be at the short-term as expansive, ongoing, and unlimited, the sense
end, their findings were quite different from of urgency is reduced, as is common in Latin
those of Hofstede’s (2001). For example, in America and the Middle East (Hall, 1960). Thus,
Trompenaars and Hampden-Turner’s study in Latin America and the Middle East it is gen-
Pakistan was ranked higher than all the erally acceptable to take more time for complet-
Confucian societies except Hong Kong in terms ing a contract than is scheduled and to arrive
of a long-term time horizon. In Hofstede’s several hours late to a scheduled meeting. In
(2001) study, Pakistan was ranked the least North America and Northern Europe one is
among 23 societies for which the Confucian expected to be on time for meetings and for
Dynamism scores were reported. Trompenaars work. Although the effects of future orientation
and Hampden-Turner measured time horizon as on the sense of urgency have not been analyzed,
an average of the time horizon for past, present, Schneider (1989) predicted that more future-
and future. They reported that societies with a oriented societies would view time with greater
higher time horizon for the past also tend to urgency. This prediction remains to be tested.
have a higher time horizon for the present and
for the future. The concept of time horizon, Organizational-Level Future
however, should not be confused with the con- Orientation: The Managerial Perspective
cept of future orientation, though they appear to
be similar on the surface. In many cases, a long- Future orientation is the fundamental decision
term time horizon may be consistent with a variable for all organizations because it repre-
lack of future orientation when the long term is sents the problem of allocation of resources over
just a ballooning of a fundamentally short-range time, or what is termed the problem of intertem-
vision. Das (1987) emphasizes that executives poral choice (Laverty, 1996). Organizational
with shorter future orientations often merely managers have to decide, for instance, whether
extend their short-term visualizations to the to exploit prior capabilities or explore new prod-
longer term instead of actually being able to uct and market opportunities even if they are dis-
perceive a distant time horizon. ruptive to their current performance; whether to
Another important dimension of time is poly- invest in incremental learning and improvements
chronicity (Bluedorn, 2000). Polychronicity is or in radical experimentations and innovations;
13-House.qxd 3/19/2004 5:20 PM Page 290
and whether to train their workforce and Cultural attitudes toward uncertainty have a
preventively maintain their machinery. In all significant influence on the organization’s
these instances, organizations choose between future orientation. In societies that find it diffi-
maximizing their current profits, which may cult to deal with uncertainty, organizations
suboptimize future profits, and developing a tend to focus on short-term, operational plans.
balance that would enhance overall profitability Horovitz (1980) found that French firms focus
over both the short and long terms. more on shorter-term plans in comparison to
Numerous researchers such as Bluedorn British firms, whereas German firms lie in
(2000), Hofstede (2001), Lim and Seers (1993), between. Interestingly, France scores higher on
Ouchi (1981), Quinn and McGrath (1985), Hofstede’s uncertainty avoidance index than
Schriber and Gutek (1987), and Trompenaars the British, whereas Germany lies in between.
and Hampden-Turner (1998) have emphasized However, Hofstede’s measure of uncertainty
the importance of the future orientation con- avoidance lacks face validity and may well be
struct for organizations. Future orientation in a measure of stress rather than uncertainty
an organizational setting involves preparing avoidance.
the organization to meet future environmental The future orientation of organizational
changes (Brommer & De La Porte, 1992) and managers is also related to their flexibility
is, therefore, an essential leadership attribute. (Tendam, 1987). Flexibility of managers, in
Indeed, future orientation is the basis of formal turn, contributes to the organization’s flexi-
organizational strategy that, by definition, bility. Tendam suggests that flexibility is the
involves the planned future actions of a firm ability of an organization to adapt to changing
(Trompenaars & Hampden-Turner, 1998). environments and to pursue new goals. An
Schriber and Gutek (1987) posit that the inflexible organization that is unable to adapt
norms of time in organizations affect the sched- to the changing environment will not be likely
uling and coordination of people and tasks, how to make above-normal profits or even survive.
organizations treat stakeholders, and how they Tendam states that three aspects of an organi-
view the firm’s strategy for the future. Webber zation’s culture determine its flexibility: A
(1972), in discussing the relation between the strong vitality, a future orientation, and a
past and future, suggested that the concern sense of mission and direction. Thus, a future-
of organizational managers with the past should oriented organization has the ability to be flex-
be in the service of a better and more fruitful ible and respond to today’s competitive and
future. Managers should be concerned about high-tech environment. One measure of flexi-
where the organization is going because nothing bility or agility is speed. Onken (1999) demon-
is more important than the organization’s future. strated that the constructs of polychronic
In this respect, Das (1986) underlined the impor- culture (which connotes a sense of urgency
tance of future orientation among executives in and decisiveness in performing multiple tasks)
relation to the strategic planning needed to pre- and speed are correlated in an organizational
pare their organizations for continued growth in context (r = 0.44, p < 0.05). In another study
complex environments. Moore (1963) empha- of top management teams, Eisenhardt (1989)
sized that for predicting the future, organiza- also found that simultaneous consideration of
tions need “to make appropriate and timely multiple alternatives led to faster decisions in
adaptations to the inevitable” (p. 97). a high-velocity environment.
Long-range planning was introduced in the The future orientation of organizations is
1950s as a method for extrapolating an organi- also associated with better performance.
zation’s past into the future (Ansoff, 1988). Schriber and Gutek (1987) identified 13 tempo-
Long-range planning is based on the assumption ral dimensions of organizational culture, includ-
that the future environment will develop in a ing future orientation. Lim and Seers (1993), in
logically incremental manner. This planning is using 5 of these dimensions, found that, among
uniquely suited for making incremental strate- others, future orientation was a positive predic-
gic development explicit, guided, and coordi- tor of organizational performance, at least
nated among different parts of an organization. among manufacturing organizations in the
13-House.qxd 3/19/2004 5:20 PM Page 291
southern United States. They concluded that (Keegan, 1983), focused or broad (Daft & Weick,
an organization’s time norms have an impact 1984), narrow or comprehensive (Fredrickson,
on the managerial values that govern work 1984), and formal or informal (Stubbart, 1982).
processes and behaviors. Brommer and De La They may also differ in terms of resources
Porte (1992) suggest that a firm’s orientation to directed at the scanning function and in their
the future aligns the organization to face an ability to absorb what they have scanned
uncertain future environment, thereby enhanc- (Cohen & Levinthal, 1990).
ing its competitive advantage and performance With regard to selecting, organizations may
prospects. On the other hand, Onken (1999) be guided by rational–objective or subjective–
found another dimension of time, polychronic- personal sources (Mintzberg, 1975), detailed
ity (the emphasis on parallel processing), had and well-documented written reports or ver-
a moderately positive effect on the perfor- bally communicated and personally observed
mance indicators of return on assets and return information (Kotter, 1982), industry–academic
on sales. experts or colleagues–friends (Kobrin, Basek,
An interesting paradox of the construct of Blank, & La Palombara, 1980), and quantitative
future orientation is that changing the temporal or qualitative approaches (Mintzberg, 1975).
dimension of an organization’s culture may be When interpreting new information, organiza-
extremely difficult because time orientation is tions may frame the selected information in
one of culture’s most fundamental values, as terms of threat or opportunity (Dutton &
well as being ingrained at the individual level Jackson, 1987); costs or benefits (Tversky &
of analysis (Bluedorn, 2000). This is a signifi- Kahneman, 1974); and they may try to fit the
cant challenge for leaders who attempt to new information intuitively with a rather philo-
change an organization’s time orientation. sophical mission or rational–analytically with
Although not impossible, any attempt to change more concrete objectives.
a value or belief is difficult, and the embed- Validating the relevance of interpreted rules
ded nature of future orientation makes this may occur through a variety of systems in
even more challenging (House et al., 1997; different societal and organizational cultures.
Schein, 1985). Some illustrative systems include authoritarian-
Each societal culture may have its own way ism (as in Southern Asia), religious scriptures
to support future orientation in organizations. (as in the Middle East), bureaucratic procedure
For instance, studies on India and Southeast and (as in Latin Europe), historical precedent (as
East Asia show a process of strategic planning in Nordic cultures), political coalitions (as in
quite distinct from the rational approach domi- Germanic cultures), achievement goals (as in
nant in the West. Haley and Tan (1999) observe Anglo cultures), and consensus (as in Confucian
that, “Strategic planning in South and Southeast cultures). Thus, in Japan (a Confucian society)
Asia has developed into a process which is extensive face-to-face interaction is encouraged
ad-hoc and reactive, highly personalized, and multiple perspectives are sought (Yoshino,
idiosyncratic to the leader, and which uses 1976). The main goal is to develop a peripheral
relatively limited environmental scanning” vision to discern subtle changes and to gather
(p. 96). The firms following such strategic 75% of strategic information through surveil-
planning approaches have also been immensely lance as opposed to search processes (Keegan,
successful; for instance, Singapore International 1983). In the United States (an Anglo society)
Airlines, which follows a similar approach, is the focus is on specialized task-related knowl-
regularly named as the best-run airline in the edge (Kagono, Nonaka, Sakakibara, & Okumura,
world. 1985). In Sweden (a Nordic society) informal
More generally, Schneider (1989) identified personal views allow confirmation of historical
five elements of strategic planning (scanning, precedents and social relevance (Hedlund &
selecting, interpreting, validating, and prioritiz- Aman, 1984). Finally, in France (a Latin
ing) and suggested that each element may European society) theoretical discussion of
vary across organizations. In terms of scanning, bureaucratic rules guides the plans of the
organizations may be active or reactive organization (Crozier, 1964).
13-House.qxd 3/19/2004 5:20 PM Page 292
and difficult environments in emerging and similar link between the Protestant religion and
transitional economies can preclude future- a range of wealth-creating behaviors, includ-
oriented time perspective (Triandis, 1984). ing frugality and hard work. To Weber, wealth
Short-term concerns may predominate, and creation in a Protestant society was driven by
change may be guided primarily by past expe- a calling from God, that inspired people to
riences, hindering attempts toward strategic dream of a prosperous future for themselves
long-term planning. On the other hand, more and to work hard independently and confidently
predictable environments in industrialized to realize that dream.
markets support planning (Jaeger & Kanungo, Gaddis (1997) describes how members of the
1990). People in these environments give less Judeo-Christian and Islamic religions believe in
regard to past social and organizational tradi- the concept of a sovereign God and humans who
tions and make decisions based on the merits are not able to comprehend, let alone influence,
for the future. the future. Catholic religious beliefs include
the view that sacrifice is helpful to salvation,
Education: Abstract thinking is a necessary being charitable to others is a virtue, and people
condition for an individual to structure the should endure the wrongs that have been done
future (Trommsdorff, 1983); therefore, Tromms- against them. The consequences of these beliefs
dorff suggests that educational level may alter are that Catholics often believe that problems or
perceptions of the future. This is supported by events are meant to be and cannot be altered
Novaky and colleagues (1994) who, in a study (Wing Sue & Sue, 1990). Similarly, Islamic
of future orientation in Hungary, found that religious beliefs put an emphasis on partnership
level of schooling has an impact on proclivity and mutual help, as opposed to planning for
to engage in activities likely to be important the future. For instance, Islamic banking, based
for the future. For example, respondents with on the Koran (the Islamic holy book) forbids
lower levels of education tended to emphasize riba—the charging of interest. It rejects time
work and not to do anything for their future. value for money, and instead upholds the con-
Respondents with secondary schooling or cept of halal—or the purity in which investors
higher strove to do something for the future and derive their income from profits on investments
emphasized the importance of both further in which they participate and share the risks
study and work, as well as the education of their with the borrowers. Currently, due to a limited
children. Thus, the general level of education in development of Islamic banking products,
a culture affects the future orientation of collec- nearly 80% of Islamic funds are invested in
tive members. short-term commodity trades, whereas a Western
bank brings together a buyer and a seller of com-
Gender Roles: Cultural expectations of gender modities at a predetermined markup (Barakat &
roles influence the future orientation of individ- Sarver, 1997).
uals and this, in turn, affects societal values However, religion may not constrain the
and expectations (Nurmi, 1991). For example, development of future orientation and, simi-
Sundberg and colleagues (1983) showed that larly, lack of future orientation may emerge
Indian adolescents had greater gender differ- across all religions. Cahill (1998) ascribes the
ence in future orientation than their American or development of future orientation in several
Australian peers, and concluded that this was societies following a Judeo-Christian tradition
reflected in the way that values related to future to a change in philosophy from a cyclical view
orientation are passed on to children. to a linear view of the cosmos that facilitates
and accepts continuous change. Similarly,
Religion: The dominant religion in a society although Weber’s analysis ascribes a future-
also determines the future orientation of societal oriented characteristic to Protestant cultures of
members (Doob, 1971). We discussed earlier the the 19th and early 20th centuries, most postwar
relationship between Confucianism and future discourse has attributed a shorter-term orienta-
orientation as postulated by Hofstede and Bond tion to many Protestant societies focused on
(1988). Previously, Weber (1958) identified a consumption (Hofstede, 2001).
13-House.qxd 3/19/2004 5:20 PM Page 296
Further, religion may have a complex correlate with survival under hypercompetitive
relationship with future orientation. For instance, markets and superior performance under more
time plays a fundamentally important role in stable and predictable markets.
Hinduism and Buddhism (Agarwal, 1993). The
principles of Hinduism and Buddhism do not Organizational Structure and Size: Organizations
encourage hard work focused exclusively on that have a more future-oriented culture are likely
the pursuit of material wealth because these to have an organic structure (Onken, 1999).
societies judge individual worth less by material According to Burns and Stalker (1961), an
achievement than by spiritual orientation, accom- organic structure is rather flexible and fluid and is
plishments, and realizations (Hofstede, 2001). capable of dealing with unstable conditions
Consequently, planning gains a paramount through a network of relationships and a shared
significance in the Hinduism and Buddhism perception of goals. In comparison, less future-
context because planning is seen as essential oriented cultures are likely to have a mechanistic
to integrate the dualities of materialism and structure that buffers the organizational units
spiritualism. from environmental contingencies, changes, and
uncertainties, and allows them to perform their
Correlates of Organizational-Level routine and specialized tasks. Similarly, organi-
Future Orientation zational ecologists suggest that smaller and entre-
preneurial organizations, being more organic,
Although societal future orientation is a tend to be more future oriented and are guided by
major factor influencing an organization’s a vision of the future. On the other hand, larger
future orientation, an additional range of strate- organizations rely more on bureaucratic and rou-
gic contingencies and forces may have an inde- tine procedures and may be less future oriented.
pendent influence. Below, we discuss four such Although larger organizations may have greater
forces: Competitive environment and technol- resources for planning, they may value exploita-
ogy, organizational structure and size, institu- tion of prior capabilities more than exploring
tional systems, and leadership behavior. new opportunities for betterment of the future
(Hannan & Freeman, 1984).
Competitive Environment and Technology: The
environment in which an organization competes
Institutional Systems: Several elements of
may influence its future orientation. Judge and
national institutions have been identified to
Spitzfaden (1995) suggested that in high tech-
contribute to, or inhibit, the development of
nology industries the management of strategic
organizational future orientation. Some notable
time horizons is crucial to a firm’s survival.
elements are highlighted below:
Onken’s (1999) questionnaire-based study con-
firmed that in hypercompetitive industries char-
A. Planning and Control Tools: Many firms
acterized by high velocity and increased pace of
evaluate projects involving capital invest-
technological advancement speed is critical for
ments, and research and development, using
the firms just to survive in that industry. Each
discount techniques such as net present value
firm must emphasize speed or be vulnerable
(future profits discounted by the expected
to failure. In contrast, in nonhypercompetitive
minimum rate of return). The formal discount
industries, the firms may use speed to realize
techniques are based on measurable benefits
competitive advantages over their competitors.
and ignore intangible benefits such as learn-
In such industries, firms that are faster in their
ing and reputation. Therefore, organizations
decision making tend to perform better than
that use such techniques are likely to become
those that are not. For enhancing speed within a
more biased toward applied research and
given time, D’Aveni (1994) emphasized the
against long-term investments (Hayes &
importance of firms undertaking multiple strate-
Abernathy, 1980).
gies and implementing them quickly. Because
future orientation can aid in integrating and exe- B. Multidivisional Structure: As organiza-
cuting multiple strategies, it may be expected to tions grow and diversify, many of them
13-House.qxd 3/19/2004 5:20 PM Page 297
adopt a multidivisional form (M-form) of E. Patient Capital: The nature of the relationships
organizational structure (Chandler, 1962). between capital markets and the organiza-
The M-form helps corporate headquarters tions may also influence organizational future
evaluate the performance of divisions. orientation. According to Porter (1992), orga-
However, because divisional activities are nizations tend to become short-term oriented
difficult to compare and because the top if the “funds supplied by external capital
managers lack specialized knowledge about providers move rapidly from company to
the divisional activities, the headquarters are company usually based on perceptions of
forced to rely primarily on quarterly account- opportunities for near-term appreciation”
ing reports for information on divisional (p. 69). Such “impatient” capital results in a
performance. Divisional managers, therefore, breakdown in the relationship between the
have fewer incentives to make long-term investors and the organizations, thus adding
investments because such efforts are not to the cost of capital of the organization and
reflected in the accounting reports and are making long-term investments less attractive.
therefore not rewarded (Loescher, 1984). In contrast, if the capital is “patient” and
“dedicated,” the owners gain better apprecia-
C. Managerial Mobility: Many institutional
tion of the firm’s capabilities and long-term
systems encourage managers to make invest-
prospects and organizational future orienta-
ments that offer rather quick paybacks,
tion is promoted (Porter, 1992).
resulting in a less future-oriented organiza-
tional culture. A system characterized by a F. Intellectual Property Rights: Institutions that
high level of managerial mobility allows the have effective systems for protecting and
managers to exit the firm before the final suc- enforcing intellectual property rights strongly
cess of the project is known and thus benefit motivate the members of their organizations
from a reputation associated with the tempo- to invest in research and developmental pro-
rary success of the projects they initiated jects that can be patented, copyrighted, or
(Rumelt, 1987). trademarked. On the other hand, weak sys-
D. Institutional Investors: Institutional investors, tems for protecting intellectual property
such as pensions and mutual funds, and rights discourage future-oriented investments
professional investment managers may dis- (North, 1990).
courage organizational future orientation
(Johnson & Kaplan, 1987). Organizations Leadership Behavior: In addition to the com-
focused on future returns that are difficult petitive, structural, and institutional factors,
to quantify and validate may experience leadership also plays an important role in orga-
decreases in current profits and be unable nizational future orientation. An important func-
to pay short-term dividends. Professional tion of leadership is envisioning the future,
investment managers and institutional providing long-term direction, and articulat-
investors dislike such stocks, and so the futur- ing it to followers (Bryman, 1996; Kouzes &
istic organizations whose stocks are domi- Posner, 1996; Yukl, 1998). Thoms and Green-
nantly held by such investors may suffer berger (1995) suggest that the ability to create a
substantial declines in their stock values. vision is partially dependent on the future-time
Lack of substantive knowledge about the perspective of the leader: A leader who is
organization’s view of future returns further focused on future orientation is already inter-
discourages institutional and professional ested in directing energy toward the future.
investment managers from investing in Such proactive thinking about the future
the companies focused on novel futuristic increases in line with the leader’s hierarchical
projects. Consequently, institutional investors level and amount of responsibility. The future-
may focus more on the short-term profits oriented concept of vision is particularly associ-
and induce the organizations to have ated with charismatic leadership (Bass, 1998).
limited future orientation (Johnson & Kaplan, Bass notes that the key characteristic of charis-
1987). matic leadership is the ability to develop a
13-House.qxd 3/19/2004 5:20 PM Page 298
vision for the future and to communicate the in general. Adolescents’ interests and attitudes
vision to followers. toward various life domains differ among
societies with different future orientations, and
this is ultimately reflected in the values and
Correlates of Individual- practices they adopt as adults.
Level Future Orientation
Age: An individual’s future orientation is
In the field of psychology, hundreds of arti-
related to his or her perceptions of remaining
cles have been published over the past three
time to live, which is generally a function of
decades examining the correlates of future
the individual’s age. According to Carstensen’s
orientation. The correlates of individual-level
(1993) socioemotional selectivity theory, an
future orientation may be classified into two
individual’s perceptions of future orientation
types: general (Gjesme, 1983b) and contextual
influence his or her priority of specific goal
(Oner, 2000). The general correlates of future
contents. In later adulthood, personal networks
orientation focus on the adaptive and motiva-
intended to support improved social function-
tional effects of future orientation. The contex-
ing tend to reduce in size. Older people seek to
tual correlates of future orientation focus on the
maintain emotionally close social partners
role of individual interpretations, culture, and
while increasingly excluding more peripheral
situational factors (Oner, 2000; Trommsdorff,
social partners (Lang & Carstensen, 1994).
1983). The study of contextual correlates is
Age differences in goal selection are related to
guided by a view that people everywhere are
an individual’s perception of future time. If
oriented toward the past, the present, and the
younger people perceive their time to be lim-
future. Yet there are differences in the meaning,
ited, such as by imagining a geographical
manifestation, and value placed on these three
move, they also prefer familiar social partners.
orientations, specifically on future orientation,
On the other hand, if older people perceive
that can be better appreciated by referring
their time to be expansive, such as by imag-
to the specific context of people’s behaviors
ining advanced medical care, they tend to
(Sundberg et al., 1983). Individual future orien-
seek new relationships (Fung, Carstensen, &
tation is not simply a reflection of societal
Lutz, 1999).
future orientation, but is guided by an indi-
vidual’s interpretation of the cultural future
Deprivation: Scholars have suggested that
orientation.
deprivation of various types, such as socio-
economic comforts and positive experiences,
may mitigate the development of an indi-
General Correlates
vidual’s future-time perspective (Agarwal,
Upbringing: Upbringing is a key factor Tripathi, & Srivastava, 1983). Individuals
influencing the capacity to be coherent and to socialized in lower socioeconomic classes may
organize events in anticipation of the future consequently have a lower future orientation,
(Lomranz, Shmotkin, & Katznelson, 1983). whereas those socialized in higher socioeco-
Nurmi (1991) has noted that most of the devel- nomic classes are likely to have a higher future
opment of individual future orientation occurs orientation.
in family settings during adolescence, although Several studies in the United States (Mehta,
Benson (1994) shows that some future-oriented Rohila, Sundberg, & Tyler, 1972; Nurmi, 1987;
processes and behaviors occur in infants as O’Rand & Ellis, 1974; Trommsdorff & Lamm,
young as 9 months. Sundberg and colleagues 1975) have shown that adolescents with a rela-
(1983) found that in Anglo-American cultures tively high socioeconomic status extend their
adolescents focused on their personal happiness, horizons farther into the future compared with
future family, and leisure activities, whereas young people from a low socioeconomic back-
adolescents from more traditional societies ground. Adolescents with a high socioeconomic
expressed greater interest in their parent’s status tend to plan their future more than youths
family, the health of others, and societal topics with a relatively low socioeconomic position
13-House.qxd 3/19/2004 5:20 PM Page 299
(e.g., Cameron, Desai, Bahador, & Dremel, gives people the incentive to perform and
1977–1978). Nurmi (1987) found this to be true achieve and to strive for certain goals. Human
especially for hopes concerning vocational motivation is one of the most prominent con-
interests. Boocock (1978) found that American cerns in all social sciences research, because
adolescents from high-status homes make greater motivation is critical to higher perfor-
major life-course transitions at a later age than mance and other outcomes desired by indivi-
their low-status peers. Poole and Cooney (1987) duals. Gjesme (1975) showed that individuals
and Trommsdorff, Lamm, and Schmidt (1979) who were strongly oriented toward the future
found that future working life is more empha- were more motivated by a future goal than were
sized in the thinking of lower-class adolescents, individuals less oriented toward the future.
whereas middle-class adolescents tend to be Future-oriented individuals modified the psy-
more interested in education, career, and leisure chological distance between present and future
activities. by perceiving the future distance in time nearer
than individuals low in future orientation.
Adaptive Capacity: Several studies on the Further, more future-oriented individuals were
general correlates of future orientation focus better influenced by persuasive communications
on its adaptive aspects. Extended future orien- advocating a distant goal, whereas less future-
tation helps develop adaptive capacity among oriented individuals were better influenced by a
individuals by encouraging behaviors such more immediate goal message (Strathman,
as learning, savings, relationship building, Gleicher, Boninger, & Edwards, 1994).
marriage, and childrearing (Alvos, Gregson, &
Ross, 1993; Barndt & Johnson, 1955; Davids,
Contextual Correlates
Kidder, & Reich, 1962; Klineberg, 1967;
Mischel, 1974; Murrell & Mingrone, 1994; Investment: The concept of motivation mani-
Stein, Sarbin, & Kulik, 1968; Teahan, 1958; fests itself in specific concrete “investment”
Volder & Lens, 1982). Optimism is another behaviors that reflect underlying motivation.
personal characteristic related to future orien- More future-oriented individuals tend to gener-
tation (Marko & Savickas, 1998). A low orien- ate greater investments. Extensive research
tation toward the future can be disruptive to an has linked the concept of future orientation
individual’s optimism about his or her future positively with achievement orientation mani-
and adversely influence his or her feelings of fested, for instance, in greater academic suc-
control and general mental health (Cohen, 1967; cess. Students with greater future orientation
Keough et al., 1999). Thus, individuals with tend to invest greater effort in their classes,
higher future orientation have been found to show greater intensity to work in the class-
show a greater delay of gratification (Davids & room, and evidence more perseverance in
Falkof, 1975) and a lower psychopathology doing their homework (Peetsma, 2000; Roede,
(Wallace, 1956). For instance, students with 1989).
a strong future focus are more academically
successful and hold fewer irrational beliefs Delay of Gratification: Delay of gratification is
(Blinn & Pike, 1989; Poole & Cooney, 1987). the ability to defer an immediately available
Morris (1992) found that adolescent leaders reward for the sake of a delayed larger reward
hold a confident view of the self and also have (Nadler, 1975). More future-oriented people
a propensity to seek long-term rather than show a better ability to delay gratification and
short-term goals. a stronger sense of self-identity and career ori-
entation. Agarwal and colleagues (1983) show
Incentive and Motivating: The concept of time that individuals who are highly future oriented
perspective is also interpreted as a motivating as well as reflective were better able to tolerate
factor in human behavior (Frank, 1939). delay of gratification. In a study of the predic-
Negative future time perspectives, or low future tors of high job turnover, Griffeth and Hom,
orientation, are identified as a key source of lack (1988) found that participants with a lower
of motivation (Nuttin, 1964). Future orientation ability to delay gratification were more prone to
13-House.qxd 3/19/2004 5:20 PM Page 300
terminate employment and were also more Societal Health: One of the most important
likely to have an external locus of control. measures of societal health linked to future ori-
Individual willingness to delay gratification entation is the development of an individual’s
may be a function of societal cultural influ- relationships. More future-oriented individuals
ences. Gallimore, Weiss, and Finney’s (1974) show a stronger ability and willingness to form
study asked Hawaiian and Japanese American and sustain new relationships and thus demon-
adolescents to indicate what they would do strate a stronger societal fit. However, future-
with a $1,500 windfall. The Hawaiians indi- oriented individuals may also be more sensitive
cated they were more prone to immediate to reflecting on whether the relationships they
expenditure than the Japanese Americans, and are forming could be sustained, and thus show
were more likely to share with family and dissatisfaction in cultures and situations in
others. The Japanese Americans, on the other which relationships have a low probability of
hand, were more prepared to allocate the money being sustained. For instance, in Turkey, a high
to education. However, MacKinnon-Slaney’s concern for future commitment in opposite-sex
(1994) study suggests that the support for relationships has been found to negatively influ-
significant others (such as employers financing ence relationship satisfaction (Oner, 2000).
the study, partners caring for the family) can Individuals with greater concern for future com-
nurture the ability to delay gratification and mitment are more likely to show emotions such
encourage pursuit of higher education among as jealousy and possessiveness and be less satis-
adults, even in societies not acculturated to a fied in their relationships (Oner, 2000). Such
future-oriented belief in higher education. dissatisfaction effect in individuals concerned
with future commitment is associated with
Psychological Health: Numerous studies have low levels of self-esteem (Oner, 2000). Further,
investigated and generally confirmed Lewin’s because future-oriented individuals seek perma-
(1951) thesis relating future orientation as a nent relationships, they are circumspect in initi-
characteristic of psychologically well-developed ating relationships and in their ongoing
and healthy individuals, and conversely relating relationships, yet they may feel inexperienced
lower future orientation with a lower quality and lonely and not enjoy the temporary relation-
of psychological health (Novaky et al., 1994; ships in which they are involved. Some experi-
Sundberg et al., 1983; Zimbardo & Boyd, ence with temporary relationships may help
1999). Future orientation of individuals has individuals with high future orientation to be
been shown to facilitate performance of health less cautious about starting new relationships
maintenance and illness prevention behaviors. and develop meaningful, steady relationships
Less future-oriented individuals, on the other (Oner, 2000).
hand, show greater tendency to ignore the
potential future risks that health-compromising
behaviors can bring (Keough et al., 1999). SUMMARY OF CONCEPTUAL OVERVIEW
Keough and colleagues demonstrated that
greater future orientation is associated with In this overview of the literature on future
lower use of alcohol, tobacco, and drugs, orientation, we have discussed the construct
whereas a lower future orientation is associated from three perspectives: (a) cross-cultural (at
with a greater abuse of alcohol, tobacco, and the societal level), (b) managerial (at the organi-
drugs. Similarly, more future-oriented people zational level), and (c) psychological (at the
are more likely to delay their first sexual act individual level).
and are likely to have fewer sexual partners At the societal level, research has consistently
once they initiate sexual activity, as compared linked time perceptions and attitudes to impor-
with those with lower future orientation. tant outcomes, including economic success (e.g.,
Consequently, more future orientation is associ- Franke et al., 1991; Hofstede, 2001), although
ated with low-HIV risk behavior (Rothspan & the links are often complex and difficult to inter-
Read, 1996). pret. Future orientation appears to be developed
13-House.qxd 3/19/2004 5:20 PM Page 301
during childhood and adolescence, but the On the basis of the above analysis, Table
impact of society, culture, and geographic fac- 13.1 provides a summary comparison in terms
tors interact to determine adult attitudes. These, of “typical” strong versus weak cultural future
in turn, are derived from socioeconomic and orientation. The extreme cases are presented to
political conditions, education, gender roles, and help explain the concept, but it should also be
religion. Socioeconomic conditions influence clear that cultures do not neatly fit into stereo-
future orientation because under socially and types. Although the table shows cultural attrib-
economically disadvantaged conditions, a utes that tend to cluster together, it does not rule
shorter future orientation may be functional. out the fact that not all of these attributes may
Political stability ensures that societies, organi- go together at all times. Thus, societies can be
zations, and leaders can confidently look beyond expected to have differing mixes of the extreme
their immediate future. Level of education cases presented in Table 13.1.
affects members’ ability to think abstractly,
which is necessary for organizations, societies,
and leaders to structure the future. Gender roles GLOBE MEASURES
influence future orientation through families and OF FUTURE ORIENTATION
groups. Spirituality in religion is inherently ori-
ented to the future, so the dominant religion and Project GLOBE uses two constructs to portray
spirituality might contribute significantly to a a society’s level of future orientation: Ques-
culture’s future orientation. tionnaire items with an As Is response format
The societal effects carry over to organiza- pertain to respondents’ assessment of the
tions and individuals. Although there has been extent to which a society actually engages in
a dearth of research dealing directly with the future orientation practices. We refer to this
effect of future orientation on organizational as Future Orientation practices. Questionnaire
outcomes, findings to date have shown that items with a Should Be response format per-
long-term orientation is associated with a more tain to respondents’ assessment of the extent
strategic perspective. Organizations having a to which a society should engage in future
long-term orientation have also been shown to orientation practices; we refer to this as Future
be more adaptive and flexible, and thus better Orientation values.
able to cope with change in today’s turbulent Separate questionnaire scales were devel-
environment. In this case, the future orientation oped to measure each construct at the societal
of organizations and their leaders is determined level and at the organizational level within a
by the nature of the competitive environment society. Therefore, a total of four scales were
and technology, the future orientation of the created to measure each construct of interest—
individuals involved in strategic planning, and two scales measured practices and two scales
the nature of leadership. For instance, industries measured values. Separate questionnaires were
focused on technologies may encourage mem- administered to two groups of managers. The
bers in societies and organizations to be more first group responded to questions concerning
focused on long-term investments. Industries their organizational practices and values. The
focused on price-sensitive competition, on the dimensions are measured as composites of
other hand, may precipitate a concern toward quantitative items asked directly from the
immediate survival and short-term profitability. participating managers.
Similarly, the definition of charismatic leader- Tables 13.2a and 13.2b show sample ques-
ship is based on a concept of vision that implies tionnaire items relating to participants’ views on
a long-term view of the world, and substantial the societal practices and values. Tables 13.3a
evidence shows that such leadership has a posi- and 13.3b show sample items measuring the par-
tive influence on independent measures of both ticipant’s views on the organizational practices
organizational and work-group performance and values. The soundness of the psychometric
(Antonakis & House, 2002; Fiol, Harris, & properties of these scales was confirmed across
House, 1999). all societies sampled in the GLOBE project.
13-House.qxd 3/19/2004 5:20 PM Page 302
Table 13.1 Higher Future Orientation Societies Versus Lower Future Orientation Societies
1. I believe that the accepted norm in this society should be to: (reverse scored)
Plan for Accept the
the future status quo
1 2 3 4 5 6 7
Standard
Variable Mean Deviation Minimum Maximum Valid N
n = 61 cultures
Band
A B C D
a South Africa (Black Sample) * Higher scores indicate greater future orientation
b Canada (English-speaking) Our response bias correction procedure identified response
c Germany (West): Former FRG bias in some countries for this scale (see endnotes).
d South Africa (White Sample)
e Germany (East): Former GDR
among the nine GLOBE practices dimensions mean of 5.48 on Future Orientation values is
of culture, next only to Gender Egalitarianism third highest among the nine GLOBE value
practices, which has the lowest average score. dimensions, next to Performance Orientation
The average of societal Future Orientation values (5.94) and In-Group Collectivism values
values, across 61 societies, is substantially (5.66).
higher at 5.48, with a range of 4.33–6.20. In all Table 13.5 lists the scores of individual soci-
GLOBE societies, the respondents prefer a more eties on the GLOBE Future Orientation societal
strategic and long-term perspective of their practices scale. The scores ranged from 2.88
problems and the available opportunities. The (Russia) to 5.07 (Singapore) with a standard
13-House.qxd 3/19/2004 5:20 PM Page 305
deviation of 0.46. These findings are consistent bands A and B. These findings contradict the
with prior research. Comparative international popular view in the literature, reviewed earlier,
studies, for instance, have shown that journal- that lower socioeconomic status encourages
ism in Russia has “lagged behind” that of other a focus on immediate concerns for survival,
nations in the sense that its dominant temporal whereas higher socioeconomic status allows
constructions were less focused on the future savings and investments for the future.
throughout the 20th century (Ekecrantz, 2001). There could be two possible explanations
Singapore, on the other hand, practices a gover- for this surprising finding. First, industrialized
nance system in which “the government always and higher-income nations may associate
opts for what works for the country in the longer future orientation with capitalistic materialism
term rather than for what will please the people and value immediate gratification that is made
in the short term” (Principles of Governance, feasible by the focus on accumulating material
2001). The decisions and actions are legitimated wealth. On the other hand, emerging and
on the basis of their implications for the society lower-income societies nations may value the
in the long run, free from the vicissitudes of need for spiritual strength to look beyond their
popular opinion (Subramaniam, 2001). current material deprivations. Interestingly,
The practices scale results are grouped Thailand, which scores highest on the Future
into four bands. (See Chapter 11 by Hanges, Orientation values scale, also has a distinc-
Dickson, and Sipe for a further explanation of tively strong emphasis on Buddhism—a reli-
the bands.) Each band represents a group of gion known to strongly integrate spiritual and
societies that do not differ significantly from material needs (Swearer, 1997). Buddhism,
one another in terms of Future Orientation along with monarchy and nation, defines the
practices. Interestingly, contrary to the predic- very identity of the Thai people and is a core
tions of Hofstede and Bond (1988), there is pillar of their society (Eng, 1997).
no evidence of Confucian societies scoring Second, industrialized and higher-income
highest on the Future Orientation practices nations may like to enjoy the present more
or Anglo societies scoring the least. Confucian because they already have accumulated substan-
societies are scattered across three of the tial wealth and material resources. The emerg-
four bands: A (Singapore), B (Japan), and C ing and lower-income nations may see a
(China). Three Anglo societies, on the other stronger need for taking a long-term perspective
hand, are in Bands A (Canada) and B (United and sacrificing for the future because they must
States and United Kingdom). Many Eastern cope with scarce and limited resources.
Europe and Latin America societies score low Comparing Tables 13.5 and 13.6, respon-
on Future Orientation practices, which may dents in all except one society value Future
partly indicate the turbulent and transitional Orientation more than they actually practice it.
conditions in those societies during the 1990s. The sole exception is Denmark, whose aggre-
Many Northern European societies, on the gated practices rating is 4.44 and values rating
other hand, score quite high on future orienta- is 4.33. Denmark in fact, as noted earlier, has
tion practices, which may partly be attributable also scored a lower rating for Future Orienta-
to their political stability and socioeconomic tion values than any other country in the
status. GLOBE sample. Denmark’s low values score
Table 13.6 shows the scores of individual for Future Orientation is reflected in the nature
societies on the GLOBE Future Orientation of Danish media. Denmark stands apart as the
society values scale. The scores ranged from only nation in the world that gives history,
4.33 (Denmark) to 6.20 (Thailand). Surpris- with very wide time frames, a prominent posi-
ingly, most industrialized and higher-income tion in the media (Ekecrantz, 2001). Danish
countries have comparatively low scores on media regularly contain stories about or ref-
Future Orientation values and fall into bands erences to history, not necessarily as a back-
B, C, and D. In contrast, emerging and lower- ground for today’s events, but as anecdotes
income societies score high on Future that appear to serve little more than entertain-
Orientation values and are mostly placed in ment value (Ekecrantz, 2001).
13-House.qxd 3/19/2004 5:20 PM Page 306
Band
A B C D
a South Africa (White Sample) * Higher scores indicate greater future orientation
b Canada (English-speaking) Our response bias correction procedure identified response
c Germany (East): Former GDR bias in some countries for this scale (see endnotes).
d South Africa (Black Sample)
e Germany (West): Former FRG
Table 13.7a Descriptive Statistics for GLOBE Industry-Level Societal Future Orientation
Standard
Variable N Minimum Maximum Mean Deviation
Future Orientation practices suffer most from the few private firms that entered the telecom
the uncertainty and unpredictability of not sector have become monopolies in their niches
addressing the longer-term fundamental issues. and have just adopted the monopolistic practices
Therefore, such societies are most conscious of of the public sector, with few incentives for effi-
the need for moving toward a more strategic and ciency (Blasko, 1998). Because strategic plan-
spiritually fulfilling perspective. ning is critical to development of technology in
the telecom sector, the respondents in the tele-
com sector, including those from Thailand, place
Overall Scores and Industry Scores
a higher value on Future Orientation.
In each society, GLOBE data were collected Table 13.7b shows the correlation between
from one or more of three industries: food pro- the overall societal Future Orientation practices
cessing, financial services, and telecommunica- score for each society and the scores for the par-
tions. Table 13.7a provides the means of societal ticipating industries. As seen in this table, all
Future Orientation practices and values for each correlations are significant, though correlations
of the industries across all the sampled societies. involving telecom industry practices of Future
The means of societal Future Orientation Orientation are of smaller size than the correla-
practices and values are not significantly differ- tions between the food and finance sectors.
ent across the three industries (p > .10). However, The Future Orientation values scores are also
the telecommunications sector has substantially highly correlated across industries, as shown
lower minimum (2.06 in Thailand) and lower in Table 13.7c. In terms of values, the telecom
maximum (4.67 in Switzerland) practices scores industry’s pattern is quite similar to the other two
than the other two sectors. The telecommuni- industries. Thus, there appears to be more cross-
cations sector also has a higher standard devia- industry convergence in Future Orientation with
tion on the Future Orientation practices scores. respect to values than with respect to practices.
Further, the telecommunications sector has
substantially higher maximum Future Orienta-
The Correlation Between
tion values scores (6.63 in Thailand), whereas
the finance sector has substantially lower mini-
Overall Practices and Value Scores
mum Future Orientation values scores (4.14 in Industry scores on the Future Orientation
Denmark). values scales show a negative correlation with
In the telecommunications industry, tradi- industry scores on the Future Orientation prac-
tional reliance on government control con- tices scales, as can be seen in Table 13.7d. The
tributed to weaker Future Orientation practices correlation is of moderate size and significance
scores in such societies as Thailand. In Thailand, for the finance (− 0.33, p < 0.05) and food
13-House.qxd 3/19/2004 5:20 PM Page 308
** p < .01.
Table 13.7d Future Orientation: Industry Correlation Matrix (Society Values and Practices)
Telecommunication −.68**
industry score values n = 32
Table 13.8 Correlations of Societal Future Orientation With Other Cultural Dimensions of Society
(− .39, p < .05) sectors, but is quite strong values scores with the other eight dimensions of
(−.68, p < .01) for the telecommunications sec- culture assessed in the GLOBE program are
tor. These findings indicate that the societies in presented in Table 13.8.
which the telecom industry was traditionally
governed by short-term practices, due to such
Future Orientation Practices
factors as protective government regulation, are
now acutely aware of the need to reform and
and Other GLOBE Dimensions
adopt more strategic approaches for infrastruc- Managerial responses to the GLOBE
ture improvement and development. questionnaire reveal that Future Orientation
practices are positively correlated with the prac-
tices of Uncertainty Avoidance, Institutional
The Correlation With Other
Collectivism, and Performance Orientation, but
Globe Society Culture Dimensions
are negatively correlated with the practices of
In this section we review the relationships Power Distance and In-Group Collectivism.
between managerial responses to the GLOBE Indeed, several societies with relatively high
societal-level Future Orientation practices scale scores on the GLOBE Future Orientation prac-
and responses to other scales. The correlations tices scale are reported to have well-developed
of societal Future Orientation practices and collective institutions. Such institutions likely
13-House.qxd 3/19/2004 5:20 PM Page 310
encourage members to consider collective Orientation. In Japan, for instance, there has
interests in making decisions about how to man- been a strong recognition of the need to develop
age information, technology, and knowledge, a more innovative orientation and to overcome
and in reducing uncertainty. Such a mind-set the traditional mind-set of imitative learning
allows the members to look far into the future to within groups.
assess the effects of their current actions
because the future uncertainties and risks are Future Orientation Values
lowered. Performance criteria can be specified
and Other GLOBE Dimensions
more clearly and rewards can be given on the
basis of these criteria. There is a limited role Table 13.8 also shows that Future Orienta-
given to power under these conditions, and the tion values are positively correlated with the
members generally do not rely on families for values of Uncertainty Avoidance, Institutional
absorbing uncertainties or meeting unantici- Collectivism, In-Group Collectivism, and
pated contingencies. Uncertainties and contin- Performance Orientation, but are negatively
gencies, even if they arise, could possibly be correlated with the values of Gender Egalita-
met through adaptive, entrepreneurial planning. rianism. Aspirations for higher future orienta-
Denmark and the Netherlands are examples tion may reflect an understanding that most
of GLOBE societies that fit the prototype uncertainties can be managed through tech-
of high practices scores of Future Orientation, niques such as scenario planning, and their
Uncertainty Avoidance, Institutional Collec- negative effects minimized through lowering of
tivism, and Performance Orientation, but low uncertainty via better information, knowledge,
practices scores of Power Distance and In- and technology. Collective safety nets of the
Group Collectivism. Gupta, Macmillan, and institutions and of the families and relation-
Surie (in press) report that in Nordic and ships may be additional vehicles valued in these
Germanic societies, such as the Scandinavian societies. However, such societies may also
countries and the Netherlands, entrepreneurial seek a more traditional role for women. Prior
leadership tends to be more effective. Entre- research suggests that women may put less
preneurial leadership manages change and effort into advancing professionally because
uncertainty through a discovery driven plan- they put lower value on status and salary and
ning mind-set, through painting a scenario, dis- more value on their families (Van Vianen &
mantling power structures, and empowerment Fischer, 2002). Consequently, women may be
of the members for collective gain. expected to provide a family safety net in soci-
Further, as seen from Table 13.8, Future eties that value future orientation.
Orientation practices scores are negatively Guatemala and Ecuador are societies that
correlated with the values of Uncertainty fit the prototype of high values scores of
Avoidance, Institutional Collectivism, and In- Future Orientation, Uncertainty Avoidance,
Group Collectivism. In societies with strong Institutional Collectivism, In-Group Collec-
future oriented practices, the leaders may have tivism, and Performance Orientation, but low
confidence in the abilities of their members and values scores of Gender Egalitarianism. In
in the collective safety net of their institutions, general, many Latin American societies identify
and may expect their members to be more inno- their female population as a “vulnerable group”
vative and tolerant of change. These leaders with only a limited capacity to engage in future
may also prefer the members to develop their planning. Barrig and Beckman (2001) observe
own personal potential, and to have freedom to that the convulsive political situations in nations
satisfy their own aspirations, and do not value a such as Ecuador, coupled with the continuing
strong institutional or in-group collectivism. lack of stable legal systems in nations such as
Japan and South Africa (Black sample) are Guatemala, make social participation, especially
examples of societies that fit the prototype of of women, largely a myth.
low values scores of Uncertainty Avoidance, In addition, Table 13.8 indicates that Future
Institutional Collectivism, and In-Group Collecti- Orientation values are positively correlated
vism, but high practices scores of Future with the practices of Power Distance, In-Group
13-House.qxd 3/19/2004 5:20 PM Page 311
Table 13.9 Spearman’s Rank Correlation Among Different Measures of Future Orientation
** p < .01.
NOTE: We thank Shalom Schwartz and Lilach Sagiv for providing the above correlations.
N = 48
** Correlation significant at p < 0.01.
Table 13.10b GLOBE Future Orientation and Inglehart’s Modernization and Postmodernization Dimensions
can only take one year at a time, with response The Trompenaars and
on an Agree-Neutral-Disagree scale. We have Hampden-Turner Study
analyzed their results on a 1–3 scale in
which 1 = Agree and 3 = Disagree. Matthews Trompenaars and Hampden-Turner (1998)
and Ornauer surveyed only 11 countries, 8 asked respondents to indicate their past,
of which overlap with the GLOBE sample. present, and future horizons by using a 1–7
However, neither the GLOBE Future Orienta- scale, in which 1 = seconds and 7 = years.
tion practices nor values scale was related They computed a long-term horizon index by
to the belief that the future is uncertain averaging the scores for the length of past,
(p > 0.10). We might speculate that in soci- present, and future horizons reported in each
eties that view the future as amenable to plan- society. Their data include 30 nations that are
ning, future-oriented practices may be greater. common with the GLOBE study. As reported
Such societies may also value more sponta- in Table 13.9, their long-term horizon index is
neous behaviors, rejecting overanalysis and significantly and negatively correlated with
overplanning, which may preclude the capac- GLOBE’s Future Orientation values scale
ity to take advantage of unpredicted trends and (rho = − .47, p < .05), though the correlation
events. with GLOBE’s Future Orientation practices
13-House.qxd 3/19/2004 5:20 PM Page 313
scale is not significant (rho = .07, p > .10). The Schwartz Study
Some societies that place a higher value on
future orientation may have a short-term hori- Schwartz (1999) identified three fundamen-
zon. Similarly, some societies that place a tal values distinguishing all cultures: (a) auton-
lower value on future orientation may have a omy versus embeddedness, (b) hierarchy versus
long-term horizon. egalitarianism, and (c) mastery versus harmony.
Of these, the first is very pertinent to the cultural
The Hofstede and Spector concept of Future Orientation.
In embedded cultures people are perceived
and Colleagues Studies
as entities bounded in the collectivity: They
Hofstede (2001) reinterpreted the Confucian find meaning in life by participating in the
Dynamism scale discussed earlier in this chap- group and identifying with its goals. Values of
ter as a measure of long-term orientation. The social order, family security, respect for tradi-
original Confucian Dynamism scale was based tion, and obedience are emphasized. On the
on Chinese Cultural Connection questionnaire other hand, in autonomous cultures, individuals
surveys administered in 23 nations (Hofstede & are perceived as autonomous, nonbounded
Bond, 1988). Hofstede (2001) added scores for entities who find meaning in life through their
11 additional nations using data from the uniqueness. Two types of autonomy are distin-
European Media and Marketing Survey 1997, guished in Schwartz’s framework. First, intel-
which contained four questions from the lectual autonomy, in which individuals are
Confucian Dynamism scale. Of these four encouraged to follow their own ideas and
questions, only two—“thrift” and “respect for intellectual directions, emphasizes creativity,
tradition”—were correlated in the expected curiosity, and broadmindedness. Second, affec-
negative direction (r = −.64, p < .01, n = 11), tive autonomy, in which people are encouraged
and interpreted by Hofstede (2001) as measures to find positive experiences for themselves,
of long-term orientation. These measures are with an emphasis on variety, excitement, and
subject to multiple interpretations and lack face pleasure in life.
validity. Hofstede (2001) also estimated com- Cultures embedded in relationships may
parable scores for 7 additional nations on the take a more long-term view of their efforts—
basis of observations and descriptive infor- relations generally encourage a quid pro quo
mation. Consequently, in total, comparable kind of behavior that is ongoing in nature.
GLOBE data are available for a subsample of On the other hand, autonomous cultures may
27 nations. As shown in Table 13.9, neither have a more short-term focus, with more con-
GLOBE Future Orientation practices (rho = tractual kinds of behavior (see Granovetter,
.03) nor GLOBE Future Orientation values 1985). Indeed, as reported in Table 13.10a,
(rho = − .06) scales show any relationship with GLOBE’s Future Orientation values scale is
Hofstede’s presumed Long Term Orientation negatively related to Schwartz’s autonomy-
scale. Data on the Confucian Dynamism scale embedded scale (r = − .50, p < .01). Further,
were also collected by Spector and colleagues GLOBE’s Future Orientation values correlates
(2001), who also referred to the scale as long- negatively with affective and intellectual
term orientation. The correlation of Spector and autonomy (r = −.43 and − .46, p < .01) and
colleagues’ scores with the GLOBE Future positively with Schwartz’s embeddedness
Orientation practices scale (rho = .28, p > .10) scale (r = .50, p < .01). However, GLOBE’s
and GLOBE Future Orientation values scale Future Orientation practices scale is not corre-
(rho = .25, p > .10) is also small and insignifi- lated with any of these measures. Thus, the
cant. These findings suggest a need for caution value of future orientation shares the belief in
in interpreting the Confucian Dynamism scale quid pro quo and long-term interaction inher-
as a proxy for long-term orientation, and ent in embeddedness, as opposed to contractual
indicate that such interpretation may not be transience and spot relationships associated
appropriate. with autonomy.
13-House.qxd 3/19/2004 5:20 PM Page 314
The Inglehart and Baker Study societal Future Orientation values scale is
negatively and significantly correlated to
Inglehart and Baker (2000, p. 21) identify
both the traditional–secular–rational authority
two dimensions in which cultures become dif-
scale (r = −.70, p < .01) and the survival–
ferentiated as a result of the differences in their
self-expression scale (r = −.69, p < .01). In socie-
development. The first, termed traditional ver-
ties that rely on traditional authority and survival
sus secular-rational authority, is “linked with
values, the respondents are more likely to endorse
early industrialization and the rise of the work-
the value of future orientation—possibly to seek
ing class” and reflects the process of modern-
a higher quality of life.
ization. The second, termed survival versus
self-expression values, “reflects the changes
linked with the affluent conditions of advanced
industrial society and with the rise of the ser- Summary
vice and knowledge sectors” and captures the The GLOBE study appears to be the first
dynamics of postmodernization. to have clearly identified future orientation
A traditional worldview, prominent in prein- as a societal cross-cultural dimension within a
dustrial and premodern societies, reflects an comprehensive theoretical framework. As we
emphasis on obedience to traditional authority, noted earlier, the Hofstede (2001) scale is
particularly deference to God, that goes with really a measure of Confucian Dynamism, and
deference to the family and the nation. On the future orientation is only a presumed facet
other hand, in a secular worldview prominent in of this measure, while the Trompenaars and
industrial and modern societies, authority is Hampden-Turner (1998) scale is a measure of
legitimated by rational–legal norms linked with time horizon, as opposed to future-oriented
an emphasis on economic accumulation and behavior. The correlations with the scales
individual achievement. The survival and self- developed by Schwartz (1999) and Inglehart
expression values dimension reflects how and Baker (2000) to capture the fundamental
unprecedented levels of wealth and the emer- values and shifts in these values further under-
gence of welfare states in a postindustrial soci- line the significance of GLOBE’s Future
ety give rise to a shift from scarcity norms Orientation scale.
emphasizing hard work and self-denial, to post-
modern values emphasizing the quality of life,
subjective well-being, and related postmaterial-
ist priorities of self-expression. According to ECONOMIC AND SOCIAL CORRELATES OF
Inglehart and Baker (2000), the societal devel- GLOBE SOCIETAL FUTURE ORIENTATION
opment process is associated with an increased
probability of a shift first from traditional to In this section, we compare and contrast the
secular–rational authority, and second, from GLOBE findings on societal Future Orientation
survival to self-expression values. Consequently, with those of other major cross-cultural and
the societies ahead on secular–rational and comparative studies focusing on different coun-
self-expression dimensions are likely to be tries’ social and economic performance. As
more future oriented. explained in Chapter 7 by Javidan and Hauser,
As reported in Table 13.10b, the GLOBE four major ongoing reports provided the rele-
societal Future Orientation practices scale is vant data: the IMD’s Global Competitiveness
positively correlated to both the traditional– Ranking, the World Economic Forum’s Competi-
secular–rational authority scale (r = .28, p > .10) tiveness Ranking, the United Nations Human
and the survival–self-expression scale (r = .61, Development Report, and the World Values
p < .01), though the correlation is significant Survey. The findings of these studies are
only with the latter. Societies reported to be grouped into eight categories: economic health,
future oriented are more likely to encourage human condition, scientific advancement, fam-
individuals to express themselves and to lead ily and friends, political ideology, spiritual atti-
a high quality of life. In contrast, the GLOBE tudes, gender attitudes, and national savings.
13-House.qxd 3/19/2004 5:20 PM Page 315
Government World
Future Economic Support for Societal Support for Competitiveness
Orientation Prosperity Prosperity Competitiveness Index
** p < .01.
These categories of data will be examined to test These results show significant positive
hypotheses that future-oriented societies tend to relationships between the economic indicators
(a) be more economically prosperous, (b) enjoy and the GLOBE practices measures, and nega-
higher levels of societal health, (c) be more tive relationships with GLOBE values measures
scientifically advanced, (d) have looser ties with of Future Orientation. Economic prosperity, for
family and friends, (e) have more open and example, is positively associated with GLOBE
progressive political ideology, (f) have more Future Orientation scale scores in practice
spiritually oriented attitudes, (g) have favorable (r = .54, p < .01), and negatively correlated with
gender attitudes, and (h) have higher rates of Future Orientation values scores (r = −.62,
national savings. p < .01). There are three possible explanations
for this finding. First, the negative sign associ-
ated with the values score and prosperity may
Future Orientation and Economic Health
just be an artifact of the negative correlations
What is the relationship between future that we noted earlier between values and prac-
orientation and economic health? Do the more tices indices of Future Orientation as measured
future-oriented societies enjoy healthier eco- in Project GLOBE. Thus, cultures that have
nomies? What about the societies with higher shorter-term practices are less prosperous, and
aspirations for future orientation? therefore aspire for a longer-term perspective
Table 13.11a provides the answers to these that would be conducive to higher prosperity.
questions. It shows the correlation coefficients Second, in some societies, value expression
between the two GLOBE dimensions of Future alone may not guarantee the implementation
Orientation, practices and values, and the vari- of the planning practices or progress on the
ous elements of economic health. These are eco- economic front.
nomic prosperity, which refers to consumption Third, GLOBE’s practices and values scales
and growth; government support for prosperity, may partly capture different aspects of the
reflecting the extent to which the government Future Orientation construct. Gupta, Sully de
and the political body are supportive of eco- Luque, and House show in Chapter 9 that
nomic progress; and societal support for com- GLOBE’s Future Orientation practices scale
petitiveness, a measure of the general social scores are correlated with the unobtrusive and
attitude toward and support for business com- other external measures dealing with advanced
petitiveness. In addition to these measures, we planning and investments in intellectual proper-
also examined the relationship between GLOBE ties. In contrast, GLOBE’s Future Orientation
findings and a well-known measure of global values scale is correlated with the World Values
competitiveness, the World Economic Forum’s Survey indicators dealing with spiritual orienta-
Global Competitiveness Index. tion. In societies that value future orientation
13-House.qxd 3/19/2004 5:20 PM Page 316
Human
Future Societal Human Life General Development
Orientation Health Health Expectancy Satisfaction Index
there is a greater emphasis on supporting both In this section, we examine the relationship
spiritual as well as material growth, whereas in between GLOBE findings on Future Orientation
societies that practice future orientation there and five different dimensions of human
tends to be an emphasis on material growth condition. Societal health refers to the quality of
only. If the practices of future orientation do not life, safety, and security in a society. Human
support spiritual growth, then planning may not health reflects the physical health of the popu-
be widely practiced even though the people may lace. Life expectancy reflects expected years of
be quite aware of the value of a futuristic per- life at birth. General satisfaction measures the
spective. On the other hand, in societies whose extent of peoples’ happiness with their lives;
members place lower values on future orienta- this term comes from the World Values Survey.
tion, the thrust toward spiritual orientation tends The Human Development Index, prepared by
to be low. Under this situation, it is potentially the United Nations, is a measure of longevity,
feasible for the societies to practice planning for education, and standard of living.
material prosperity. (For a more detailed discus- As shown in Table 13.11b, a pattern of
sion, see Chapter 9) significant correlations provides a qualified
Our results provide a qualified support for support for the second major hypothesis:
the first hypothesis. Although Future Orienta- Countries that have highly future-oriented
tion practices are aligned with societal eco- practices tend to enjoy higher levels of societal
nomic success, Future Orientation values are health. Specifically, GLOBE Future Orientation
inversely related to societal economic success. practices are positively associated with all five
In other words, low level of societal economic of the measures of human condition, and the
success appears to contribute to a support for correlation with two measures is significant:
delay in gratification, and thus values of future Societal health and general satisfaction.
orientation. GLOBE Future Orientation values, on the other
hand, are associated negatively with all five
of the measures, and the correlation with four
Future Orientation
measures is significant: Societal health, life
and Human Condition
expectancy, general satisfaction, and human
Economic health is only one aspect of a development. These findings are similar to
society’s well-being. Another dimension is those found for economic health. It is possible
the general health of the populace of the soci- that the managerial reports of Future Orien-
ety. What is termed human condition here tation values reflect an expression of depri-
refers to the general quality of life in a partic- vation, frustration, or dissatisfaction with the
ular society. status quo of Future Orientation practices.
13-House.qxd 3/19/2004 5:20 PM Page 317
Future Orientation and Political Ideology government. In societies in which people value
more future orientation, people are less likely to
The next domain of interest is political ideol- express their voice and be politically active,
ogy. We use six measures of political ideology: and the government is more likely to play an
Disdain for democracy, passiveness, lack of voice, important role in the governance process. As
dislike for democracy, role of government, and noted earlier, Singapore is an illustration of
stability. such a society.
As shown in Table 13.11e, all six measures of We thus find support for the fifth hypothesis:
political ideology are negatively and significantly More Future Orientation practices are associ-
associated with GLOBE’s Future Orientation ated with more active political ideology at the
practices scale. People in societies with stronger mass level, whereas more Future Orientation
future-oriented practices are less likely to be pas- values are associated with more active political
sive and to reject democracy, and are more likely ideology at the government level.
to express their voice. They are also less likely
to support a strong role for the government, and
Future Orientation and Spiritual Attitudes
more likely to reject stability in favor of devel-
opment. In essence, societies with stronger As discussed previously, the concept of
future-oriented practices may encourage stronger future orientation has gained prominence with
opportunities for people to be involved and the rise of an interest in the spiritual aspects of
engaged in the political and governance process. religion. A long-term orientation is expected to
In contrast, the GLOBE’s Future Orientation support an integration of the material and the
values scale score is positively associated spirit and allow people to discover more mean-
with all six measures of political ideology, ing and fulfillment in their life.
and the correlation is significant for three mea- We use two measures of spiritual attitudes.
sures: Passiveness, lack of voice, and role of Religious devotion refers to being devoted to
13-House.qxd 3/19/2004 5:20 PM Page 319
soul-searching and purifying rituals. Religious scores also tend to be less economically and
dogma refers to belief in the matters of spirit. socially prosperous. Consequently, the empow-
As shown in Table 13.11f, the GLOBE Future erment of women may become particularly dif-
Orientation values scale is positively and signif- ficult in such societies. On the other hand, the
icantly associated with both measures of spiri- findings also suggest that if societies adopt
tual attitudes. In contrast, the GLOBE Future future-oriented practices, they likely recognize
Orientation practices scale does not show any the need and benefits of empowering women
correlation with either of the measures. and make stronger efforts toward realization of
These findings support the sixth hypothesis gender egalitarianism.
that more future-oriented aspirations are associ- The findings thus appear to be consistent
ated with spiritually oriented attitudes. with the hypothesis that future orientation prac-
tices are conducive to positive attitudes toward
gender.
Future Orientation and Gender Attitudes
Gender is another factor of interest in rela-
Future Orientation and National Savings
tion to GLOBE’s cultural dimension of Future
Orientation. Table 13.11g correlates Future As noted earlier, Hofstede (2001) has identi-
Orientation practices and values scales with a fied a national propensity to save as an impor-
measure of gender equality. The correlation is tant validation indicator for the construct of
positive for Future Orientation practices (.40, long-term or future orientation, and several
p < .05), but negative for Future Orientation scholars, such as Edwards (1995), have identi-
values (−.44, p < .01). As noted previously, fied social attitudes as one important determi-
societies with higher Future Orientation values nant of thrift. One may expect future orientation
13-House.qxd 3/19/2004 5:20 PM Page 320
Table 13.11h Relationship Between Future Orientation and National Savings Rate (as percentage of Gross
Domestic Product)
National Savings Rate Future Orientation Societal Values Future Orientation Societal Practices
Tropical wet and dry (savanna) 7 3.61 0.30 0.11 5.99 0.13 0.05
Marine west coast (maritime) 10 4.15 0.43 0.14 4.98 0.32 0.10
NOTE: SD = Standard deviation of the mean estimate. 95% confidence interval = Mean ± 1.96* Standard Error (SE).
(Molles, 2001). Further, higher solar input analysis of variance (not shown) indicated that
favors higher rates of photosynthesis in the trop- climatic clusters account for 20% of the
ics, which increases the productivity of the plant between-society differences in GLOBE Future
resources and, hence, of the formation of plant- Orientation practices (p > .05), and 57% of the
based minerals and other resources. With more between-society differences in GLOBE Future
total resources and more diversity in resources, Orientation values (p < .01). The differences in
more total opportunities exist and more diver- societal Future Orientation practices are not sig-
sity in these opportunities exist. As a result of nificantly related to the physical climates of
these factors, future orientation becomes less those societies. However, the differences in val-
costly and more beneficial for the people in ues of societal Future Orientation are signifi-
tropical societies. The opportunity cost of cantly related to physical climates.
resources would be lowered so that delay in the Table 13.12a shows the grand mean scores
gratification of any specific kind of resource or for Future Orientation values and practices for
even total resources should be perceived less each climatic region. We used Analysis of
costly and demanding on the people. On the Variance Contrast tests to assess if the means in
other hand, future orientation would be per- a climatic cluster are significantly higher or
ceived as costly for the people in nontropical lower compared with all other climatic clusters.
societies because limited resource availability Societal values of Future Orientation are signifi-
should make the opportunity cost of those cantly low in maritime climates (mean = 4.98,
resources high. Therefore, we hypothesize that contrast = −.61, p < .01) and in continental
the values of future orientation will be signifi- climates (mean = 5.27, contrast = −.29, p < .01);
cantly lower in the nontropical societies than in and are significantly high in savanna climates
the tropical societies. As described in Chapter (mean = 5.99, contrast = .51, p < .01).
10, Gupta & Hanges sorted participating The above findings are broadly consistent
GLOBE societies into seven clusters on the with our hypothesis that nontropical societies
basis of climate: tropical humid, tropical wet tend to favor a lower future orientation due to a
and dry (savanna), subtropical humid, subtropi- higher perceived opportunity cost of resources.
cal wet and dry (Mediterranean), desert, marine The lower Future Orientation values ratings for
west coast (maritime), and continental. An maritime and continental climates indicate
13-House.qxd 3/19/2004 5:20 PM Page 322
NOTE: SD = Standard deviation of the mean estimate. 95% Confidence interval = Mean ±1.96* Standard Error (SE).
higher aspirations for consumption, and may Table 13.12b shows the grand mean scores
reflect respondents’ desire to enjoy the present. of Future Orientation values and practices
In tropical wet and dry (i.e., savanna) climates, for each geographical region. We used the
stronger planning orientation is valued. Analysis of Variance Contrast test to assess
whether the means in a cultural cluster are
significantly higher or lower compared with
all other cultural clusters. Future Orientation
Future Orientation and Regional Clusters
practices scores are significantly higher in
As discussed in Chapter 8 by Hanges Germanic Europe (contrast = .54, p < .01)
and Dickson, there are significant differences and Nordic Europe (contrast = .49, p < .05);
in values and practices for every cultural they are significantly lower in Eastern Europe
dimension among the 62 societal cultures (contrast = −.58, p < .01), Latin America (con-
studied. Gupta and Hanges in Chapter 10 have trast = −.41, p < .01) and the Middle East
also categorized GLOBE societies into 10 (contrast = −.36, p < .05).
cultural or regional clusters: Eastern Europe, Future Orientation societal values scores are
Latin America, Latin Europe, Nordic Europe, significantly higher in sub-Saharan Africa (con-
Germanic Europe, Anglo, Middle East, trast = .47, p < .01), Southern Asia (contrast =
Southern Asia, Sub-Saharan Africa, and .46, p < .01), the Middle East (contrast = .43,
Confucian Asia. An analysis of variance p < .01), and Latin America (contrast = .33,
(ANOVA) revealed that cultural clusters p < .01); they are significantly lower in Nordic
accounted for 66% of the between-society Europe (contrast = −.75, p < .01) and Germanic
differences in Future Orientation practices Europe (−.48).
(p < .01), and 71% of the between-society dif- Respondents’ reports of strong practices
ferences in Future Orientation values (p < .01). and weak values for Future Orientation in the
13-House.qxd 3/19/2004 5:20 PM Page 323
Nordic and Germanic Europe clusters could be Asia and Anglo clusters are not significantly
partly explained by their reliance on “profes- higher or lower than those for the other clusters
sional bureaucracy” (Mintzberg, 1979; Weber, and are similar to the international average. This
1958). Professional bureaucratic work culture finding casts concerns on trying to measure
in Nordic and Germanic Europe involves an the construct of Future Orientation using the
emphasis on the skills and knowledge of operat- Confucian Dynamism scale (Hofstede, 2001).
ing professionals to function. Professionals are
formally trained and granted considerable auton-
omy over their own work (Weber, 1958). They GLOBE FINDINGS ON
work rather independently of their teams, with ORGANIZATIONAL FUTURE ORIENTATION
whom they coordinate through well-planned
standardized rules and systems. Despite these GLOBE used a different set of questions to
well-planned standards, professional aspirations measure organizational culture. This question-
are usually too complex to be fully codified naire was administered to a different set of
through formal and advance planning; a sense of respondents than those who completed the
instant discretionary decision making grounded societal culture questionnaire. To compute orga-
in professional know-how is always valued for nizational scores on Future Orientation we used
enjoying the present (Mintzberg, 1979). data from only those organizations having at
In contrast, weak practices and strong values least seven respondents in the GLOBE sample.
scores for Future Orientation in the Middle Table 13.13 gives the summary statistics for
East and Latin America clusters are indicative organizational Future Orientation for a sample
of their peculiar approach to relationships. In of 276 organizations. Below, we discuss the
these societies, people value efforts and plans GLOBE findings with respect to organizational
to develop family-like relational bonding and practices organizational values, interindustry
believe that these family-like relationships differences, and interorganizational variability.
provide an insurance against any future con-
tingencies and, thus, are less likely to practice Organizational Practices: As shown in
planning for the future. Moreover, Middle Table 13.13, the mean Future Orientation
Eastern and Latin American societies have a organizational practices score is 4.61, which
multifocus or polychronic orientation toward is higher than the mean of Future Orientation
the future (Sharda & Miller, 2001). They would societal practices score of 3.85 (see Table 13.4).
like to keep future well-being in mind, but in Respondents perceive their organizations to
practice are highly responsive to situational be more future oriented in practice than their
demands. They may, for instance, sit for several societies.
hours chatting with friends of friends just to
strengthen the bonds with their friends. Organizational Values: As shown in Table 13.13,
Furthermore, a high priority on spiritual ori- the mean Future Orientation organizational
entation and a holistic view toward matter and values score is 5.66, which is only marginally
spirit could account for high values scores of higher than the mean Future Orientation societal
Future Orientation in the Sub-Saharan Africa values score of 5.48 (see Table 13.4).
and Southern Asia clusters. Finally, traditional
reliance on government planning could explain Interindustry Differences: As shown in
low practices scores for Future Orientation in Table 13.13, the means of Future Orientation
the Eastern Europe cluster. Also, the turbulent organizational practices in the three industries—
post-Soviet environment has possibly con- financial, food processing, and telecommunica-
tributed to difficulties in practicing future plan- tions—are not significantly different from one
ning, making people less confident of their another.
abilities to shape their future through advance
planning and deliberated decisions. Interorganizational Variability: Finally, as is
Interestingly, the Future Orientation societal evident from Table 13.13, standard deviations
practices and values scores for the Confucian of organizational Future Orientation scores are
13-House.qxd 3/19/2004 5:20 PM Page 324
Standard
Variable N Min Max Mean Deviation
much higher for the reported practices than conducted two hierarchical linear models
for values. This is true for the overall as well as (HLMs) to test these hypotheses for organiza-
industry-specific scores. The respondents to the tional Future Orientation practices and values.
organizational questionnaire share a strong con- We tested the GLOBE hypothesis regarding
sensus on the value of Future Orientation, but the effect of societal culture on organizational
they vary substantially more in their reports culture by conducting HLM analyses in which
of Future Orientation practices. The differences organizational Future Orientation was predicted
in ability to use Future Orientation effectively by societal Future Orientation. These analyses
could have significant influence on the differ- supported our hypotheses that societal Future
ences in competitive advantage of the organiza- Orientation practices have a significant and
tions, especially in light of the high value given strong positive relationship with organizational
to Future Orientation by the sampled organiza- Future Orientation practices ( p < .01). We
tions. Further research is needed to investigate found a similar significant and strong relation-
barriers to future planning in many organiza- ship between societal Future Orientation values
tions around the world, and to disseminate the and organizational Future Orientation values
best practices in this regard. Increased uncer- (p < .01). Both analyses support a principal
tainty and volatility in the environments, proposition in the GLOBE theoretical model
because of globalization and crisis-like situa- (i.e., Proposition 3, Figure 2.1, Chapter 2, by
tions, could be partly responsible for the House & Javidan): Societal cultural values and
difficulties in using traditional approaches to practices affect organizational cultural values
planning. and practices.2
1 7
Greatly Contributes
Inhibits 2 3 4 5 6 Greatly to
Leadership Leadership Somewhat Slightly No Contributes Contributes Leadership
Attributes Effectiveness Inhibits Inhibits Impact Slightly Somewhat Effectiveness
Inspirational Inspires emotions, beliefs, values, and behaviors of others; inspires others to be
motivated to work hard
Anticipatory Anticipates, attempts to forecast events; considers what will happen in the future
Intellectually Encourages others to think and use their minds; challenges beliefs, stereotypes and
stimulating attitudes of others
effective leadership. One of the scales is Some notable societies on the top end of
Charismatic I: Visionary, or simply Visionary band A include Canada, the United States, the
Leadership. This scale consists of nine items, Philippines, Israel, and Ireland. Visionary
shown in Table 13.14, and reflects visionary leadership directed at transformational initia-
leaders’ readiness for future events, their tives has been widely identified to be strongly
emphasis on anticipating future events, and their effective in the United States and the Canadian
capability to do so successfully with experience, context (Bass, 1998). The Philippines experi-
their imaginativeness and vision, their goal- enced a substantial exposure to the American
based planning and action behavior, and their model of leadership during the 20th century,
ability to inspire and motivate others to work which could account for its high emphasis on
hard. This leadership scale is one of six others visionary leadership. Both Israeli and Irish
that comprise the global leadership dimension people live by a strong emphasis on building a
labeled Charismatic/Value-Based and will be unique identity for their nations, which may
discussed later. translate into a high endorsement of visionary
leadership.
Some less expected societies at the top
Society Scores on Visionary Leadership
include Ecuador (with the highest mean of
Table 13.15 shows the scores on Visionary 6.50), Colombia, and Iran. In these societies,
Leadership for the 61 societies in the GLOBE deregulation, trade liberalization, privatization,
sample. The societies are grouped into four and globalization have resulted in slow progress
bands, with band A consisting of significantly by comparable regional and international stan-
higher scores, and band D consisting of signifi- dards. Yet, strong charismatic leaders have been
cantly lower scores. able to offer an alternative vision to people in
13-House.qxd 3/19/2004 5:20 PM Page 326
Band
A B C D
a Canada (English-speaking)
b South Africa (White Sample)
c Switzerland (French-speaking)
d Germany (West): Former FRG
e Germany (East): Former GDR
f South Africa (Black Sample)
these societies. For instance, the Ayatollah leading a popular revolution that overthrew the
Ruhollah Khomeini, spiritual leader of Iran U.S.-supported monarch, Mohammed Reza
from 1979 to 1989, is believed by many to have Shah, who had been ruling Iran since 1953
displayed a charismatic vision in single-handedly (Shirley, 1994).
13-House.qxd 3/19/2004 5:20 PM Page 327
Standard
Variable Mean Deviation Minimum Maximum Valid N
Charismatic I: Visionary
overall societal score 6.02 0.36 4.62 6.50 61
Charismatic I: Visionary
finance industry 6.03 0.42 4.03 6.49 56
Charismatic I: Visionary
food industry 6.04 0.32 4.71 6.55 45
Charismatic I: Visionary
telecommunications
industry 6.06 0.33 5.02 6.45 32
At the bottom, there are three societies in the GLOBE sample, ranging from 6.03 (finance
band D. Qatar has the lowest score of 4.62, fol- sector) to 6.06 (telecommunications sector).
lowed by Morocco (4.84) and France (5.06). Further, as can be seen from Table 13.17, the
In Qatar, for several decades nearly four fifths correlations between the three industries and the
of the workforce has consisted of expatriates, overall score are both significant and large.
which may make the enactment of visionary
leadership quite difficult (Ali, 1995). Similarly,
Societal Demographics
several scholars have underlined French dis-
trust of visionary leaders, and their emphasis
and Visionary Leadership
on bureaucratic rules that minimize the possi- To explore the impact of societal demo-
bilities of any person imposing his or her view graphics, we examined the relationship between
of the future (see, for instance, Crozier, 1964). Visionary Leadership and two variables: Physi-
We may also note that even the societies in cal climate and regional clusters. Although we
band D report the effectiveness of Visionary found no significant differences in Visionary
Leadership at above the midpoint of 4.0, reflect- Leadership scores across different climate types
ing the universal reported effectiveness of (p > .10), regional clusters did explain a signifi-
visionary leadership. cant 32.7% (F = 2.75, p < .05) of the between-
society differences in Visionary Leadership.
Table 13.18 presents grand means for
Overall Scores and Industry
Visionary Leadership across the 10 regional
Scores on Visionary Leadership
clusters. The Analysis of Variance Contrast
Table 13.16 shows the overall and industry tests indicated that Visionary Leadership is
averages across all societies sampled. On a 7- endorsed significantly more strongly in the
point scale, the average is 6.02, which is very Anglo cluster at 6.26 (contrast = .27, p < .05),
high, reflecting the participants’ belief that and significantly less strongly in the Middle
Visionary Leadership is a highly effective East cluster at 5.49 (contrast = −.57, p < .01).
leader behavior. The universal view among the Indeed, Anglo societies are known for their
GLOBE sample is that charismatic visionary visionary leadership, possibly as a transforma-
leaders are effective leaders. tional framework for directing short-term profit
Table 13.16 also shows that the mean scores oriented behaviors and to allow for individual
for the Visionary Leadership dimension are autonomy (Bass, 1998). On the other hand,
almost identical across the three industries in Khadra (1990) observes that in the Middle East
13-House.qxd 3/19/2004 5:20 PM Page 328
Visionary Telecommunication
Leadership Food Industry Score Finance Industry Score Industry Score
* p < .05.
** p < .01.
the commonly effective form of leadership is of goals and plans, and inspiring behaviors that
the “caliphal model” based on an authoritarian go beyond the stereotypes, visionary leadership
leadership. should theoretically be more effective if a soci-
etal culture is future oriented.
Societal and Organizational Table 13.19 shows the correlation between
Future Orientation Culture GLOBE Future Orientation society practices
and society values scales and the Visionary
and Visionary Leadership
Leadership scale. Both correlations are near
Because the concept of visionary leadership is zero and are nonsignificant, suggesting that
rooted in anticipation of the future, development leaders who are able to anticipate the future,
13-House.qxd 3/19/2004 5:20 PM Page 329
Table 13.19 Correlation Between Future Orientation and Charismatic I: Visionary Leadership
plan ahead, have foresight, and are well prepared GLOBE FUTURE ORIENTATION AND
for the future may be effective in societies with CULTURALLY ENDORSED LEADERSHIP
a varying range of societal Future Orientation
practices and values. Visionary leadership Below we examine relationships between
may offer a framework within which sponta- Future Orientation scores and the six CLTS.
neous behaviors may be enacted successfully Competitive tests of all culture dimensions and
and rewards ensured (in less future-oriented CLTs are presented in Chapter 21 by Dorfman
societies), as well as a framework for long-term and colleagues. In general, we expect that soci-
planning and goal-oriented behavior (in more etal and organizational values will be more
future-oriented societies). strongly related to CLT leadership dimensions
Within the organizational context, visionary than societal and organizational practices. As
leadership may be particularly relevant for nur- indicated previously, our notions of values and
turing a future-oriented culture in organizations, CLT leadership dimensions represent idealized
as described in the earlier review section on the concepts of how the world “Should Be” in con-
correlates of organizational Future Orientation. trast to practices that represent the world “As
To test this hypothesis, we aggregated reported Is.” As you read through the results discussed
Visionary Leadership scores across respondents below, it may be helpful to view Figure 13.1 for
for each organization separately. For a sample a visual summary. The figure, however, only
of 271 organizations, data were available for shows results regarding cultural values, not
both the Visionary Leadership scale as well as practices. (All HLM coefficients are presented
for GLOBE’s organizational Future Orientation in Table 21.10 of Chapter 21 by Dorfman et al.)3
values and practices scales. We tested for the relationship between
As reported in Table 13.19, the Visionary culture and the CLT leadership dimensions
Leadership scale had no correlation with the by using hierarchical linear modeling (HLM).
organizational Future Orientation practices An overview of HLM analyses and a detailed
scale (r = −.09). However, the correlation with discussion of how we conducted these analyses,
the organizational Future Orientation values as well as how to interpret the R2 information
scale is both positive and significant (r = .35, in HLM, is provided in Chapter 11 by Hanges,
p < .01). The respondents who value more Dickson, and Sipe. In this next section, we dis-
Future Orientation in their organizations also cuss the results of the HLM analyses exploring
tend to believe in the effectiveness of Visionary the relationship between organizational and
Leadership. Thus, one possible way in which societal culture and the content of CLTs.
Visionary Leadership can be promoted in orga- In the present analyses we examine the simul-
nizations may be through a strategic institution- taneous associations of organizational and soci-
alization of a culture of future orientation. etal GLOBE Future Orientation scores in a single
Next, we examine how future orientation at analysis. The total amount of organizational and
the society and organizational levels influences societal variance explained by Future Orientation
alternative forms of culturally implicit theories scales ranged from 0–19.2%. Future Orientation
of leadership (CLTs). values or practices were found to be more
13-House.qxd 3/19/2004 5:20 PM Page 330
Charismatic/VB (O)
strongly associated with some of the CLTs (e.g., leadership dimension and explained a total of
Team-Oriented leadership) than for others (e.g., 19.2% of the organizational and societal vari-
Autonomous leadership). More specifically, ance for this dimension. Approximately 37.7%
of this total explained variance was associated
• Charismatic/Value-Based leadership. with forces operating at the organizational level
Future Orientation cultural practices and values of analysis. The remaining portion of the
scores were significantly related to the explained variance (62.3%) was associated
Charismatic/Value-Based leadership dimension with forces operating at the societal level of
and explained a total of 6.7% of the organiza- analysis.
tional and societal variance for this dimension.
All of this explained variance was associated The organizational Future Orientation cul-
with forces operating at the organizational tural values scores were positively related
level of analysis. The organizational Future ( p < .01) to the Team-Oriented leadership
Orientation cultural values scores were posi- dimension. That is, Team-Oriented leadership is
tively related ( p < .01) to the Charismatic/ more likely to be a part of a shared leadership
Value-Based leadership dimension. Charismatic/ belief system in organizations reported to
Value-Based leadership is more likely to be espouse Future Orientation values. The societal
a part of a shared leadership belief system Future Orientation cultural practices scores
in organizations reported to espouse Future were negatively related (p < .05) to the Team-
Orientation values. Oriented leadership dimension. Team-Oriented
• Team-Oriented leadership. Future Orien- leadership is less likely to be a part of the shared
tation cultural practices and values scores were leadership belief system in societies reported to
significantly related to the Team-Oriented espouse Future Orientation practices.
13-House.qxd 3/19/2004 5:20 PM Page 331
almost universally valued. Across the entire control may have contributed to perceptibly
study, on a 7-point scale, the GLOBE measure weaker Future Orientation practices reported by
of Future Orientation as a societal value scored the telecommunications sector in some societies
5.48, whereas the Future Orientation in prac- such as Thailand. The negative correlation
tices score was 3.85. The respondents in each between practices and values of societal Future
society in the GLOBE sample, with the sole Orientation is very prominent for the telecom-
exception of Denmark, value more Future munications sample. The societies in which this
Orientation than they report actually practicing. industry was traditionally governed by short-
Also very important, in this chapter we have term practices, due to such factors as protective
demonstrated that organizations reflect the government regulation, appear to be acutely
culture (practices and values) in the society in aware of the need to reform and to adopt more
which they are embedded. Societies that are strategic approaches for infrastructure improve-
future oriented likely have organizations within ment and development. The GLOBE data show
them that are also future oriented. a strong cross-industry convergence in Future
We found a strong negative correlation Orientation values and practices.
between Future Orientation practices and val- Several societies with relatively high scores
ues. The respondents in societies with weaker on the GLOBE Future Orientation practices
practices of Future Orientation have stronger scale are reported to have well-developed col-
aspirations for Future Orientation. It may be that lective institutions. In these societies, the lead-
the societies lacking future oriented practices ers have confidence in the abilities of their
suffer most from the uncertainty and unpre- members and in the collective safety net of
dictability of not addressing the longer-term their institutions, and they expect their members
fundamental issues. Therefore, respondents in to be more innovative and tolerant of change.
such societies are most conscious of the need for As illustrated by the case of some Nordic and
moving toward a more strategic perspective. Germanic societies in the GLOBE sample,
Surprisingly, most industrialized and higher- under these conditions more entrepreneurial
income countries have comparatively low leadership may be promoted and power struc-
scores on Future Orientation values. In contrast, tures may be dismantled.
emerging and lower-income markets score high Similarly, the societies with stronger aspira-
on Future Orientation values. These findings tions for Future Orientation also value collective
contradict the popular view in the literature that safety nets of the institutions and of families and
lower socioeconomic status encourages a focus relationships, and seek to lower the uncertainty in
on immediate concerns for survival, whereas their environments. Here the role of women in
higher socioeconomic status allows savings and supporting families becomes even more salient.
investments for the future. It is likely that indus- As illustrated by the case of many Latin American
trialized and higher-income nations enjoy the societies in the GLOBE sample, the women in
present more and are not overly concerned such societies may become a vulnerable group
about a future orientation because they have engaged primarily in microfamily issues, with
already accumulated substantial wealth and limited capacity or opportunity to participate in
material resources. The emerging and lower- macrolevel strategic planning. Further, in some
income nations may, on the other hand, see a societies, such as in the Middle East, people may
stronger need for taking a long-term perspective aspire toward future orientation to reform the
and sacrificing for the future because they must authoritarian, kinship-oriented, and fragmented
cope with scarce and limited resources. institutional fabric in their cultures.
The means of societal Future Orientation Interestingly, neither GLOBE Future Orient-
practices and values are not significantly differ- ation practices (rho = .03) nor GLOBE Future
ent across the three industries sampled in the Orientation values (rho = −.06) scales show
GLOBE project (food processing, financial ser- any relationship with Hofstede’s Confucian
vices, and telecommunications). Still, tradi- Dynamism scale. Also, contrary to the predic-
tional stability in the market structure and lack tions of Hofstede and Bond (1998), there is no
of visionary leadership under government evidence of Confucian societies scoring highest
13-House.qxd 3/19/2004 5:20 PM Page 333
on Future Orientation practices, or Anglo The analysis of regions showed that most
societies scoring the least. China, for instance, Nordic and Germanic societies report strong
scores less than Canada, the United States, and practices, but weak values, of Future Orien-
the United Kingdom on reported Future tation. These findings may be attributable to
Orientation practices. high levels of political stability and socio-
Still, the GLOBE Future Orientation values economic status, and a great emphasis on pro-
scale does appear to capture the belief in quid fessional bureaucracy in the Northern European
pro quo and long-term interaction inherent in region. The Northern European societies tend to
Schwartz’s cultural embeddedness scale, as have well laid-out procedures for long-term
opposed to contractual transience and spot rela- planning. At the same time, improvising by pro-
tionships associated with Schwartz’s cultural fessionals to deal with emergent opportunities
autonomy scale (1999). Further, the societies is well respected and valued. In contrast, most
that are reported to be future oriented are actu- Latin American and Middle Eastern societies
ally more likely to encourage individuals to report strong values, but weak practices, of
express themselves and to lead a high quality Future Orientation. In these societies, people
of life, as evidenced by a strong positive corre- value long-term relationships but often believe
lation of GLOBE societal Future Orientation that the relationships can help one deal with any
practices scale with Inglehart and Baker’s unplanned future contingencies, and people are
(2000) survival versus self-expression scale therefore less likely to plan ahead.
(r = .61, p < .01). Conversely, in societies that Societies in Sub-Saharan Africa and Southern
depend more on traditional authority or on sur- Asia stand out for their high value on Future
vival motivation, the value of future orientation Orientation, possibly reflecting the high priority
becomes more salient. they set on spiritual orientation and a holistic
The societies reported to have stronger prac- view favoring matter that is integrated with spirit
tices of future orientation tend to show better on a long-term basis. On the other hand, Eastern
economic and societal health, more scientific Europe is distinctive in its low reported Future
advancement, more democratic political ideals, Orientation practices, which may partly indicate
more empowered gender status, and greater the turbulent and transitional conditions in that
domestic savings. Under such conditions, fam- region during the 1990s.
ily and friends may play a less prominent role in Consistent with other related findings, societal
the life of people. On the other hand, the aspira- practices and values of future orientation for the
tions for future orientation are stronger in soci- Confucian Asia and Anglo clusters are not sig-
eties that have weak economic and societal nificantly higher or lower than the other clusters,
health, less scientific advancement, less democ- and are similar to the international average. This
ratic political ideals, and less empowered finding casts doubt the effectiveness of measur-
gender status. These aspirations for future ori- ing the construct of future orientation using the
entation are interrelated with a strong emphasis Confucian Dynamism scale (Hofstede, 2001).
on family and friends and on spiritual orienta- At the organizational level, practices are
tion. Family, friends and spiritual orientation reported to be more future oriented within orga-
may offer critical support under weak eco- nizations than throughout societies. In addition,
nomic, societal, scientific, political, and gender the respondents who report their societies to have
conditions. Future Orientation practices are also likely to
Physical climate plays an important role in report their organizations as practicing future ori-
defining the societal values of Future Orienta- entation. Similarly, the Future Orientation values
tion. Societies in the relatively unpredictable at the organizational level are strongly and direc-
nontropical maritime and continental climates tly correlated with the Future Orientation values
tend to report lower values of Future Orienta- at the societal level (r = .52, p < .01), although
tion, whereas respondents from the societies the values for future orientation are reported to
in more predictable tropical wet and dry (i.e., be equally strong for both organizations as well
savanna) climates report greater preference for as societies. The reported future orientation of
Future Orientation. organizations does not vary by industry.
13-House.qxd 3/19/2004 5:20 PM Page 334
The organizations sampled in the GLOBE chose to emphasize the relationships of CLT
program share a strong consensus on the value leadership dimensions with cultural values
of Future Orientation, but they vary substan- rather than practices for reasons discussed pre-
tially more in their abilities to currently practice viously. Both cultural values and CLT leader-
it. It is probable that differences in ability to ship dimensions reflect idealized constructs
effectively use future orientation could have rather than actual cultural practices or leader-
significant influence on the differences in com- ship behaviors. When organizations are viewed
petitive advantage of the organizations, espe- as valuing Future Orientation, the CLT
cially in light of the high value given to Future endorsed at the organizational level is more
Orientation by the sampled organizations. likely to be composed of Participative, Humane-
The organizations that report stronger values Oriented, Team-Oriented, and Charismatic/
of Future Orientation are more likely to endorse Value-Based leadership styles. Interestingly,
the effectiveness of Visionary Leadership, societal Future Orientation values are also
though the reported organizational practices associated with preferences for a more Self-
of Future Orientation are not associated with Protective leadership. Viewed in its totality, this
the endorsement of Visionary Leadership. cultural dimension was related to five of the six
Similarly, visionary leaders who are able to CLT leadership dimensions, and was second
anticipate the future, plan ahead, have foresight, only to Performance Orientation in its impor-
and are well-prepared for the future may be tance to idealized leadership styles.
effective in societies with varying ranges of
societal Future Orientation practices and values.
Visionary Leadership is endorsed signifi- ENDNOTES
cantly more strongly in the Anglo cluster at
6.26 (contrast = .27, p < .05), and significantly 1. Our response bias correction procedure identi-
less strongly in Middle East cluster at 5.49 (con- fied response bias in some countries for this scale.
trast = −.57, p < .01). Anglo societies are known We recomputed the predicted response bias corrected
for their visionary leadership, possibly as a scale score for each country. Response bias corrected
transformational framework for directing short- scores are:
term profit-oriented behaviors and to allow for Practices: France, 3.74 (no change in band);
individual autonomy. Israel and Ireland also Qatar, 4.08 (moves from band C to band B); and
report high scores on Visionary Leadership: Taiwan, 3.65 (moves from band B to band C).
Both Israeli and Irish people live by a strong Values: Ecuador, 5.62 (moves from band A to
vision of promoting a unique identity for their band B); France, 5.35 (moves from band C to band
nations, which may translate into a high B); Morocco, 6.33 (no change in band); and New
endorsement of visionary leadership. Zealand, 5.90 (moves from band B to band A).
On the other hand, in the Middle East, For a complete discussion of this procedure and
authoritarian leadership is commonly more all response bias corrected scores, see Appendix C,
effective. In Qatar, for instance, nearly four this volume.
fifths of the workforce has been composed of 2. As reported in Chapter 20 by Brodbeck,
other countries’ expatriates for several decades, Hanges, Dickson, Gupta, and Dorfman, we found that
which may make the enactment of visionary all the cultural dimensions of organizational cultural
leadership quite difficult (Ali, 1995). France values and practices significantly differed across soci-
also stands out for its low endorsement of eties. Although important, this prior analysis did not
visionary leadership. The French often distrust identify the particular aspect of societal differences
visionary leaders and emphasize bureaucratic that was related to organizational culture. In the pre-
rules that minimize the possibilities of any per- sent chapter, we found that societal and organizational
son pushing his or her idiosyncratic view of the Future Orientation practices were significantly related
future. (R2 Total = 3.9%, R2 Societal = 42.5%, p < .01). We
We also examined in depth the relationships also found strong results for societal and organiza-
between Future Orientation and dimensions of tional Future Orientation values (R2 Total = 4.6%, R2
leadership identified in the GLOBE study. We Societal = 32.9%, p < .01). As discussed in Chapter 11
13-House.qxd 3/19/2004 5:20 PM Page 335
by Hanges, Dickson, and Sipe, the R2 Total considers Barakat, M., & Sarver, E. (1997, January 30).
all levels of analysis (i.e., individual, organizational, Western banks taking 1st steps into Islam’s “no
and societal) whereas the R2 Societal isolates the interest” world. American Banker, 162(20), 9.
societal-level portion of the dependent variable and Barndt, R. J., & Johnson, D. M. (1955). Time orien-
indicates the percentage of variance accounted for by tation in delinquents. Journal of Abnormal and
the predictor at only this level. Although we have Social Psychology, 51, 343–345.
primarily taken the conservative approach and reported Barrig, M., & Beckman, E. (2001). Latin American
the R2 Total in GLOBE, several scholars suggest feminism. NACLS Report on the Americas,
that R2 Societal provides a more accurate description 34(5), 29–36.
of aggregated relationships. For further discussion, see Bass, B. M. (1998). Transformational leadership:
the paper by Lance and James (1999). Industrial, military, and educational impact.
3. Results between the single HLM and multiple Mahwah, NJ: Lawrence Erlbaum.
HLM tests will likely differ somewhat. The differences Becker, G. (1965). A theory of the allocation of time.
between the results of the multiple HLMs and single The Economic Journal, 75, 493–517.
HLMs are conceptually similar to the differences Benson, J. B. (1994). The origins of future orientation
between a multiple regression analysis and a correla- in the everyday lives of infants and toddlers.
tion coefficient. Table 21.10 in Chapter 21 by Dorfman In M. M. Haith, J. B. Benson, R. R. Roberts, &
and colleagues presents both single and multiple HLM B. Pennington (Eds.), The development of
coefficients. In addition, the relationships for all cul- future-oriented processes (pp. 375–407).
ture dimension values are summarized in Chapter 3. Chicago: University of Chicago Press.
Bentham, J. (1970). An introduction to the principles
of morals and legislation. London: Methuen.
REFERENCES (Original work published 1789)
Blasko, J. C. (1998). Overcoming the legal and his-
Agarwal, A. (1993). Time, memory, and knowl- torical obstacles to privatization: The telecom-
edge representation: The Indian perspective. In munications sector in Thailand. Case Western
J. Altarriba (Ed.), Cognition and culture: A cross- Reserve Journal of International Law, 30(2/3),
cultural approach to psychology (pp. 45–55). 507–539.
Amsterdam: Elsevier Science. Blinn, L. M., & Pike, G. (1989). Future time perspec-
Agarwal, A., Tripathi, K. K., & Srivastava, M. tive: Adolescents’ predictions of their interper-
(1983). Social roots and psychological implica- sonal lives in the future. Adolescence, 24(94),
tions of time perspective. International Journal 289–301.
of Psychology, 18(5), 367–380. Bluedorn, A. C. (2000). Time and organizational
Ali, A. J. (1995). Cultural discontinuity and Arab culture. In N. M. Ashkanasy, C. P. Wilderom, &
management thought. International Studies of M. F. Peterson (Eds.), Handbook of organiza-
Management & Organization, 25(3), 7–30. tional culture and climate (pp. 117–128).
Alvos, L., Gregson, R. A. M., & Ross, M. W. (1993). Thousand Oaks, CA: Sage.
Future time perspective in current and previous Boocock, S. S. (1978). The social organization of the
injecting drug users. Drug and Alcohol Depen- classroom. Annual Review of Sociology, 4, 1–28.
dence, 31(2), 193–197. Brandt, R. B. (1995). Morality, utilitarianism and
Ansoff, H. I. (1988). The new corporate strategy. rights. Cambridge, UK: Cambridge University
New York: John Wiley. Press.
Antonakis, J., & House, R. J. (2002). An analysis of Brommer, M., & De La Porte, R. (1992). A context
the full-range leadership theory: The way for- for envisioning the future. National Productivity
ward. In B. J. Avolio & F. J. Yammarino (Eds.), Review, 11, 549–552.
Transformational and charismatic leadership: Bryman, A. (1996). Leadership in organizations. In
The road ahead (pp. 3–33). New York: JAI/ S. Clegg, C. Hardy, & W. Nord (Eds.),
Elsevier Science. Handbook of organization studies (pp. 276–292).
Ariarajah, S. W. (1986). Hindu spirituality: An invi- London: Sage.
tation to dialogue? Ecumenical Review, 38(1), Burns, T., & Stalker, G. M. (1961). The management
75–81. of innovation (1st ed.). London: Tavistock.
13-House.qxd 3/19/2004 5:20 PM Page 336
Cahill, T. (1998). The gifts of the Jews. New York: Doob, L. W. (1971). The patterning of time. New
Doubleday. Haven, CT: Yale University Press.
Cameron, P., Desai, K. G., Bahador, D., & Dremel, G. Dutton, J. E., & Jackson, S. E. (1987). Categorizing
(1977–1978). Temporality across the life-span. strategic issues: Links to organizational action.
International Journal of Aging and Human Academy of Management Review, 12, 76–90.
Development, 8, 229–259. Edwards, S. (1995). Why are saving rates so different
Carstensen, L. L. (1993). Motivation for social across countries? An international comparative
contact across the life span: A theory of socio- analysis (Working Paper No. 5097). Cambridge,
emotional selectivity. Nebraska Symposium on MA: NBER.
Motivation, 40, 209–254. Eisenhardt, K. M. (1989). Making fast strategic
Chandler, A. D., Jr. (1962). Strategy and structure: decisions in high-velocity environments.
Chapters in the history of the industrial enter- Academy of Management Journal, 32, 543–576.
prise. Cambridge: MIT Press. Ekecrantz, J. (2001, August). Postmodern times? A
Cohen, J. (1967). Psychological time and health and comparative study of temporal constructions.
disease. Springfield, IL: Charles C Thomas. Paper presented at the 15th Nordic Conference
Cohen, W. M., & Levinthal, D. A. (1990). Absorptive on Media and Communication Research,
capacity: A new perspective on learning and Reykjavík, Iceland.
innovation. Administrative Science Quarterly, Eng, P. (1997). Thai democracy: The people speak.
35,128–152. Washington Quarterly, 20(4), 169–189.
Cottle, T. (1967). The Circles Test: An investigation Ferrari, J. R., & Emmons, R. A. (1995). Methods of
of perception of temporal relatedness and procrastination and their relation to self-control
dominance. Journal of Projective Technique and and self-reinforcement: An exploratory study.
Personality Assessments, 31, 58–71. Journal of Social Behavior and Personality,
Cottle, T. J., & Klineberg, S. L. (1974). The present 10(1), 135–142.
of things future. New York: Free Press. Fiol, C. M., Harris, D., & House, R. J. (1999).
Cross, T. (2001). Spirituality and mental health: A Charismatic leadership: Strategies for affecting
Native American perspective. Focal Point, social change. Leadership Quarterly, 10,
15(2), 37–38. 449–482.
Crozier, M. (1964). The bureaucratic phenomenon. Fraisse, P. (1963). The psychology of time (J. Leith,
Chicago: Chicago University Press. Trans.). Westport, CT: Greenwood.
Daft, R. E., & Weick, K. E. (1984). Towards a model Frank, L. K. (1939). Time perspectives. Journal of
of organizations as interpretation systems. Social Philosophy, 4, 293–312.
Academy of Management Review, 9, 284–295. Franke, R. H., Hofstede, G., & Bond, M. H. (1991).
Das, T. K. (1986). The subjective side of strategy Cultural roots of economic performance: A
making: Future orientations and perceptions of research note. Strategic Management Journal,
executives. New York: Praeger. 12, 165–173.
Das, T. K. (1987). Strategic planning and individual Fredrickson, J. W. (1984). The comprehensiveness of
temporal orientation. Strategic Management strategic decision processes: Extension, obser-
Journal, 8, 203–209. vations, future directions. Academy of Manage-
D’Aveni, R. A. (1994). Hyper-competition. New York: ment Journal, 27, 445–466.
Free Press. Fung, H. H., Carstensen, L. L., & Lutz, A. M. (1999).
Davis, E. P., & Steil, B. (2001). Institutional Influence of time on social preferences: Impli-
investors. Cambridge: MIT Press. cations for life-span development. Psychology
Davids, A., & Falkof, B. B. (1975). Juvenile delin- and Aging, 14, 595–604.
quents then and now: Comparison of findings Gaddis, J. L. (1997). We now know: Rethinking cold
from 1959 and 1974. Journal of Abnormal war history. Oxford, UK: Clarendon Press.
Psychology, 84, 161–164. Gallimore, R., Weiss, L. B., Finney R., & Tharp,
Davids, A., Kidder, C., & Reich, M. (1962). Time ori- R. G. (1974). Cultural differences in delay
entation in male and female juvenile delinquents. of gratification: A problem of behavior classi-
Journal of Abnormal and Social Psychology, fication. Journal of Personality & Social
64, 239–240. Psychology, 30, 72–80.
13-House.qxd 3/19/2004 5:20 PM Page 337
Inglehart, R., Basanez, M., & Moreno, A. (1998). Journal of International Business Studies, 11,
Human values and beliefs: A cross-cultural source- 32–47.
book. Ann Arbor: University of Michigan Press. Kotter, J. P. (1982). The general managers.
Jaeger, A., & Kanungo, R. (1990). Introduction: The New York: Free Press.
need for indigenous management in developing Kouzes, J. M., & Posner, B. Z. (1996). Envisioning
countries. In A. Jaeger & R. Kanungo (Eds.), your future: Imagining ideal scenarios. The
Management in developing countries (pp. 1–23). Futurist, 30(3), 14–20.
New York: Routledge. Lance, C. E., & James, L. R. (1999). A proportional
Johnson, H. T., & Kaplan, R. S. (1987). Relevance variance accounted for index for some
lost. Boston: Harvard Business School Press. cross-level and person-situation research designs.
Judge, W. Q., & Spitzfaden, M. (1995). The Organizational Research Methods, 2, 395–418.
management of strategic time horizons within Lang, F. R., & Carstensen, L. L. (1994). Close emo-
biotechnology firms: The impact of cognitive tional relationships in late life: Further support
complexity on time horizon diversity. Journal of for proactive aging in the social domain.
Management Inquiry, 4(2), 179–196. Psychology and Aging, 9, 315–324.
Kagono, T., Nonaka, I., Sakakibara, K., & Okumura, Lang, F. R., & Carstensen, L. L. (2002). Time counts:
A. (1985). Strategic vs. evolutionary manage- Future time perspective, goals, and social relation-
ment: A U.S.-Japan comparison of strategy and ships. Psychology and Aging, 17(1), 125–139.
management. Amsterdam: North-Holland. Larwood, L., Falbe, C. M., Kriger, M., & Miesing, P.
Kant, I. (1965). Critique of pure reason. London: (1995). Structure and meaning of organizational
Macmillan. (Original work published 1781) vision. Academy of Management Journal, 38,
Kastenbaum, R. (1961). The dimensions of future 740–769.
time perspective: An experimental analysis. Lasane, T. P., & Jones, J. M. (1999). Temporal ori-
Journal of General Psychology, 65, 203–218. entation and academic goal-setting: The mediat-
Kastenbaum, R. (1964). Cognitive and personal ing properties of a motivational self. Journal of
futurity in later life. Journal of Individual Social Behavior and Personality, 14, 31–44.
Psychology, 19, 216–219. Lauenstein, M. C. (1985). SMR Forum:
Keegan, W. J. (1983). Strategic market planning: The Diversification—The hidden explanation of suc-
Japanese approach. International Marketing cess. Sloan Management Review, 27(1), 49–55.
Review, 1, 1–15. Laverty, K. J. (1996). Economic “short-termism”:
Keough, K. A., Zimbardo, P. G., & Boyd, J. N. (1999). The debate, the unresolved issues, and the impli-
Who’s smoking, drinking, and using drugs? Time cations for management practice and research.
perspective as a predictor of substance use. Basic Academy of Management Review, 27, 825–860.
and Applied Social Psychology, 21(2), 149–164. Lay, C. H., & Schouwenburg, H. C. (1993). Trait pro-
Khadra, B. (1990). The prophetic-caliphal model of crastination, time management, and academic
leadership: An empirical study. International behavior. Journal of Social Behavior and
Studies of Management and Organization, 20, Personality, 8(4), 647–662.
37–51. Leach, E. R. (1961). Two essays concerning the sym-
Klineberg, S. L. (1967). Changes in outlook on the bolic representation of time. In Rethinking anthro-
future between childhood and adolescence. pology (pp. 124–136). London: Athlone Press.
Journal of Personality and Social Psychology, 7, Lewin, K. (1926). Intention, will and need. In
185–193. D. Rapaport (Ed.), Organization and pathology
Klineberg, S. L. (1968). Future time perspective and of thought (pp. 95–151). New York: Columbia
the preference for delayed reward. Journal of University Press.
Personality and Social Psychology, 8, 253–257. Lewin, K. (1942). Time perspective and morale. In
Kluckhohn, F. R., & Strodtbeck, F. L. (1961). G. Watson (Ed.), Civilian morale (pp. 48–70).
Variations in value orientations. Evanston, IL: Boston: Houghton Mifflin.
Row, Peterson. Lewin, K. (1951). Field theory in social science.
Kobrin, S., Basek, J., Blank, S., & La Palombara, J. New York: Harper & Brothers.
(1980). The assessment and evaluation of Lim, Y. M., & Seers, A. (1993). Time dimensions of
noneconomic environments by American firms. work: Relationships with perceived organizational
13-House.qxd 3/19/2004 5:20 PM Page 339
performance. Journal of Business and Psychology, and other correlates. Adolescence, 27(105),
8(1), 91–102. 173–181.
Loescher, S. M. (1984). Bureaucratic measure- Murrell, A. J., & Mingrone, M. (1994). Correlates of
ment, shuttling stock shares, and shortened time temporal perspective. Perceptual and Motor
horizons: Implications for economic growth. Skills, 78, 1331–1334.
Quarterly Review of Economics and Business, Nadler, A. (1975). Delay of gratification: Review and
24, 8–23. suggestions for future research. Catalog of
Lomranz, J., Shmotkin, D., & Katznelson, D. B. selected documents in psychology, 5(2), 79.
(1983). Coherence as a measure of future time North, D. C. (1990). Institutions, institutional change
perspective in children and its relationship and economic performance. Cambridge, UK:
to delay of gratification and social class. Cambridge University Press.
International Journal of Psychology, 18(5), Novaky, E., Hideg, E., & Kappeter, I. (1994). Future
407–413. orientation in Hungarian society. Futures, 26(7),
MacKinnon-Slaney, F. (1994). The adult persistence 759–770.
in learning model. Journal of Counseling and Nurmi, J. (1987). Age, sex, social class, and quality
Development, 72(3), 268–275. of family interaction as determinants of adoles-
Malinowski, B. (1926–1927). Lunar and seasonal cents’ future orientation: A developmental task
calendar in the Trobriands. Journal of the interpretation. Adolescence, 22, 977–991.
Anthropological Institute of Great Britain and Nurmi, J. (1991). How do adolescents see their
Ireland, 56–57, 203–215. future? A review of the development of future
Marko, K. W., & Savickas, M. L. (1998). Effectiveness orientation and planning. Developmental
of a career time perspective intervention. Journal Review, 11, 1–59.
of Vocational Behavior, 52, 106–119. Nuttin, J. (1964). The future time perspective
Matthews, D., & Ornauer, H. (1976). The year 2000 in human motivation and learning. Acta
questionnaire and the marginals of 11 nations. Psychologica, 23, 60–82.
In H. Ornauer, H. Wiberg, A. Sicigski, & O’Rand, A., & Ellis, R. A. (1974). Social class and
J. Galtang (Eds.), Images of the world in the social time perspective. Social Forces, 53, 53–62.
year 2000 (pp. 637–702). Atlantic Highlands, Oner, B. (2000). Future time orientation and relation-
NJ: Humanities Press. ships with the opposite sex. The Journal of
Mehta, P. H., Rohila, P. K., Sundberg, N. D., & Psychology, 134, 306–314.
Tyler, L. E. (1972). Future time perspective of Onken, M. H. (1999). Temporal elements of organi-
adolescents in India and the United States. zational culture and impact on firm perfor-
Journal of Cross-Cultural Psychology, 3, mance. Journal of Managerial Psychology,
293–302. 14(3/4), 231–244.
Mill, J. S. (1986). A system of logic. New York: Ouchi, W. (1981). How American business can
Classworks. (Original work published 1843) meet the Japanese challenge. Reading, MA:
Mintzberg, H. (1975). The manager’s job: Folklore Addison-Wesley.
and fact. Harvard Business Review, 53(4), Peetsma, T. T. D. (1993, August 31- September 5).
49–61. Future time perspective as an attitude: The val-
Mintzberg, H. (1979). The structuring of organiza- idation of a concept. Paper presented at the 5th
tions. Englewood Cliffs, NJ: Prentice Hall. Conference of the European Association for
Mischel, W. (1974). Processes in delay of gratifica- Research on Learning and Instruction (EARLI),
tion. In W. Mischel & L. Berkowitz (Eds.), Aix-en-Provence, France.
Advances in experimental social psychology Peetsma, T. T. D. (2000). Future time perspective as a
(pp. 249–288). San Diego, CA: Academic Press. predictor of school investment. Scandinavian
Molles, M. C. (2001). Ecology: Concepts and appli- Journal of Educational Research, 44(2), 177–192.
cations (2nd ed.). New York: McGraw-Hill. Pervin, L. A. (Ed.). (1989). Goal concepts in person-
Moore, W. (1963). Man, time and society. New York: ality and social psychology. Hillsdale, NJ:
Wiley. Lawrence Erlbaum.
Morris, G. B. (1992). Adolescent leaders: Rational Piaget, J. (1952). The origins of intelligence in children.
thinking, future beliefs, temporal perspectives New York: International Universities Press.
13-House.qxd 3/19/2004 5:20 PM Page 340
Poole, M. E., & Cooney, G. H. (1987). Orientations Rothspan, S., & Read, S. J. (1996). Present versus
to the future: A comparison of adolescents in future time perspective and HIV risk among het-
Australia and Singapore. Journal of Youth and erosexual college students. Health Psychology,
Adolescence, 16(2), 129–151. 15, 131–134.
Porter, M. E. (1992). Capital disadvantage: Rumelt, R. P. (1987). Theory, strategy, and entrepre-
America’s failing capital investment system. neurship. In D. J. Teece (Ed.), The competitive
Harvard Business Review, 70(5), 65–82. challenge (pp.137–158). New York: Harper &
Principles of governance. (2001). Retrieved January Row.
3, 2001, from the Singapore government website Sagie, A., Elizur, D., & Yamauchi, H. (1996). The
http://www1.moe.edu.sg/ne/About_NE/Govern structure and strength of achievement motiva-
ance/governance& - 9 5;principles.htm tion: A cross-cultural comparison. Journal of
Quinn, R. E., & McGrath, M. R. (1985). The transfor- Organizational Behavior, 17, 431–444.
mation of organizational cultures: A competing Schein, E. H. (1985). Organizational culture and
values perspective. In P. J. Frost, L. F. Moore, leadership. San Francisco: Jossey-Bass.
M. R. Louis, C. C. Lundberg, & J. Martin (Eds.), Schneider, S. C. (1989). Strategy formulation: The
Organizational culture (pp. 315–334). Beverly impact of national culture. Organization Studies,
Hills, CA: Sage. 10, 149–168.
Ralston, D. A., Gustafson, D. J., Elsass, P. M., Cheung, Schriber, J. B., & Gutek, B. A. (1987). Some time
F., & Terpstra, R. H. (1992). Eastern values: A dimensions of work: Measurement of an under-
comparison of managers in the United States, lying aspect of organizational culture. Journal of
Hong Kong, and the People’s Republic of China. Applied Psychology, 72(4), 642–650.
Journal of Applied Psychology, 77, 64–671. Schwartz, S. H. (1999). A theory of cultural values
Raynor, J. O. (1969). Future orientation and motiva- and some implications for work. Applied
tion of immediate activity: An elaboration of Psychology: An International Review, 48,
the theory of achievement motivation. Psycho- 23–47.
logical Review, 76, 606–610. Seginer, R., & Schlesinger, R. (1998). Adolescents’
Raynor, J. O. (1974). Future orientation in the study future orientation in time and place: The case of
of achievement motivation. In J. W. Atkinson & the Israeli kibbutz. International Journal of
J. O Raynor (Eds.), Motivation and achievement Behavioral Development, 22, 151–167.
(pp. 219–236). Washington, DC: Winston. Seijts, G. H. (1998). The importance of future time
Raynor, J. O., & Entin, E. E. (1982). Motivation, perspective in theories of work motivation. The
career striving, and aging. Washington, DC: Journal of Psychology, 132, 154–168.
Hemisphere. Sharda, B. D., & Miller, G. A. (2001). Culture and
Read, R. (1993). Politics and policies of national organizational structure in the Middle East: A
economic growth. Unpublished doctoral disser- comparative analysis of Iran, Jordan and the
tation, Stanford University, CA. USA. International Review of Sociology, 11(3),
Roede, E. (1989). Explaining student investment: An 309–324.
investigation of high school students’ retrospec- Shirley, E. G. (1994, Fall). The Iran policy trap.
tive causal accounts of their investment in Foreign Affairs, 96, 75–93.
school. Amsterdam: Stichting Centrum voor Shouval, R., Kav Venaki, S., Bronfenbrenner, U.,
Onderwijsonderzoek van de Universiteit van Devereux, E. C., & Kiely, E. (1986). Anomalous
Amsterdam. reactions to social pressure of Israeli and Soviet
Roer-Strier, D., & Rosenthal, M. K. (2001). children raised in family versus collective settings.
Socialization in changing cultural contexts: A In T. R. Horowitz (Ed.), Between two worlds:
search for images of the “adaptive adult.” Social Children from the Soviet Union in Israel (pp.
Work, 46, 215–228. 67–90). New York: University Press of America.
Rosenberg, M. J., & Hovland, C. I. (1960). Cognitive, Skinner, B. F. (1937). Two types of conditioned
affective, and behavioral components of atti- reflex: A reply to Konorski and Miller. Journal
tudes. In C. I. Hovland & M. J. Rosenberg (Eds.), of General Psychology, 16, 272–279.
Attitude organization and change (pp. 1–14). Sorokin, P., & Merton, R. (1990). Social time: A
New Haven, CT: Yale University Press. methodological and functional analysis. In
13-House.qxd 3/19/2004 5:20 PM Page 341
Hassard, J. (Ed.), The sociology of time Thrift, N. (1990). The making of capitalist time
(pp. 56–66). Basingstoke, UK: Macmillan. consciousness. In J. Hassard. (Ed.), The sociol-
Spector, P. E., Cooper, C. L., Sparks, K., Bernin, P., ogy of Time (pp. 105–129). Basingstoke, UK:
Büssing, A., Dewe, P., et al. (2001). An interna- Macmillan.
tional study of the psychometric properties of Toda, M. (1983). Future time perspective and human
the Hofstede Values Survey Module 1994. cognition: An evolutional view. International
Applied Psychology: An International Review, Journal of Psychology, 18(5), 351–365.
50, 269–281. Trempala, J., & Malmberg, L. (1998). The antici-
Stein, K. B., Sarbin, T. R., & Kulik, J. A. (1968). pated transition to adulthood: Effects of culture
Future time perspective: Its relation to the and individual experience on Polish and Finnish
socialization process and the delinquent role. adolescents’ future orientations. Journal of
Journal of Consulting and Clinical Psychology, Psychology, 132, 255–267.
32, 257–264. Triandis, H. C. (1984). Toward a psychological the-
Strathman, A., Gleicher, F., Boninger, D. S., & ory of economic growth. International Journal
Edwards, C. S. (1994). The consideration of of Psychology, 19, 79–95.
future consequences: Weighing immediate and Trommsdorff, G. (1983). Future orientation and
distant outcomes of behavior. Journal of socialization. International Journal of Psycho-
Personality and Social Psychology, 66, 742–752. logy, 18, 381–406.
Stubbart, C. (1982). Are environment scanning Trommsdorff, G., & Lamm, H. (1975). An analysis
units effective? Long Range Planning, 15(3), of future orientation and some of its social deter-
139–145. minants. In J. T. Fraser & H. Lawrence (Eds.),
Subramaniam, S. (2001). The dual narrative of “good The study of time II (pp. 343–361). New York:
governance”: Lessons for understanding politi- Springer.
cal and cultural change in Malaysia And Trommsdorff, G., Lamm, H., & Schmidt, R. W.
Singapore. Contemporary Southeast Asia: A (1979). A longitudinal study of adolescents’
Journal of International and Strategic Affairs, future orientation (time perspective). Journal of
23(1), 65–80. Youth and Adolescence, 8, 131–147.
Sundberg, N. D., Poole, M. E., & Tyler, L. E. (1983). Trompenaars, F., & Hampden-Turner, C. (1998).
Adolescents’ expectations of future events—A Riding the waves of culture: Understanding
cross-cultural study of Australians, Americans, cultural diversity in global business (2nd ed.).
and Indians. International Journal of Psychology, New York: McGraw-Hill.
18, 415–427. Tu, W. M. (1991). The continuity of being: Chinese
Swearer, D. K. (1997). The worldliness of Buddhism. visions of nature. In J. B. Callicott & R. T. Ames
Wilson Quarterly, 21(2), 81–93. (Eds.), Nature in Asian traditions of thought:
Takyi-Asiedu, A. (1993). Some socio-cultural factors Essays in environmental philosophy (pp. 67–78).
retarding entrepreneurial activity in Sub-Saharan New Delhi: Sri Satguru Publications.
Africa. Journal of Business Venturing, 8, 91–98. Tversky, A., & Kahneman, D. (1974). Judgment
Teahan, J. E. (1958). Future time perspective, opti- under uncertainty: Heuristics and biases.
mism, and academic achievement. Journal of Science, 185, 1124–1131.
Abnormal and Social Psychology, 57, 379–380. Van Vianen, A. E., & Fischer, A. H. (2002).
Teather, E. K., & Chow, C. S. (2000). The geogra- Illuminating the glass ceiling: The role of orga-
pher and the feng shui practitioner: So close and nizational culture preferences. Journal of
yet so far apart? Australian Geographer, 31(3), Occupational and Organizational Psychology,
309–332. 75, 315–337.
Tendam, H. W. (1987). Managerial flexibility: A strate- Volder, M. M. de, & Lens, W. (1982). Academic
gic asset. Leadership & Organization Develop- achievement and future time perspective as a
ment Journal, 8(2), 11–16. cognitive-motivational concept. Journal of
Thoms, M. A., & Greenberger, D. B. (1995). The Personality and Social Psychology, 42, 566–571.
relationship between leadership and time orien- Wallace, M. (1956). Future time perspective in schiz-
tation. Journal of Management Inquiry, 4, ophrenia. Journal of Abnormal and Social
272–292. Psychology, 52, 240–245.
13-House.qxd 3/19/2004 5:20 PM Page 342
Wallace, M., & Rabin, A. I. (1960). Temporal World Bank. (2000). Towards a measure of genuine
experience. Psychological Bulletin, 57 (3), savings. Retrieved September 2000 from http://
213–236. www.worldbank.org/data/wdi2000/pdfs/tab3_
Watson, J. B. (1913). Psychology as the behaviorist 15.pdf
views it. Psychological Review, 20, 158–177. Yi, C. H. (1998). Yi Chung-Hwan’s t’aengniji: The
Webber, M. J. (1972). Impact of uncertainty on Korean classic for choosing settlements
location. Cambridge: MIT Press. (University of Sydney East Asian Series No. 12;
Weber, M. (1958). The Protestant ethic and the spirit I. C. Yoon, Ed. & Trans.). Sydney, Australia:
of capitalism. New York: Scribner’s Press. Wild Peony. (Original work published c. 1750)
Williams, J. E., Satterwhite, R. C., & Saiz, J. L. Yoshino, M. Y. (1976). Japan’s multinational enter-
(1998). The importance of psychological traits: prises. Honolulu: Hawaii University Press.
A cross-cultural study. New York: Plenum. Yukl, G. (1998). Leadership in organizations (4th
Wing Sue, D., & Sue, D. (1990). Counseling ed.). Upper Saddle River, NJ: Prentice Hall.
American Indians. In Counseling the culturally Zaleski, Z., Chlewinski, Z., & Lens, W. (1994).
different: Theory and practice (pp. 175–188). Importance of and optimism-pessimism in pre-
New York: John Wiley. dicting solution to world problems: An inter-
Winnbust, J. A. M. (1975). Het Westerse Tijds- cultural study. In Z. Zaleski (Ed.), Psychology
syndroom: Conceptuele integratie en eerste of future orientation (pp. 207–228). Lublin,
aanzet tot construct-validatie van eeen reeks Poland: Towarzystwo Naukowe KUL.
molaire tijdsvariabelen in de psychologie [The Zimbardo, P. G., & Boyd, J. N. (1999). Putting time
Western Time Syndrome: Conceptual integra- in perspective: A valid, reliable individual-
tion and first attempt to construct validation differences metric. Journal of Personality and
of a series of molar time variables in psychol- Social Psychology, 77, 1271–1288.
ogy]. Nijmegen, The Netherlands: Stichting
Studentenpers Nijmegen.
14-HOUSE.qxd 3/19/2004 4:47 PM Page 343
14
CROSS-CULTURAL
DIFFERENCES IN GENDER
EGALITARIANISM
Implications for Societies,
Organizations, and Leaders
CYNTHIA G. EMRICH
FLORENCE L. DENMARK
AUTHORS’ NOTE: We would like to thank Erica Heitner for her assistance in the preparation of this chapter.
343
14-HOUSE.qxd 3/19/2004 4:47 PM Page 344
society: that society’s specific answer to a detail the development and nature of Hofstede’s
general problem with which any human society masculinity/femininity dimension to underscore
has to cope” (Hofstede, 1998, p. 10). Hofstede’s the differences between it and the measure of
masculinity/femininity dimension addressed Gender Egalitarianism developed in Project
the problem of the “duality of female versus GLOBE. Ultimately, we argue that masculinity/
male” (p. 11). Perhaps because societies’ femininity (Hofstede, 1980, 1998, 2001) con-
answers to this problem are both widely debated founds assertiveness, gender egalitarianism,
and wide-ranging in impact, Hofstede (1998) humane orientation, and achievement orienta-
characterized masculinity/femininity as the tion, thereby yielding findings that are difficult
“taboo dimension of national cultures.” He to interpret.
observed that the taboo placed on this dimen-
sion was greatest among masculine countries,
presumably because of “concerns that lead to HOFSTEDE’S MASCULINITY/FEMININITY
stress on political correctness in writing DIMENSION OF SOCIETAL CULTURE
about gender and to vigilance against sexual
harassment” (p. 209). In reality, the problem Conceptually, masculinity/femininity (Hofstede,
of the duality of female and male is twofold 1980, 1998, 2001) appears to encompass at
in nature. A society must decide both whether least two distinct aspects of societal culture.
to emphasize and reward behaviors that are The first aspect reflects differences among
stereotypically masculine versus feminine and societies in the extent to which each emphasizes
how to allocate social roles between the genders. and rewards “tough” or “masculine” values such
We consider these two problems in turn. as assertiveness, success, and competition
The first problem associated with the duality versus “tender” or “feminine” values such as
of female and male involves a society’s nurturance and solidarity. The emphasis that
general preoccupation with the “assertiveness” societies place on masculine versus feminine
(a stereotypically masculine trait) versus the values manifests itself in a variety of ways.
“nurturance” (a stereotypically feminine trait) For example, “masculine” cultures have been
of its members (Hofstede, 1980, p. 278). Some shown to score higher in achievement motiva-
societies emphasize and reward the pursuit of tion and in acceptance of a “machismo style” of
recognition and advancement, whereas others management (Triandis, 1994), and to value indi-
emphasize and reward cooperation and caring. vidual achievement over solidarity, confrontation
As such, societies can be ordered along a con- over cooperation, and independent thought over
tinuum, ranging from highly assertive to highly honoring moral obligations (Doney, Cannon, &
nurturing, to reflect the behaviors and traits that Mullen, 1998).
they value in their members. A second aspect of masculinity/femininity
The second problem associated with the (Hofstede, 1980, 1998, 2001) reflects differ-
duality of the sexes involves a society’s beliefs ences among societies in their beliefs about the
about the way in which social roles should behavior that is appropriate for males versus
be allocated between women and men. Some females. In more masculine cultures, males are
societies prescribe differentiated roles, whereas expected to be assertive and tough, whereas
other societies prescribe overlapping or egalitar- females are expected to be modest and tender.
ian roles. Some societies compel men to work In more feminine cultures, both females and
outside the home and women to work inside the males are expected to be modest and tender.
home (visible roles), whereas others encourage These expectations are presumably conveyed
men and women to participate equally in both through powerful socialization agents such as
venues. In addition, some societies encourage families, peers, schools, the media, and so forth
men to engage in “ego” roles and women to (Hofstede, 1980). Although cultures do vary in
engage in “social” roles in the home and in the their socialization of males and females,
community, whereas others encourage men and Hofstede (1980) and others (e.g., Coltrane,
women to engage equally in both emotional 1992; Williams & Best, 1982, 1990b) assert that
roles. In the next section, we discuss in some the most common socialization pattern is one in
14-HOUSE.qxd 3/19/2004 4:47 PM Page 345
which males are expected to be assertive, but masculinity/femininity (Hofstede, 1980, 1991).
females are expected to be nurturing. Several For the masculinity pole of this dimension,
studies have confirmed this pattern of socia- greatest importance was attached to the work
lization. In one 45-nation study (Barry, Bacon, goals of earnings, recognition, advancement,
& Child, 1957), for example, males were and challenge. In contrast, for the femininity
higher on self-reliance, achievement, and pole, greatest importance was attached to hav-
independence, whereas females were higher ing a good working relationship with one’s
on nurturance, responsibility, and obedience. manager, cooperation among peers, living in a
This pattern emerged in all 45 nations. More desirable area, and employment security.
recently, scholars have discovered “virtually Hofstede (1980) chose the masculinity/
perfect” correspondence between “sex differ- femininity label for this dimension, because
ences in socialization emphases and sex differ- it was the only one on which men and women
ences in behavior” (Segall, Dasen, Berry, & in the IBM sample scored differently. Men in
Poortinga, 1990, p. 250). These expectations this sample attached greatest importance to
shape not only the behavior of males and work goals such as earning and advance-
females, but their opportunities as well: ment (hence, “masculinity”), whereas women
attached greatest importance to goals such as
Organizations in masculine societies provide cooperation and a good working relationship
unequal opportunities for men and women to with manager (hence, “femininity”). Hofstede
advance in the managerial echelon and stress (1991) concluded, “The importance of earnings
work centrality over family life, independence and advancement corresponds to the masculine,
over dependence, decision over intuition, assertive, and competitive social role. The
assertiveness over consideration, results over importance of relations with the manager and
process, equity over equality and an adversarial with colleagues corresponds to the feminine,
over a mutual style of conflict resolution and caring, and social-environment role” (p. 82).
negotiation. (Erez, 1994, p. 573)
Replicating the
Development of the Masculinity/Femininity Dimension
Masculinity/Femininity Index
One criticism of Hofstede’s (1980) original
Masculinity/femininity was measured origi- study was that the cultural dimensions were
nally by surveying a sample of IBM employees derived exclusively from the responses of IBM
in 40 countries in the late 1960s and early 1970s employees. In response, several scholars have
(Hofstede, 1980). Additional data were col- conducted follow-up studies to gauge the extent
lected in the early 1970s that replicated the to which similar dimensions and findings would
original findings. The survey itself contains emerge if markedly different samples of respon-
questions related to 14 “work goals” such as dents were surveyed. For example, Hoppe (1998)
earnings, challenge, cooperation, and employ- noted that very few of Hofstede’s (1980) respon-
ment security. Respondents were asked to con- dents held leadership positions at IBM, and so he
sider these work goals and then, “Try to think conducted a comprehensive follow-up study of
of . . . which would be important to you in an 1,500 higher-level managers from 19 countries
ideal job; disregard the extent to which they are (Hoppe, 1990, 1993, 1998). These respondents
contained in your present job. How important were alumni of an international study center
is (each work goal)” (Hofstede, 2001, p. 256). called the Salzburg Seminar. On average, there
Respondents rated each of the 14 work goals were more women in Hoppe’s (19%) than in
on a 5-point scale, with “1” being “of utmost Hofstede’s (9%) sample, and Hoppe’s respon-
importance to me” and “5” being of “very little dents were older and more highly educated. The
or no importance.” Hofstede (1980) identified masculinity/ femininity scores obtained in this
two clusters in his analysis of respondents’ rat- follow-up study were correlated (r = .36, p > .05)
ings: individualism/collectivism (see Gelfand, with Hofstede’s study (1980). They were more
Bhawuk, Nishii, & Bechtold, Chapter 16) and highly correlated (r = .83, p < .01), however, with
14-HOUSE.qxd 3/19/2004 4:47 PM Page 346
a reformulated version of Hofstede’s index that both dimensions were presaged by the early
(1998, 2001). In Hofstede’s original formulation, anthropological work of Inkeles and Levinson
four work values comprised the masculinity/ (1954/1969, 1954/1997), and that the original
femininity dimension: advancement and earning research data confirmed that both dimensions
for masculinity and security and cooperation for reflected important aspects of societal culture.
femininity. The reformulated index retains Hofstede and Vunderink (1994) have argued, too,
advancement (masculinity) and cooperation that some scholars’ inappropriate interpretation
(femininity) but replaces the other two goals with of findings related to the masculinity/femininity
two new items. For masculinity, the new item dimension are partly to blame for the confusion
measures individuals’ beliefs in a just world by surrounding it: “In quite a few cases, differences
gauging the extent to which they agree with in psychological functioning across cultures
this statement: “When people have failed in life that conceptually clearly relate to the mas/fem
it is often their own fault.” For femininity, the distinction have indiscriminately been attributed
new item measures the extent to which indi- to individualism/collectivism: Masculinity at the
viduals agree that “Most people can be trusted.” individualist pole and femininity at the collec-
Although this reformulated index yielded tivist pole” (p. 330). Another criticism is that
stronger correlations between respondents at masculinity/femininity does not yield as “intui-
IBM and the Salzburg Seminar, the gender tive” a clustering as do other dimensions. For
differences that initially prompted Hofstede to example, individualism/collectivism typically
label the dimension masculinity/femininity failed yields clear contrasts between the Western and
to replicate in the latter sample. Specifically, the Asian cultures, whereas masculinity/femininity
men and women at the Salzburg Seminar created does not. The latter “bundles countries that are
nearly identical rank orderings of Hofstede’s 14 not engaged in common institutional dialogues, as
work goals (rho = 0.96, p < .01). is the case for Asian countries versus the West”
A more recent study of male airline pilots (Hofstede & Vunderink, 1994, p. 330).
from 19 countries also failed to replicate Masculinity/femininity has also been criti-
Hofstede’s (1980) masculinity/femininity index cized as overemphasizing conventional gender
(Merritt, 2000). One possible explanation for roles that may be less valid for self-description
this finding is that these male pilots ranked (Bem, 1975; Fagenson, 1990; Martin, 1987)
“employment security”—a feminine work value— given declining differences in the socialization
so highly due to the volatility in the aviation of males and females in many countries (Segall
industry. Further, the fact that male pilots are et al., 1990). The nonsignificant sex differences
employed in a relatively financially rewarding in Hoppe’s (1998) aforementioned study sug-
occupation may account for their relatively low gest that this criticism may be a valid one. To
rankings of “earnings”—a masculine work value this criticism, Hofstede and Vunderink (1994)
(Merritt, 2000). Taken together, the results from responded that masculinity/femininity is a
Hoppe’s (1998) and Merritt’s (2000) studies dimension of national culture, and
suggest limitations in the generalizability of the
masculinity/femininity dimension. . . . [is] not meant to describe individuals, but
dominant patterns of socialization (“mental pro-
Other Criticism of the gramming”) in nations; these dominant patterns
will affect different individuals to different
Masculinity/Femininity Dimension
degrees, and some components of a national cul-
Masculinity/femininity, like individualism/ ture pattern may be found in one individual, while
collectivism, affects the way that individuals other complementary components will be found in
function in their social environments. Hofstede other individuals within the same society. (p. 331)
and Vunderink (1994), noted, however, that
though the individualism/collectivism dimension
has been largely accepted by scholars, the same is Correlates of Masculinity/Femininity
not true for the masculinity/femininity dimension, In Hofstede’s (1980) original study, masculin-
which has received far more criticism. They argue ity/femininity was not significantly correlated
14-HOUSE.qxd 3/19/2004 4:47 PM Page 347
with the other three dimensions of societal with human-heartedness, in contrast, is somewhat
culture, with one exception. When Hofstede dis- puzzling, especially given that items that load
tinguished between wealthier and poorer coun- positively on this measure seem to reflect more
tries, masculinity/femininity and uncertainty feminine values such as “kindness,” “patience,”
avoidance were positively correlated (r = .43, and “courtesy.”
p < .05) in the wealthier countries, but nega- To summarize, Hofstede’s (1980) original
tively correlated (r = −.21, p > .05) in the poorer study represents a milestone in the research on
countries (Hofstede, 1980). Moreover, the com- national cultures. However, the findings to date
bination of these two dimensions was the best are mixed with regard to the reliability and
predictor of McClelland’s (1961) “need for validity of his masculinity/femininity dimen-
achievement.” Presumably, societal norms for sion. Conceptually, this dimension appears to
masculinity and uncertainty avoidance influence encompass at least two distinct subdimensions:
individuals’ values with regard to potential specifically, the extent to which societies foster
motivators. Masculinity/femininity reflects the and reward stereotypically masculine versus
opposition of ego needs versus affiliation needs, feminine behavior among members in general
respectively, whereas uncertainty avoidance and the extent to which they advocate differen-
reflects the opposition of success motivation tiated versus egalitarian roles for women and
(low uncertainty avoidance) versus fear motiva- men more specifically. To further complicate
tion (high uncertainty avoidance). matters, stereotypical masculine behaviors
Masculinity/femininity has been signifi- include success striving and assertiveness,
cantly correlated with numerous societal- whereas stereotypical feminine behaviors
level measures. For example, more masculine include nurturance and gender egalitarianism.
societies were found to be less permissive As a result, Hofstede’s masculinity/femininity
than more feminine cultures (Hofstede, 1991). measure confounds at least four dimensions
Governments in more masculine societies of societal culture and possibly others.
were more likely to sacrifice the environment Consequently, it is very difficult to interpret
in pursuit of economic growth (Hofstede, 1991). the masculinity/femininity measure itself and
Masculinity/femininity has been significantly its correlations with other measures. For the
correlated with various viewpoints as well purposes of Project GLOBE, we chose to
(Hofstede, 1980, 1991). For example, respon- empirically investigate cultural dimensions
dents from more masculine societies more of Performance Orientation (see Javidan,
strongly endorse viewpoints such as “individu- Chapter 12) in lieu of success striving, Humane
als make better decisions than groups,” “it is Orientation (Kabasakal & Bodur, Chapter 18)
better to work for a large corporation than a in lieu of nurturance, as well as Assertive-
small one,” and “employees will avoid work if ness (Den Hartog, Chapter 15) and Gender
they can.” Egalitarianism as two distinct cultural dimen-
Finally, Hofstede’s (1980, 1998, 2001) sions. The results of Project GLOBE empiri-
masculinity/femininity dimension has been cally confirm the conceptual division of these
significantly correlated with various value constructs.
dimensions, although not always in directions Project GLOBE’s Gender Egalitarianism
that make intuitive sense. As one might expect, measure constitutes the focus of this chapter. At
masculinity/femininity was positively corre- its core, this measure reflects societies’ beliefs
lated with “mastery,” or the self-assertion to about whether members’ biological sex should
control one’s social and physical environments determine the roles that they play in their
in a 23-nation study (Schwartz, 1994). But homes, business organizations, and communi-
masculinity/femininity was also positively ties. Societies with greater gender egalitarian-
correlated with “human-heartedness” or the ism rely less on biological sex to determine
social awareness and need to be courteous to the allocation of roles between the sexes.
others in another, 22-nation study (Chinese Henceforth, then, we refer to societies that seek
Culture Connection, 1987). The correlation with to minimize differences between the roles of
mastery makes intuitive sense. The correlation females and males in homes, organizations, and
14-HOUSE.qxd 3/19/2004 4:47 PM Page 348
communities as gender egalitarian. In principle (Hofstede, 1980, p. 262). These absolute and
and in practice, any distinction among societies statistical biological differences shape, in part,
with respect to this dimension is continuous societal norms regarding the behavior, activi-
rather than dichotomous in nature. Thus, gender ties, and occupations that are deemed suitable
egalitarianism is a relative term, with any given for each gender.
society being more or less egalitarian than other Of course, history offers numerous examples
societies in the same study. In the next section, of women who managed to circumvent the
we discuss the larger context of the Gender prevailing cultural norms of their societies to
Egalitarianism construct and its relevance to pursue careers that were prohibited to them. For
individuals, organizations, and societies. example, the Brontë sisters (“Acton, Currer, and
Ellis Bell”), Mary Ann Evans (“George Eliot”),
and Amantine Aurore Dudevant (“George Sand”)
THE LARGER CONTEXT assumed male noms de plume to pursue cele-
OF GENDER EGALITARIANISM brated literary careers (Jordan & Patten, 1995;
Walton, 2000). In medicine, Dr. Mary Putnam
Scholars from multiple perspectives recognize Jacobi initially wrote anonymously for the presti-
the universality of the division of roles between gious New York Medical Record. Over time,
the sexes. Some anthropologists regard the however, she made fewer concessions to the lim-
duality of female and male as one of the most its placed on her by the male-dominated medical
fundamental dualities in the human existence— profession and pioneered the use of survey data in
a duality that dictates reproduction and, hence, medical research (Wells, 2001). Though notable,
the perpetuation of societies (Braidotti, 1994; women such as these constitute exceptions
Gilmore, 1996; Linke, 1992; Roschrhomberg, (McElvaine, 2000). Moreover, societal norms
1994). Women not only bear children, but also regarding appropriate roles for females and males
feed and nurture them. Hence, the responsibility affect not only their career choices and success
for nurturing people generally, and children, (Heilman, 1995; Stroh, Brett, & Reilly, 1992;
more specifically, falls on women. Men, on the Walker & Fennell, 1986), but also their mental
other hand, are placed in charge of economic and physical health (Girard, 1993; Peterson &
matters. Smith, 1997; Piccinelli & Simon, 1997), mortality
From a psychological perspective, sex ranks (Kattler & Williamson, 1988; Pampel & Park,
alongside age as a “universal dimension for 1986; Pattnayak & Shai, 1995), parent–child rela-
differentiating people” (Fiske & Taylor, 1991, tionships (Coltrane, 1988, 1992; Katz & Konnor,
p. 121). Psychological differentiation of the 1981; Mead, 1949), and a host of other outcomes
sexes occurs by assigning distinct and nonover- that dictate the nature and quality of human life.
lapping personality traits to females and males. Our focus in this chapter is on the cultural norms
The duality of male assertiveness and female that shape the roles and, hence, lives of females
nurturance is a common stereotypic portrait. and males in societies around the globe.
Studies have found that women are considered Gender egalitarianism can best be explored
more communal and expressive than men, in the context of its attitudinal domain and its
whereas men are deemed more competent, agen- behavioral manifestations. Figure 14.1 depicts
tic, and instrumental than women (Denmark, the two components of Gender Egalitarianism
Rabinowitz, & Sechzer, 2000). as conceptualized and measured in Project
From a biological perspective, only child- GLOBE. Its attitudinal domain relates to the
bearing and child begetting constitute absolute fundamental values, beliefs, and attitudes held
sex differences. All other differences are by members of a society with regard to gender
statistical with, for example, the average man stereotypes and gender-role ideology. The
being taller, stronger, and heavier than the aver- behavioral manifestations are actions and
age woman, and the average woman having behaviors observed in a society in relation to
greater finger dexterity, longer life-expectancy, gender egalitarianism—for example, gender
and faster metabolism than the average man discrimination and gender equality.
14-HOUSE.qxd 3/19/2004 4:47 PM Page 349
p. 255). In this way, gender-role ideologies and partnership in its firm. In this way, gender
gender egalitarianism are linked in theory and in stereotypes, and the discrimination they fos-
reality. tered, served to reinforce societal norms regard-
ing appropriate roles for women and men in the
American workplace.1
Behavioral Manifestations
The final concept to consider is gender
A third concept associated with gender egali- equality (Hendrix, 1994). Studies of gender
tarianism is gender discrimination. Both gender equality focus on the degree to which women
stereotypes and gender-role ideologies reflect and men are represented equally in the labor
individuals’ internalized beliefs regarding the force and in positions of authority (e.g.,
behavior, qualities, roles, and abilities of females Blackburn, Jarman, & Siltanen, 1993; Charles,
and males and, as such, comprise the attitudinal 1992; Clark & Anker, 1993; Jacobs & Lim,
domain of gender egalitarianism. By contrast, 1992; Kukreja, 1992; Moore & Shackman,
gender discrimination constitutes an external act 1996; Nuss & Majka, 1983; Wright, Baxter, &
(i.e., a behavioral manifestation) that prevents Birkelund, 1995). Studies of gender equality
members of one sex from gaining the recogni- also examine women’s and men’s relative
tion, occupations, positions, and status accorded contributions to child rearing and housework
to equally qualified members of the opposite (e.g., Berk & Berk, 1979; Chafetz & Hagan,
sex. These concepts are associated, because it 1996; Coverman & Sheley, 1986; Hochschild,
seems likely that gender stereotypes and gender- 1989; Reskin & Padavic, 1994; South & Spitze,
role ideologies play key roles in gender discrim- 1994; Szalai, 1972). Thus, the concept of gen-
ination. The Supreme Court of the United States der equality, like gender discrimination, consti-
gave legal standing to this argument when it tutes a behavioral manifestation of societies’
ruled that gender stereotyping was a precursor to beliefs about the appropriate allocation of roles
gender discrimination and, as such, constituted a between the sexes, or gender egalitarianism.
violation of women’s civil rights: The more that a society seeks to minimize
differences in the roles allocated to the sexes,
In the specific context of stereotyping, an the more equality we would expect to find in
employer who acts on the basis of a belief that a women’s and men’s participation in the labor
woman cannot be aggressive, or that she must not force, positions of authority, child rearing, and
be, has acted on the basis of gender. . . . We are housework. The more that a society seeks to
beyond the day when an employer could evaluate maximize differences in the roles allocated to
employees by assuming or insisting that they the sexes, the more inequality we would expect
matched the stereotype associated with their to find in women’s and men’s participation in
group. . . . An employer who objects to aggres- these same areas.
siveness in women but whose positions require To summarize, we began this section with a
this trait places women in an intolerable Catch 22: discussion of different perspectives—anthropo-
out of a job if they behave aggressively and out of logical, psychological, and biological—on the
a job if they don’t. (Price Waterhouse v. Hopkins, duality between the sexes. We moved next to a
1989, pp. 1790–1791) consideration of the two key components of
gender egalitarianism: its attitudinal domain and
Gender egalitarianism and the concepts of behavioral manifestations. By taking into con-
gender stereotypes, gender-role ideology, and sideration the larger context of gender egalitari-
gender discrimination are interwoven because anism, it is possible to anticipate its causes and
beliefs about what females and males can and consequences. As a cultural norm, gender egal-
should do reinforce a society’s preferences itarianism reflects individuals’ stereotypes and
regarding the allocation of roles between them. In ideologies regarding females and males. These
the case cited above, gender stereotypes that beliefs play a key role in gender discrimination,
depicted, and even demanded, passivity in women an act that promotes gender inequality in homes,
led Price Waterhouse to deny Ann Hopkins, an workplaces, and communities. In this way, the
aggressive and achievement-oriented woman, relationships among gender stereotypes and
14-HOUSE.qxd 3/19/2004 4:47 PM Page 351
gender-role ideology on the one hand and Confounded explanations such as these
gender discrimination and gender inequality on may never be resolved to our satisfaction.
the other comprise a recurrent feedback loop. Nonetheless, we devote this section to a dis-
Specifically, gender stereotypes and ideologies cussion of the potential causes of cross-cultural
about what is possible or appropriate for differences in gender egalitarianism. It has been
females and males limit their social roles, suggested that “our reasoning about sex roles
thereby affecting their participation in the labor and inequality has been woefully oversimplis-
force and their contributions to their families. tic,” and that “we need to at least think about
Gender inequality in the home, workplace, and multiple causes and feedback relations as we
society shapes, in turn, individuals’ gender try to develop our theories further” (Hendrix &
stereotypes and ideologies, bringing us full Hossain, 1988, p. 452). With this critique and
circle. The concept of gender egalitarianism recommendation in mind, we cover a number of
grounds this feedback loop in powerful and potential causes of gender egalitarianism. The
overarching cultural norms about the division of majority of these have not been examined in the
roles between the sexes. These norms, though context of gender egalitarianism per se, but in
not explicit in Figure 14.1, are the focus of this the context of associated constructs such as gen-
chapter. Next, we consider past research on der stratification, gender ideology, and gender
the causes of cross-cultural differences in gen- equality. In this section, we attempt to make the
der egalitarianism and explore its relevance to case for a causal relationship between these
individuals, organizations, and societies. potential cultural drivers and gender egalitarian-
ism specifically (see Table 14.1).
limited than males in the number of offspring help to create a new set of beliefs about the
that they can produce; therefore, their parental roles that are possible and appropriate for
investment is necessarily greater: women and men. In a sample of 90 nonindus-
trial societies (Coltrane, 1988), the more active
In mammals the female is uniquely adapted to care men’s involvement in child rearing, the greater
for the young, and the male is not necessarily joined women’s decision making and authority in
to the mother-offspring unit. In most mammals the their communities. Thus, the greater men’s
parents do not remain together after copulation, and role in child rearing in a society and the closer
the male contributes no parental investment except their proximity to their young children, the
the sperm. (Katz & Konnor, 1981, p. 159) more affectionate their relationships with their
children and the greater the public prestige of
The balance between the parental invest- women in that society. This relationship held
ments of women and men varies across cultures. even after controlling for several factors that
In the past 20 years, scholars have focused on were believed to influence women’s status such
men’s participation in child rearing as a poten- as a male social structure, external warfare, and
tial explanation of cross-cultural differences societal complexity.
in the division of roles between the sexes A more recent study of 92 nonindustrial soci-
(Coltrane, 1988, 1992; Katz & Konnor, 1981). eties (Coltrane, 1992) demonstrated a strong
For example, numerous explanations have and consistent relationship between men’s
been proposed to account for women’s limited parental investment and key measures of
public authority in most countries. Common to women’s prestige in the home and in society,
virtually all of these explanations “is the notion thereby replicating the results reported above
that women’s primary responsibility for child (Coltrane, 1988). Results showed that men’s
rearing, while culturally conditioned, constrains parental investment was negatively correlated
their ability to exercise public power” (Coltrane, with women’s deference to men, husbands’
1988, p. 1061). Earlier, we argued that gender dominance over their wives, and an ideology of
stereotypes, gender-role ideologies, gender dis- women’s inferiority. The closer men’s proxim-
crimination, and gender inequality combine to ity to their children, the more active their role in
form a recurrent feedback loop. Coltrane (1988, caring for their children: the more affectionate
1992) and others (e.g., Chodorow, 1978; Mead, their relationships with their children, the more
1949; Parsons, 1964; Slater & Slater, 1965) have egalitarian the treatment of and beliefs about
argued similarly that the “sexual division of labor women in that society. These results pinpoint
in child rearing” (Coltrane, 1988, p. 1061) rein- potential mediators of the earlier-reported
forces gender-role stereotypes and ideologies (Coltrane, 1988) relationship between men’s
by exposing children to women and men in their parental investment and women’s status.
traditional roles of caretakers and providers, Specifically, the quality of men’s relationships
respectively. Through exposure to women and with their children may affect the way they treat
men in traditional gender roles, children come to and view women (Coltrane, 1992), which, in
internalize these in the form of stereotypes and turn, may affect women’s access to decision
ideologies. In this way, gender roles are passed making and to positions of authority in society
down through successive generations. (Coltrane, 1988).
This argument has led some to speculate
that, by increasing men’s parental investment
Climate or Geographical Latitude
(Trivers, 1972) or involvement in child rearing,
it might be possible to increase the public pres- The ambient temperature of a society,
tige and authority of women. Greater paternal as measured by its geographical latitude or
investment in child rearing may not only give distance from the equator, may be another key
women more time to pursue nontraditional determinant of gender egalitarianism. As
social roles, it may also expose children to more Hofstede (1980) noted, ambient temperature
women and men in nontraditional social roles. “clearly comes first. . . . In more moderate
In this way, increased paternal investment may climates, survival presupposes the mastery of
14-HOUSE.qxd 3/19/2004 4:47 PM Page 353
complex skills by both men and women, which Several scholars have examined the relation-
makes extreme inequality between the sexes ship between extreme ambient temperature and
unlikely” (p. 203). In other words, both women Hofstede’s masculinity/femininity dimension of
and men must attain similar complex skills in societal culture (Peterson & Smith, 1997; Van de
less-hospitable climates to ensure survival. In Vliert, 1998; Van de Vliert et al., 1999). For
fact, the relationship between Hofstede’s mas- example, Van de Vliert and his colleagues
culinity/femininity dimension and geographical (1999) found a positive relationship between
latitude was significant in his original survey ambient temperature and cultural masculinity in
(1980): Nations closer to the equator had more 53 countries: The greater the mean ambient tem-
masculine cultures, whereas nations closer to perature of a society’s capital, the more mascu-
the poles had more feminine cultures. This rela- line its culture. This relationship was significant
tionship was moderated, however, by national even after controlling for such factors as popula-
wealth. Specifically, the relationship between tion size, population density, economy, and
geographical latitude and masculinity/femininity democracy. This and other studies (Hofstede,
was stronger among poor than among wealthy 1980; Peterson & Smith, 1997; Van de Vliert,
countries (Hofstede, 1980, p. 203). Presumably, 1998) suggest that extreme ambient temperature
the greater struggle for existence (Darwin, 1859/ is a key antecedent of “the culturally pro-
1958) in poor rather than in rich countries grammed tendency of a society’s male inhabi-
increased the salience of biological determinants tants to manifest masculinity” (Van de Vliert,
of gender roles. 1998, p. 128). On the basis of Triver’s (1972)
More recent investigations have provided parental investment theory, it seems likely that
additional insight into the role of ambient tem- ambient temperature also plays a role in the cul-
perature in gender roles (Peterson & Smith, 1997; turally programmed tendency to assign females
Van de Vliert, 1998; Van de Vliert, Schwartz, and males distinct, gender-differentiated roles in
Huismans, Hofstede, & Daan, 1999; Van de homes, organizations, and communities.
Vliert & Van Yperen, 1996). In each of these The fundamental conclusion of this section is
studies, Triver’s (1972) parental investment that there is a higher level of cultural masculin-
theory plays a key explanatory role, because a ity in warmer than in colder climates. Parental
society’s ambient temperature is theorized to investment theory (Trivers, 1972) seems to pro-
affect the degree to which both parents must vide a powerful explanatory hypothesis for this
invest in offspring if the offspring are to survive. phenomenon. An alternative hypothesis—based
In tropical, hunter–gatherer societies, women on resource dependency theory—would claim
could gather nearby vegetation year-round and that men are granted more status in colder cli-
thereby “adequately feed themselves and their mates because their dexterity and physical attrib-
children without male provisioning” (Miller, utes are more important to the survival of the
1994, p. 228). In societies that lie closer to the family. To date, however, the data most clearly
poles, however, females depend on males to pro- favor Trivers’ parental investment theory.
vide meat during the winter, a time when vege-
tation is dormant. The greater need for male
Religion
provisioning in colder climates ensures that males
will have both greater investment in as well as Scholars continue to debate the potential
closer proximity to their offspring. Simply put, it causal role of religion in gender egalitarianism
is more demanding to meet basic needs of food, and related constructs. Monotheism—or the
safety, and security in cold climates than in warm belief in one god—has been linked to the low
climates (Van de Vliert & Van Yperen, 1996). To status of females in societies (Stover & Hope,
raise children in cold climates, then, requires 1984). The low status of females and monothe-
cooperation between men and women. This type istic beliefs are theoretically linked by the fact
of “cross-gender exchange, complementary divi- that virtually all monotheistic cultures worship a
sion of labor, and maintenance of mixed-gender male, rather than a female, god (Paxton, 1997).
interest groups could result in more egalitarian In contrast, polytheistic cultures worship a
interactive behavior” (Coltrane, 1992, p. 94). variety of gods, which often includes both male
14-HOUSE.qxd 3/19/2004 4:47 PM Page 354
and female figures. Some have argued that the women’s inferiority in matters of politics and
superior status accorded to males in monotheistic administration and specific religious ideologies,
religions causes female’s low status. However, Paxton argued that, although there was variation
some sociologists—for example, Durkheim— among Protestant religions, they were generally
have argued the reverse: That “monotheism is less patriarchal than the Catholic or Islamic
a reflection of the causally prior low status of religions. This difference may be due to the
women” (Gray, 1987, p. 1121; see also Verweij, fact that religious leaders in both Catholicism
Ester, & Nauta, 1997). and Islam are uniformly male, whereas other
To further complicate matters, obtained religions permit women to hold leadership
correlations between monotheism and gender positions. Geographical region was also a
equality are mixed. In one cross-cultural study, significant predictor of women’s representation
women’s status and monotheism were negatively in national legislatures (Paxton, 1997). Their
correlated such that women were accorded representation was lowest in the Middle East,
lower status in monotheistic societies (Stover & Southern Africa, South America, and Asia. The
Hope, 1984). However, Gray (1987) reanalyzed introduction of religion into the equation sig-
these findings to examine the possibility that nificantly reduced the predictive power of
only a subset of geocultural regions was respon- geographical region. As predicted, Catholicism
sible for this correlation: was negatively associated with women’s repre-
sentation in national legislatures; however,
Most societies in the Circum-Mediterranean region Islam was not. One of the regional variables that
exhibit both active high gods [i.e., monotheism] and remained significant even after the introduction
low female status. Societies in the Insular Pacific of the religious ideology variables was that of
region usually lack high gods and award high sta- the Middle East, which lead Paxton to speculate
tus to women. The societies in these two regions that “the interpretation of Muslim doctrine may
are so uniform that they alone might account for be more severe in the Middle East” (p. 461).
the statistically significant relation between gender This is true in many countries in the Middle East
bias and monotheism. (p. 1125) where religious groups hold political power.
Perhaps Islam is more strictly interpreted in the
Gray discovered, in fact, that monotheism Middle East than in other regions because it is
predicted low status for women in the African, the center of Islamic orthodoxy.
Circum-Mediterranean, East Eurasian, and The relationship between religion and gender
South American regions of the world but pre- egalitarianism becomes further complicated by
dicted high status for women in the Insular the argument that a society’s norms concerning
Pacific and North American regions of the masculinity/femininity affect its religious men-
world. Thus, monotheism’s effect on women’s tality rather than the reverse (Verweij et al.,
status was not a uniform one worldwide. 1997). Building on prior research and theory
Up to this point, our focus has been on (Hofstede, 1991), Verweij and his colleagues
the role of religion generally in concepts related argued that a society’s gender-role distinc-
to gender egalitarianism—specifically, women’s tions affect the cognitions and attitudes of its
status and prestige, as well as beliefs about members:
women’s roles in society. However, the pres-
ence or absence of “high gods” (Gray, 1987) or In masculine countries, God—represented in a
overall religiosity of a culture does not tell the male image as Father, King, Shepherd—is felt to
full story, because different religions hold be more important in people’s everyday lives than
different beliefs about appropriate gender roles in feminine countries. In a feminine culture, rela-
in society. In one study, for example, Paxton tions are more important than distinctions and
(1997) examined the relationship between positions, and therefore in the domain of religion
varying dominant religions (i.e., Protestant, a hierarchical God, or a more outspoken male or
Islam, Roman Catholic, and Other) and the rep- female representation of God, is less acceptable
resentation of women in national legislatures. than a more nondescript, individualized image of
Drawing linkages between an ideology of God. (p. 312)
14-HOUSE.qxd 3/19/2004 4:47 PM Page 355
Thus, according to Verweij and his for the impact of economic development on
colleagues, the culturally programmed tendency women’s roles and status in societies (e.g.,
to prescribe differentiated gender roles precedes Smelser, 1970 vs. Boserup, 1970).
a society’s religious beliefs, thoughts, and One 162-nation study, for example, exam-
practices. Data from the 1990 European Values ined the relationship between national economic
Survey of 16 Western, industrialized nations development and the integration of women into
is consistent with this prediction. Nations’ the labor force (Nuss & Majka, 1983). Gross
masculinity/femininity scores constituted “one National Product (GNP), or the value of all
of the most important single society characteris- products and services produced by a nation,
tics explaining cross-national differences in sec- served as the measure of national economic
ularization” (Verweij et al., 1997, p. 322), such development, and crude labor force participa-
that the more feminine a nation’s culture, the less tion rates of women served as a measure of
religiously oriented and orthodox its members. their integration into the labor force. There
In summary, scholars continue to debate was no overall linear or curvilinear relationship
both the nature and causality of the relationship between GNP and women’s labor force partici-
between religion and constructs related to pation. Instead, higher per capita GNP was
gender egalitarianism. Nonetheless, there is suf- associated with increases in women’s participa-
ficient empirical evidence and sound theory to tion only in traditionally female occupations
regard societies’ religious beliefs, thought, and in the clerical and communication sectors of
behavior as potentially key antecedents of their nations’ economies. From these findings,
tendencies to use members’ biological sex to Nuss and Majka (1983) concluded that, “a divi-
determine their social roles. sion of labor between women and men is a
global phenomenon. . . . Economic develop-
ment has no predictive value for the integra-
Economic Development
tion of women into the essential sectors of the
Nations vary in the nature and health of their economy. . . . The same is true for important
economies. Several past studies have examined occupations” (p. 44).
the link between various economic indicators More recently, a study of 32 developing
and the status and roles of women in societies. nations found that higher levels of direct
Although it is difficult to argue conclusively investment in a nation led to a “feminization
that economic factors play a causal role in gen- of the tertiary” or service sector of the economy
der egalitarianism and related concepts, it is (Kukreja, 1992). Thus, economic development
perhaps easier to argue than the reverse. did not create new roles for women in these
Moreover, past scholars have typically argued countries. Instead, it “pushed” them into per-
that economic factors play a causal role, and forming more of the same types of services they
we do so here as well (e.g., Menard, 1990; had performed in the past. Although women
Moore & Shackman, 1996; Nuss & Majka, now received financial remuneration for these
1983; Shen & Williamson, 1999; Sigelman & services, their roles in the labor force were lim-
Tsai, 1985). ited, nevertheless, to those traditionally deemed
The underlying, causal assumption of most suitable for them.
past research in this area is that an “acceleration Given the mixed findings from the modern-
of economic development is accompanied ization perspective, researchers examined it and
by inevitable improvement in other spheres of an alternative, economic competition perspec-
society” (Nuss & Majka, 1983, p. 30). Accor- tive in one study of 100 nations from all regions
ding to this modernization perspective (Moore of the world (Moore & Shackman, 1996). This
& Shackman, 1996), economic development alternative perspective argues that gender
creates new opportunities for women and other inequality persists or even increases with growth
marginalized groups because it both increases in industrialization because it stimulates compe-
available jobs and changes the social and tition between women who are relatively power-
cultural climate. Although theoretically and less and men who wish to retain their power.
intuitively compelling, there is mixed evidence The findings from this study supported the
14-HOUSE.qxd 3/19/2004 4:47 PM Page 356
economic competition perspective rather than women have relatively few valued resources
the modernization perspective: GNP had a nega- to contribute to their marriages when they are
tive impact on women’s participation in admin- prevented from both inheriting their families’
istrative occupations, but no impact on women’s wealth and living close enough to draw on
representation in parliament. Therefore, economic their support, status, and power. By contrast, the
development was associated with either a nega- presence of matrilineal descent and matrilocality
tive change or no change in women’s represen- are theorized to promote more modern gender ide-
tation in traditionally male positions in commerce ologies and gender equality in societies, because
and in politics, respectively. women have more economic and familial
In summary, the impact of a nation’s wealth resources and, hence, power at their disposal.
or economic development on gender equality in There are relatively few societies that trace
the labor force is more equivocal than originally individuals’ lineages and the like through the
envisioned. One reason may be the types of maternal line, however, just as there are rela-
measures used. GNP is a confounded measure tively few truly gender egalitarian societies
of economic development because of the influ- (Coltrane, 1992). Furthermore, as with other
ence of population size. Regardless, evidence potential antecedents of gender egalitarianism,
gathered to date suggests that economic devel- it is difficult to claim unequivocally a causal
opment does not create new roles for women relationship between a society’s structure or
but, instead, simply provides more venues— resource control and its allocation of roles
“pink-collar ghettos” (Nuss & Majka, 1983, between the sexes. It seems reasonable to con-
p. 42)—for them to engage in their traditional, ceptualize societal structure as playing a causal
service-oriented roles. role in gender egalitarianism, nonetheless,
because the manner in which a society traces
members’ ancestries and transfers wealth from
Social Structure and Resource Control
one generation to the next is both enduring and
The conflict over scarce resources constitutes highly institutionalized. Thus, this societal
one of the most basic struggles in human exis- aspect likely takes temporal and, hence, causal
tence. In many societies, men form strong precedence over other more changeable aspects
fraternal interest groups that link “coresident, of societies. With a causal relationship in mind,
related men” (Coltrane, 1992) to maintain then, we review several key studies of the rela-
control over scarce resources and to enhance tionship between societal structure and various
their power and prestige: concepts related to gender egalitarianism.
A study of 100 nonindustrialized societies
In a system of patrilineage, not only is descent reck- was conducted to identify the “circumstances
oned through the male line, but there is a tendency or conditions under which wives may have
for inheritance to flow disproportionately to men. greater or lesser decision-making authority in
. . . Patrilocality requires women to move near their marriage” (Warner et al., 1986, p. 122). The
husband’s kin groups at marriage and is commonly researchers examined women’s access to mater-
seen as enabling men to appropriate women’s labor ial resources, as operationalized by lineage and
and products while it enhances the authority of locality, and to social resources in two family
senior men. (Coltrane, 1992, p. 93) structures: nuclear and extended. Following
on Whytes’s (1978) earlier work, it was hypoth-
According to resource theory (McDonald, esized that the presence of multiple adults of
1980; Safilios-Rothschild, 1970), strong frater- both sexes in extended families promotes
nal interest groups—created and reinforced by greater role differentiation between the sexes. If
patrilineal descent and patrilocality—promote a wife is not able to care for the children in an
traditional gender ideologies and gender extended family, then this responsibility need
inequality in societies, because they limit not fall to the husband. A grandmother, aunt, or
women’s ability to “contribute valued resources sister can assume responsibility for the children.
to the marriage” (Warner, Lee, & Lee, 1986, In contrast, if the wife is not able to care for the
p. 121). The rationale for this assertion is that children in a nuclear family, the husband must
14-HOUSE.qxd 3/19/2004 4:47 PM Page 357
care for them himself. In Warner and colleague’s extent and nature of women’s representation in
study, five coders read ethnographic material on formalized political roles. Contrary to the pre-
more than 100 societies and then rated each on dictions of resource theory, women’s kinship
the degree to which wives exercised decision- power, property control, and labor value were
making authority over their own and family not significant predictors of women’s access to
members’ behavior, including that of their formalized political roles. Consistent with
husbands. Wives had a high degree of power resource theory, women’s solidarity was a sig-
in 46.3% of societies with a nuclear family nificant predictor. Thus, predictions associated
structure, but in only 29.2% of societies with an with resource theory were only partially sup-
extended family structure. In addition, wives ported in this study of women’s political partic-
had a high degree of power in 70% of societies ipation. Despite mixed findings, this study
with matrilineal descent and matrilocality, offers an important contribution to the literature
but in only 32.5% of societies with patrilineal in its demonstration of the pivotal role that
descent and patrilocality. Together, these two psychological resources such as women’s soli-
findings suggest that access to material resources darity can play in promoting gender equality in
alone could not account for cross-cultural societies by “transform[ing] a collectivity and
differences in gender egalitarianism. Social prepar[ing] it for various types of socio-political
resources in the form of an interdependency and action” (Welch, 1983, p. 69).
exchange of traditional husband and wife roles To summarize, the social structures put in
also played a key explanatory role. place to control scarce resources appear to play
A second study of 93 nonindustrial societies a key role in the division of roles between the
(Coltrane, 1992) confirmed the dual importance sexes. Women are shown greater deference by
of resource control and male–female inter- men, exert greater decision-making authority,
dependence in gender egalitarianism: Women’s and are more likely to gain access to key
control of property was negatively correlated religious and political roles when they have
with men’s displays of manliness, women’s greater control over critical economic, social,
deference to men, and husbands’ domination of and psychological resources.
their wives. The more control women had over
property, inheritances, and the proceeds from
Mode of Production
productive labor, the less “belligerence and
bravado” (pp. 102–103) men displayed. However, In a seminal treatise, Sacks (1979) theorized
it was only when women both controlled prop- that women’s control over the means of pro-
erty and shared child rearing responsibilities duction determines their status in societies (see
with men that they were “likely to avoid the also Hendrix & Hossain, 1988). In this treatise,
harassment and humiliation that comes from women’s control of production is considered
being simultaneously feared and denigrated by the key mediator in the relationship between
men” (p. 105). mode of production (defined below) and
Two additional studies examined the issue of women’s social status. Hendrix and Hossain
social structure and resource control, one in the (1988) empirically tested Sack’s (1979) causal
religious domain (Welch, 1982) and the other in model in a worldwide sample of 93 tribal and
the political domain (Welch, 1983). The tenets historical societies. In this study, they opera-
of resource theory found mixed support in the tionalized mode of production as a combination
study of women’s religious roles. As predicted, of two variables: Type of descent system (i.e.,
greater “kin power” (e.g., matrilocality) was matrilineal, patrilineal, or both) and level of
positively associated with women’s selection social stratification (i.e., recognizing neither
into the shaman role. However, counter to class nor wealth distinctions vs. recognizing
prediction, greater “property control” (e.g., one but not the other). By combining these two
matrilineage) was negatively associated with variables, Hendrix and Hossain created a tri-
women’s selection into the shaman role (Welch, chotomy of mode of production: The commu-
1982). A second study (Welch, 1983) examined nal mode, the kin corporate mode, and the class
the utility of resource theory for explaining the mode. Their dependent variable reflected three
14-HOUSE.qxd 3/19/2004 4:47 PM Page 358
Table 14.2 Higher Gender Egalitarianism Societies Versus Lower Gender Egalitarianism Societies
Societies That Score Higher on Gender Societies That Score Lower on Gender
Egalitarianism Tend to: Egalitarianism Tend to:
• Have more women in positions of authority • Have fewer women in positions of authority
• Accord women a higher status in society • Accord women a lower status in society
• Afford women a greater role in community • Afford women no or a smaller role in
decision making community decision making
• Have a higher percentage of women • Have a lower percentage of women
participating in the labor force participating in the labor force
• Have less occupational sex segregation • Have more occupational sex segregation
• Have higher female literacy rates • Have lower female literacy rates
• Have similar levels of education of females • Have a lower level of education of females
and males relative to males
of cultural drivers of gender egalitarianism tentative and general, the portraits presented
because, as noted earlier, the solidarity created in Table 14.2 reflect the pervasive effects of
through such movements provides a key psycho- cultural norms regarding gender egalitarianism
logical resource in women’s attempts to assume on males, females, and nations (see also House
leadership positions in societies (Welch, 1983). et al., 1999). We devote the next two sections of
In short, both the nature and openness of a this chapter to describing the measurement of
nation’s political system may play important the Gender Egalitarianism dimension in Project
and potentially causal roles in women’s success GLOBE and to exploring the relationship between
in attaining leadership roles traditionally it and key political, economic, religious, and
reserved for men. Altogether, we have identified social indicators.
seven potential antecedents of cross-cultural
differences in gender egalitarianism: Parental
investment, climate, economic development, GLOBE MEASURES
social structure and resource control, mode of OF GENDER EGALITARIANISM
production, religion, and political system.
In GLOBE, Gender Egalitarianism was mea-
sured at both the societal and organizational
Summary of Cultural
levels. At the societal level, participating man-
Drivers of Gender Egalitarianism
agers completed two scales: One that assessed
From the preceding discussion, one gains a their perceptions of the current (As Is, or prac-
sense of the complex nature of gender egalitari- tices) degree of Gender Egalitarianism in their
anism and related constructs. Gender egalitari- societies, and another that assessed their percep-
anism is complex in the sheer number and range tions of the ideal (Should Be, or values) degree
of antecedents that drive cross-cultural differ- of Gender Egalitarianism in their societies.
ences in the division of roles between the Managers also completed two scales—practices
sexes. It is complex, too, in that studies within a and values—that assessed their perceptions
particular area sometimes yield contradictory of the degree of Gender Egalitarianism in
findings. Even so, portraits begin to emerge their organizations. Thus, Gender Egalitaria-
of societies that seek to minimize gender-role nism was measured using four scales, each with
differences as opposed to those that seek to maxi- three to five quantitative items that survived the
mize these differences. Although necessarily statistical procedures described by Hanges and
14-HOUSE.qxd 3/19/2004 4:47 PM Page 360
Table 14.3 Sample Items From the Gender Egalitarianism Scale-Societal Practices (As Is)
1-17. In this society, boys are encouraged more than girls to attain a higher education:
Strongly Neither agree Strongly
agree nor disagree disagree
1 2 3 4 5 6 7
1-36. In this society, who is more likely to serve in a position of high office?
Men Women
1 2 3 4 5 6 7
Table 14.4 Sample Items From the Gender Egalitarianism Scale-Societal Values (Should Be)
3-17. I believe that boys should be encouraged to attain a higher education more than girls:
Strongly Neither agree Strongly
agree nor disagree disagree
1 2 3 4 5 6 7
Dickson in Chapter 8. Each item was measured scale. This item is worded exactly as adminis-
on a 7-point scale. All four scales were vali- tered, with emphasis added to highlight the
dated in all societies involved in the GLOBE focus on practices.
project to ensure the comparability of the Table 14.6 contains a sample item from the
Gender Egalitarianism measures cross-culturally. organizational Gender Egalitarianism values
When interpreting results involving As Is and scale. This item is worded exactly as adminis-
Should Be scales, we henceforth use the terms tered, with emphasis added to highlight the
practices and values, respectively, to emphasize focus on values.
that participating managers’ As Is responses Much cross-cultural research has been con-
reflect their perceptions of current practices, ducted using Hofstede’s masculinity/femininity
whereas their Should Be responses reflect their dimensions. As noted earlier in this chapter, how-
underlying values with regard to an ideal soci- ever, Hofstede’s original dimension encompasses
ety on this dimension. at least two GLOBE dimensions—particularly,
Table 14.3 contains sample items from the Aggressiveness and Gender Egalitarianism. Later
societal practices scale that was developed to in this chapter we compare the current find-
measure participating managers’ perceptions of ings on Gender Egalitarianism with Hofstede’s
current Gender Egalitarianism practices in their (1998, 2001) most recent findings on masculinity/
societies. Items are worded exactly as adminis- femininity.
tered, with emphases added to highlight the
focus on practices.
Table 14.4 contains sample items from the GLOBE FINDINGS
societal values scale. These items are worded ON GENDER EGALITARIANISM
exactly as administered, with emphases added
to highlight the focus on values. Mean scores for the GLOBE Gender Egalitari-
Table 14.5 contains a sample item from the anism scales must be interpreted differently from
organizational Gender Egalitarianism practices mean scores for other GLOBE cultural scales.
14-HOUSE.qxd 3/19/2004 4:47 PM Page 361
Table 14.5 Sample Item From the Gender Egalitarianism Scale-Organizational Practices (As Is)
1-17. In this organization, men are encouraged to participate in professional development activities
more than women.
Strongly Neither agree Strongly
agree nor disagree disagree
1 2 3 4 5 6 7
Table 14.6 Sample Item From the Gender Egalitarianism Scale-Organizational Values (Should Be)
All GLOBE cultural items were measured on education. A score of 4 indicates indifference
7-point scales, and all are constructed such that (“neither agree nor disagree”), but a score of
7 is the conceptual maximum response, with the 7 indicates “strong” disagreement that boys
exception of Gender Egalitarianism. Its con- should be favored in this regard. Thus, for this
ceptual maximum for most items is 4, or the item, the most gender egalitarian response is not
midpoint of the scale. For these items, a score of one in which the respondent indicates essentially
1 indicates strong male domination—for no opinion but, instead, one in which the respon-
example, “strong” agreement that leadership dent registers strong disagreement.
opportunities are (in the case of practices) or The interpretive complexity of the items used
should be (in the case of values) more available to measure Gender Egalitarianism relative to the
for men than for women, whereas a score of 4 on other cultural dimensions in GLOBE reflects
this same item indicates that leadership opportu- its ongoing evolution. The response scales
nities are or should be equally available for men for Gender Egalitarianism items were origi-
and women. A score of 7 indicates a belief that nally written to reflect Hofstede’s masculinity/
there is or should be, in contrast, preferential femininity dimension: that is, to differentiate
treatment of women who seek positions of lead- societies in terms of the degree to which they
ership. Whether this reflects a belief that women were masculine or feminine. This rationale is
are better suited to lead or a desire to redress past reflected in the original GLOBE name for this
discrimination is uncertain. Regardless, a score dimension: Gender Differentiation. But as we
of 4 represents equal treatment of men and discussed in Chapter 8 (by Hanges & Dickson),
women or maximum gender egalitarianism for analyses performed on data from the two pilot
this item. This midpoint as the conceptual maxi- studies quickly clarified that our Gender Differen-
mum is also true for the majority of items that tiation items did not measure a single cultural
comprise the Gender Egalitarianism scales. construct but two conceptually distinct, albeit
There are a couple of exceptions to this rule, correlated, ones. The Hofstede masculinity/
however. One example is the item that addresses femininity scale correlated with the subset of
opportunities for boys versus girls to attain a GLOBE cultural items that were subsequently
higher education. For this item, a score of 1 indi- labeled Assertiveness, whereas the Schwartz
cates “strong” agreement that boys are (in the egalitarianism scale correlated with the subset of
case of practices) or should be (in the case of val- GLOBE cultural items that were subsequently
ues) encouraged more than girls to attain a higher labeled Gender Egalitarianism.
14-HOUSE.qxd 3/19/2004 4:47 PM Page 362
Standard
Variable Mean Deviation Minimum Maximum Valid N
Gender Egalitarianism
society practices 3.37 0.37 2.50 4.08 61
Gender Egalitarianism
society values 4.51 0.48 3.18 5.17 61
Thus, our understanding of the Gender among respondents in their perceptions of current
Egalitarianism dimension evolved over time, gender egalitarian practices in their societies than
and we would argue that it will and should con- in their perceptions of current practices associ-
tinue to evolve over the course of future studies. ated with any other cultural dimensions, with the
One recommendation for future studies would exception of Assertiveness.
be to construct items that use the full 7-point
response scale, with 7 reflecting the greatest
Societal Values
degree of Gender Egalitarianism.
GLOBE respondents indicate that their soci-
eties should be (M = 4.51) less male dominated
Societal Practices than they are now (M = 3.37). This can be inter-
Across all societies surveyed in GLOBE preted to mean that GLOBE respondents gener-
(N = 61), the mean (M = 3.37) and standard devi- ally prefer their societies not to favor men in
ation scores (SD = 0.37) for Gender Egalitarian- attainment of higher of education or leadership
ism societal practices are lower than for all other positions.
cultural dimensions. The mean scores for all other
cultural dimensions range from 3.85 to 5.17, Summary of Societal
which results in a grand mean of 4.25 for all
Practices and Values
dimensions for all GLOBE countries.
The maximum societal practices score is 4.08. Table 14.7 contains the summary statistics
This score does not differ significantly from the for both of the Gender Egalitarianism societal-
midpoint (4.00). Thus, no society in GLOBE is level scales. Taken together, these two mea-
perceived to be female dominated to the point of, sures revealed a paradoxical world in which
for example, encouraging girls, more so than individuals indicate that biological sex dictates,
boys, to attain a higher education or of having in part, the roles that are appropriate for
more women than men in positions of high office. females and males, despite believing that bio-
When interpreted in the context of the specific logical sex should not dictate these roles.
items that comprise the Gender Egalitarianism However, these results must be interpreted with
measure, the mean score of 3.37 indicates a mod- caution because it is unclear whether managers
est, cross-cultural reliance on biological sex in the responded in socially desirable ways when
allocation of roles between females and males. reporting how society should be with regard to
Interestingly, the standard deviation of 0.37 gender egalitarianism. In addition, the results
for Gender Egalitarianism societal practices is tied may be affected by the current composition of
with Assertiveness for the lowest standard devi- managers in these societies. Those groups who
ation score among the GLOBE societal culture benefit most from gender egalitarian practices
dimensions. Because standard deviation is an may endorse such practices, whereas those groups
indication of variability, the lower score for Gender who perceive themselves as either unaffected
Egalitarianism indicates greater agreement or harmed by gender egalitarian practices may
14-HOUSE.qxd 3/19/2004 4:47 PM Page 363
report a desire for a less gender egalitarian overall societal practices score for each society
society. However, if the societal values and the scores for the participating indus-
measures accurately reflect managers’ attitudes tries, collapsing across all societies surveyed.
and if these managers are in positions to effect As seen in this table, all coefficients are sig-
change, we may begin to see a movement nificant, thereby indicating consistency across
in many societies toward greater gender egali- industries in managers’ perceptions of their soci-
tarianism. For example, we may begin to see eties’ current practices with respect to gender
girls and boys encouraged more equally to egalitarianism.
attain a higher education and to see more Table 14.9 lists the correlation coefficients
women in leadership positions. for the overall societal values score for each
society and the scores for the participating
Overall Scores and Industry Scores. As noted industries, collapsing across all societies sur-
earlier, GLOBE society co-investigators gathered veyed. Again, all coefficients are significant,
societal culture data in three industries: Food, indicating that the managers’ perceptions of
finance, and telecommunications. Table 14.8 how gender egalitarian their societies should
below lists the correlation coefficients for the be are generally consistent across industries. In
14-HOUSE.qxd 3/19/2004 4:47 PM Page 364
other words, all managers surveyed report unrelated to their values with regard to Gender
similar feelings about how their societies should Egalitarianism.
divide roles between women and men, regard-
less of the industry in which they work.
GLOBE GENDER EGALITARIANISM
The Correlation Between Overall Practices AND OTHER ECONOMIC AND
and Values Scores. Managers’ perceptions of SOCIAL INDICATORS
their societies’ practices and values with respect
to Gender Egalitarianism are significantly In this section, we will compare and contrast
correlated (r = .32, p < .05). This coefficient the GLOBE findings on Gender Egalitarianism
constitutes a medium-sized effect (Cohen, 1988). with those of other major cross-cultural and
Thus, the more gender egalitarian a society’s comparative studies that focused on key eco-
current practices, the more gender egalitarian a nomic and social indicators. As Javidan and
manager’s values. Of course, this coefficient Hauser explained in Chapter 7, data for these
does not provide insight into the discrepancy indicators were taken from the following five
(D) between practices and values in a society. reports: The Human Development Report, the
As noted earlier, managers overall express Global Competitiveness Report, the World
a desire for their societies to be more gender Competitiveness Yearbook, the World Develop-
egalitarian than they are currently (D = 1.14). ment Indicators, and the World Values Survey.
Managers from Zambia and the Philippines Key indicators from these studies are grouped
are most prototypical in this respect. Their data and then examined in the following two cate-
points fall closest to the regression line, reflect- gories: Economic Health and Human Condition.
ing their desire for their societies to be some-
what more gender egalitarian than they are now
Gender Egalitarianism
(Ds = 1.45 and .94, respectively). Less proto-
and Economic Health
typical are managers from Qatar, who express a
desire for their society to be slightly less gender In our review of the literature, we discovered
egalitarian than it is currently (D = −.25). that the relationship between economic devel-
Tables 14.10a and 14.10b contain Gender opment and the status and roles of women in
Egalitarianism practices for both sets of scores, societies was more equivocal than scholars orig-
and countries are sorted into “bands.” Countries inally envisioned. The modernization perspec-
in band A score higher on Gender Egalitarianism tive had dominated much of past research, with
than do countries in band B, which in turn score its assumption that economic development cre-
higher on Gender Egalitarianism than do coun- ates opportunities for women to take on nontra-
tries in B and C, and so forth. Within a given ditional roles in society in general and in the
band—for example, Hungary and Hong Kong in labor force in particular (Moore & Shackman,
band A (see Table 14.10a)—there are no signif- 1996). However, evidence gathered to date sug-
icant differences in countries’ scores on Gender gests that economic development may instead
Egalitarianism. Across different bands—for propel women into performing their traditional
example, Hungary in band A and New Zealand (service) roles for pay in what some have
in band B—society scores differ significantly. termed “pink-collar ghettos” (Nuss & Majka,
1983, p. 42), a finding that is consistent with the
Overall Scores and Industry Scores. Although economic competition perspective. Although
overall scores on the societal level practices past research is equivocal with respect to the
and values scales are significantly correlated relationship between economic health and the
(r = .32, p < .05), industry scores on these two division of roles between women and men, we
scales are not significantly correlated except nonetheless hypothesize the following:
for the food industry, as displayed in Table
14.11. Thus, within financial and telecom • Hypothesis 14.1: Societies that are more
industries, the participating managers’ assess- gender egalitarian would be more prosperous
ments of their societies’ current practices are economically.
14-HOUSE.qxd 3/19/2004 4:47 PM Page 365
Band
A B C
Band
A B C D
level of consumption and growth in a society. societies’ practices with regard to Gender
The second is economic productivity, which Egalitarianism are not correlated with any of
reflects the favorableness of the labor environ- these three indicators. It is possible that the low
ment and gains in worker productivity. The standard deviation associated with this measure
third indicator, GNP per capita, represents the limits the potential to uncover links between it
Gross National Product per person in a society. and these measures. In contrast, managers’ val-
Participating managers’ perceptions of their ues positively correlate with all three indicators,
14-HOUSE.qxd 3/19/2004 4:47 PM Page 367
Telecommunication .29
industry score values N = 32
indicating that the more gender egalitarian a health and a desire to minimize gender-role
society’s values, the healthier its economy. Two differences. The modernization perspective
additional indicators reflect the support for eco- assumes that economic growth causes a cultural
nomic progress in a society. The first indicator, shift in attitudes and practices surrounding the
government support for prosperity, reflects the participation of women and other disadvan-
extent to which government programs and regu- taged groups in the labor force. The reverse
lations facilitate economic progress. Societal causal explanation is that a cultural shift in gen-
support for competitiveness, the second indica- der egalitarianism fuels economic growth by
tor, reflects society members’ attitudes toward prompting a society to capitalize more fully on
economic progress. Neither Gender Egalitarian- the talents of all members. In light of the pre-
ism measure is significantly correlated with sent findings, this latter explanation seems less
these support measures. Finally, the world com- plausible than the modernization perspective’s
petitiveness index—a measure of the extent to one, because economic health is positively cor-
which various factors such as a society’s related with egalitarian values, but not prac-
domestic economy, government, and infra- tices, in Project GLOBE. The notion that a
structure support economic competitiveness—is cultural shift in gender egalitarianism precedes
also uncorrelated with the two measures of economic development would be more plausi-
Gender Egalitarianism. ble if current practices were also positively cor-
To summarize, the findings in this section related with the various measures of economic
provide partial support for Hypothesis 14.1. health.
The three significant and positive correlations
are consistent with the modernization perspec-
Gender Egalitarianism
tive discussed earlier (Moore & Shackman,
and Human Condition
1996). This perspective contends that economic
growth creates new opportunities and roles for Indicators of economic health reflect one
women, thereby minimizing gender-role differ- aspect of overall well-being in any society.
ences. Of course, the correlational nature of the Another important aspect is the general health
present findings precludes any conclusions of a society’s members. Consistent with Hauser
about the direction or, even, presence of a causal and Javidan’s terminology in this volume, we
relationship between a society’s economic use the term human condition here to describe
14-HOUSE.qxd 3/19/2004 4:47 PM Page 368
Human
Gender Societal Human Life Psychological General Development
Egalitarianism Health Health Expectancy Health Satisfaction Index (HDI)
the general quality of life in a society. We Gender Egalitarianism are more consistently and
explore six indicators in this section: Societal strongly correlated with various human condi-
health, human health, life expectancy, psycho- tion indicators than are their perceptions of the
logical health, general satisfaction, and the 1998 current practices in their societies. Specifically,
Human Development Index to test the following managers’ perceptions of current practices in
correlational hypotheses: their societies significantly correlate with only
one of the indicators—the 1998 Human
• Hypotheses 14.2 and 14.3. Societies that are Development Index. Thus, the more gender
more gender egalitarian would enjoy higher egalitarian a society’s current practices, the
levels of human development and psycho- greater its members’ longevity, knowledge, and
logical health. standard of living. Managers’ values are more
strongly correlated, however, with the 1998
In Table 14.13 we find a pattern of correla- Human Development Index. These same values
tions similar to those found between the Gender are also significantly and positively correlated
Egalitarianism measures and various indicators with two other indicators of human condition—
of countries’ economic health. Specifically, life expectancy and general satisfaction. These
participating managers’ values with regard to correlations indicate that the more gender
14-HOUSE.qxd 3/19/2004 4:47 PM Page 369
egalitarian a society’s values, the longer its between religion and concepts related to gender
members’ life expectancies, and the happier and egalitarianism. Gender Egalitarianism practices
more satisfied their lives. and values scores are not significantly corre-
To summarize, managers’ values with regard lated with the strength of religious devotion or
to Gender Egalitarianism are more strongly and dogma in the societies sampled in the GLOBE
consistently associated with key indicators of project (p > .10).
human condition than are managers’ percep- In contrast to the results associated with
tions of their societies’ practices. The more gen- religious ideology, the Gender Egalitarianism
der egalitarian a manager’s values, the greater measures are significantly correlated with
the longevity in their societies, the more knowl- several measures of societies’ political ideolo-
edgeable and satisfied their societies’ members, gies. In Table 14.14, we find that the members
and the greater their standard of living. This pat- of societies with more Gender Egalitarian
tern of findings suggests that modernization practices express both greater disdain for and
more generally, rather than economic modern- dislike of democracy. At first glance, these
ization specifically (Moore & Shackman, 1996), correlations seem counterintuitive. Earlier in the
may serve to minimize gender-role differences. chapter, though, we described a study that
As discussed in the previous section on eco- examined the impact of countries’ political
nomic health, however, the causal chain of systems on women’s representation in national
events may run in reverse such that minimized legislatures (Paxton, 1997). Paxton and others
gender-role differences may prompt positive (e.g., Lovenduski & Hills, 1981; Lovenduski &
changes in the overall well-being of a society’s Norris, 1989; Rule, 1987) argued that simple
members. Again, however, the fact that man- plurality systems such as the United States’ pose
agers’ values with regard to Gender Egalitarian- an obstacle to female officeholding, because the
ism are more strongly and consistently related focus on individual candidates (rather than par-
to the various measures of human condition ties) makes candidate gender salient in the
than are their perceptions of current practices minds of voters. To the extent that voters stereo-
renders this latter explanation less convincing type women as less politically astute or able than
than the one offered in the modernization men, or embrace more traditional gender-role
perspective (Moore & Shackman, 1996). ideologies that place women in the home rather
than in political office, increasing the salience
of women candidates’ gender would decrease
Religion and Political Ideology
their representation in national legislatures.
Earlier, we reviewed several studies that pre- Consistent with this argument, Paxton (1997)
sented a conflicting account of the relationship found greater percentages of women legislators
14-HOUSE.qxd 3/19/2004 4:47 PM Page 370
Standard Standard
Climate Mean N Deviation Mean N Deviation
Standard Standard
Climate Mean N Deviation Mean N Deviation
expressed the most Gender Egalitarian values. feminine its culture. Parental investment theory
In contrast, managers from Iran, Qatar, and (Trivers, 1972) has been invoked as a potential
Turkey, which experience a desert climate explanation for this relationship, with scholars
(M = 4.14), expressed the least Egalitarian val- noting that the colder and less hospitable the
ues. Table 14.15 provides summary statistics for climate, the greater the need for both women
all seven climates. and men to invest in their offspring (e.g., Van de
Vliert & Van Yperen, 1996). This greater need
for cooperation between men and women is
Ambient Temperature
thought to result in a more egalitarian division
Earlier in the chapter we noted that ambient of roles between the sexes (Coltrane, 1992,
temperature has been identified as a key p. 94).
antecedent of cross-cultural differences in mas- To examine this possibility, we correlated
culinity (Hofstede, 1980; Peterson & Smith, Gender Egalitarianism practices and values
1997; Van de Vliert, 1998; Van de Vliert et al., scores with recent measures of the average
1999; Van de Vliert & Van Yperen, 1996). Past ambient or daytime temperature (in Fahrenheit)
studies have demonstrated that the lower a soci- in countries’ capital cities. Consistent with past
ety’s average daytime temperature, the more research in this area and parental investment
14-HOUSE.qxd 3/19/2004 4:47 PM Page 374
theory (Trivers, 1972), ambient temperature is Nonetheless, there were significant cross-
significantly and negatively correlated with regional differences in Gender Egalitarianism
both Gender Egalitarianism practices and values values. The Germanic Europe (M = 4.91) and
(see Table 14.16). Thus, the lower a society’s Anglo (M = 4.91) clusters scored highest on this
average daytime temperature, the more gender scale, whereas the Middle East (M = 3.65) clus-
egalitarian the society. ter scored lowest. Interestingly, although the
Eastern Europe cluster scored highest on Gender
Egalitarianism practices, they fell midrange rel-
Geographic Region
ative to the other clusters’ Gender Egalitarianism
Another key demographic characteristic is a values scores. In contrast, the Germanic Europe
society’s geographic region. Geographic region cluster scored among the lowest on Gender
is a broader demographic measure than is ambi- Egalitarianism practices but scored the highest
ent temperature, because countries within the on Gender Egalitarianism values. Thus, when
same geographic region often not only share talking about potential cross-regional differ-
similar climates, but also similar histories, reli- ences in gender egalitarianism, it is critical to
gions, and peoples compared with countries in distinguish between practices and values with
different geographic regions. In GLOBE, we regard to women’s and men’s roles in society.
classified each society into one of the follow-
ing 10 culture clusters: Eastern Europe, Latin
America, Latin Europe, Confucian Asia, Nordic GENDER EGALITARIANISM AND
Europe, Anglo, Sub-Saharan Africa, Southern OTHER GENDER-RELATED MEASURES
Asia, Germanic Europe, and Middle East. The
results of separate one-way analyses of variance In the introduction to this chapter, we noted that
(ANOVA) for both Gender Egalitarianism gender egalitarianism is related to several con-
practices (F[9,51] = 4.39, p < .01) and values cepts, including masculinity/femininity, gender-
(F[9,51] = 9.59, p < .01) yield significant cross- role stereotypes, gender-role ideology, gender
regional differences. discrimination, and gender equality. In this
The nature of these differences can be seen section, we examine the relationship between
in Table 14.17. The Eastern Europe (M = 3.84) GLOBE’s measures of Gender Egalitarianism
cluster scored highest on Gender Egalitarianism and several of these concepts.
practices, followed by the Nordic Europe (M =
3.71) cluster. In contrast, the Middle East (M = Hofstede’s Masculinity/
2.95), Confucian Asia (M = 3.18), and Germanic Femininity Dimension
Europe (M = 3.14) clusters scored lowest on this
same scale. Managers in every culture cluster Hofstede (1980) developed the masculinity/
reported that they would like their societies to be femininity dimension of societal culture by
more gender egalitarian than they are currently. asking managers in his IBM sample to rate the
14-HOUSE.qxd 3/19/2004 4:47 PM Page 375
Standard Standard
Geographic Region Mean N Deviation Mean N Deviation
(Continued)
14-HOUSE.qxd 3/19/2004 4:47 PM Page 376
Standard Standard
Geographic Region Mean N Deviation Mean N Deviation
Hofstede adopted the masculinity/femininity roles between men and women. This did not
label for two reasons. First, he argued that past prove to be the case, however. Earlier in this
research on work goals had demonstrated con- chapter, we asserted that Hofstede’s measure of
sistent differences between the women’s and masculinity/femininity confounds assertiveness,
men’s work goals, with women emphasizing gender egalitarianism, humane orientation, and
social goals associated with building strong achievement orientation, thereby yielding find-
and caring relationships and men emphasizing ings that are difficult to interpret. To the extent
ego goals or personal achievement. Second, he that this assertion is true, there may simply be
found that masculinity/femininity was the only too much variance in masculinity/femininity
dimension of societal culture that yielded signif- that is unrelated to Gender Egalitarianism to
icant and consistent differences between women uncover any variance that is shared between the
and men respondents in the IBM sample two measures.
(Hofstede, 1980, 1998, 2001).
In an earlier section, we explained that Gender Stereotypes
Hofstede’s masculinity/femininity dimension and Gender-Role Ideologies
was divided into two separate dimensions in
GLOBE: Assertiveness (see Den Hartog, Earlier in the chapter we introduced a two-
Chapter 15) and Gender Egalitarianism. Despite component model to illustrate the larger con-
this separation, we expected to discover a sig- text of Gender Egalitarianism (refer back to
nificant, albeit imperfect, correlation between Figure 14.1). The first component—the attitu-
Hofstede’s (1998, 2001) society scores for mas- dinal domain—reflects the fundamental val-
culinity/femininity and GLOBE’s society scores ues, beliefs, and attitudes that members of a
for Gender Egalitarianism practices and values. society hold with regard to gender stereotypes
To test this prediction, we computed the corre- and gender-role ideologies. Gender stereotypes
lations among these measures for the 47 coun- and gender-role ideologies represent parallel
tries included in both research projects. Contrary constructs, because stereotypes about the psy-
to our prediction, Hofstede’s masculinity/ chological attributes of women and men and
femininity scores are uncorrelated with scores ideologies about the roles that women and
on both Gender Egalitarianism scales. The cor- men can and should play in society are both
relation between Hofstede’s dimension and grounded in a powerful feedback loop (Best &
Gender Egalitarianism practices occurs in the Williams, 1993). Throughout this chapter, we
predicted direction, however, with the more have argued that this feedback loop occurs
feminine societies tending also to engage in within the larger context of societal norms
more gender egalitarian practices (r = −.17, regarding the division of roles between the
p > .10). This coefficient approaches, but does sexes. In this section, we examine gender
not achieve, significance, and so it should be stereotypes and gender-role ideologies sepa-
interpreted cautiously. In contrast, the corre- rately as potential key correlates of these
lation between masculinity/femininity and societal norms.
Gender Egalitarianism values is essentially zero
(r = −.02, p > .10). The absence of any link Gender Stereotypes. In our earlier discussion
between these two measures is particularly of gender stereotypes, we briefly described a
unexpected given that managers were instructed 25-nation study in which university students
to think in terms of an “ideal” when completing completed a 300-item adjective checklist
both—the ideal imaginary job for Hofstede’s (Williams & Best, 1990a). For each adjective
masculinity/femininity measure and ideal soci- or trait, students indicated whether it was
etal practices for GLOBE’s Gender Egalitarian- “more frequently associated with men than
ism values measure. Thus, we had greater with women, more frequently associated with
reason to predict a significant correlation women than with men, or not differentially
between Hofstede’s measure and values, rather associated with the two sexes” (p. 21). They
than practices, with regard to the division of then asked the students to rate these same traits
14-HOUSE.qxd 3/19/2004 4:47 PM Page 378
Table 14.18 Relationship Between Gender Egalitarianism, Gender Stereotypes, and Gender-Role
Ideologies
Activity Strength
Favorability of of Male’s Female’s
Gender of Female Female Female Gender-Role Gender-Role
Egalitarianism Traits+ Traits+ Traits+ Ideologies++ Ideologies++
+ Measure calculated by subtracting ratings of female traits from ratings of male traits; thus, a lower score indicates more
positive stereotypes of females.
++ Higher scores indicate more modern gender-role ideologies.
** Correlation significant at p < .05 level (2-tailed).
** Correlation significant at p < .01 level (2-tailed).
on the following three dimensions: Favorability, ranging from −.04 to .05). Thus, the model
activity, and strength. These ratings revealed proposed earlier (see Figure 14.1) receives par-
that male-stereotyped traits such as “adult” and tial support: The more egalitarian a society’s
“achievement-oriented” were rated more posi- values (but not practices), the more favorable
tively on these three dimensions than were members’ stereotypes of females relative to
female-stereotyped traits such as “child-like” their stereotypes of males.
and “affiliation-oriented,” although there were
cross-cultural differences. Based on the concep- Gender-Role Ideologies. The second facet of the
tual model outlined earlier (see Figure 14.1), we attitudinal domain of gender egalitarianism is
proposed that this gender gap would be smaller gender-role ideologies (see Figure 14.1). As
in societies whose practices and values are more noted earlier, individuals’ gender-role ideolo-
(vs. less) gender egalitarian. gies vary along a continuum of “traditional” to
Correlation analyses reveal that the measure “modern” (Best & Williams, 1993). Traditional
of Gender Egalitarianism values is, in fact, ideologies view men as more important than
strongly and negatively correlated with the gaps women and advocate relationships in which
or mean differences in both the perceived favor- men dominate and control women. In contrast,
ability (r = −.58, p < .01) and strength (r = −.53, modern ideologies view men and women as
p < .01) of traits ascribed to males versus equals and advocate egalitarian relationships
females (see Table 14.18). No significant corre- between them. For this reason, we hypothesized
lation was found for the activity dimension, a significant and positive relationship between
however. Thus, as predicted, the more gender GLOBE’s measures of Gender Egalitarianism
egalitarian a society’s values, the more favor- and the modernity of gender-role ideologies.
ably and strongly its university students per- In a 14-nation study of gender-role ideolo-
ceived “female” (relative to “male”) traits to be gies, approximately 100 university students
in Williams and Best’s study. In contrast, completed a 30-item survey of gender-role rela-
Gender Egalitarian practices scores are unre- tionships, using a 7-point scale (Williams &
lated to students’ perceptions of the favorability, Best, 1990b). The higher their overall ratings,
activity, and strength of these same traits (rs the more modern their gender-role ideologies.
14-HOUSE.qxd 3/19/2004 4:47 PM Page 379
These ratings were computed separately for men are treated equally but recognize that
men and women. Across countries, women women and men are not, in fact, treated
(M = 4.75) espoused more modern ideologies equally. From a purely statistical standpoint,
than did men (M = 4.41). Most relevant to the weaker correlations between various mea-
our investigation, correlational analyses yield sures and Gender Egalitarianism practices (vs.
partial support for the hypothesized positive values) may stem from the relatively low stan-
relationship between Williams and Best’s dard deviation associated with managers’ per-
(1990b) measures of gender-role ideologies and ceptions of current practices. As noted earlier,
GLOBE’s measures of Gender Egalitarianism. the standard deviation for Gender Egalitarian-
The correlation coefficients associated with ism practices (SD = 0.37) is the lowest among
managers’ perceptions of current practices and the nine GLOBE dimensions (SDs ranging
both men’s and women’s gender-role ideologies from 0.37 to 0.73). This low standard deviation
are nonsignificant (see Table 14.18). In contrast, may account for the dearth of significant rela-
the coefficients associated with managers’ tionships for this measure compared with the
values and both men’s and women’s gender- values measure (SD = 0.48).
role ideologies constitute “medium-to-large” To summarize, we argued earlier that gender
and “large” effects (Cohen, 1988), respectively, egalitarianism reflects and affects individuals’
although the former does not achieve statistical ideologies and stereotypes regarding women
significance. With only 14 countries repre- and men. We referred to this as the attitudinal
sented in Williams and Best’s study, the power domain of the gender egalitarianism construct.
to detect significant correlations with their The results from the previous two sections
measures and the GLOBE measures of Gender provide moderate support for this argument.
Egalitarianism is necessarily quite low. There- Stereotypes of women are more favorable and
fore, Cohen’s effect sizes labels may provide strong in societies in which managers report
the most meaningful gauge of the links between a desire to minimize gender-role differences.
these measures. These links suggest that Moreover, gender-role ideologies advocate
members of societies in which gender egalitari- greater importance, rights, and freedom for
anism is highly valued espouse greater equality women in these same societies. Taken together,
of women and men. then, we find strong convergent validity between
Similar to gender stereotypes, then, gender- GLOBE’s measure of Gender Egalitarianism
role ideologies were more closely related to values, and both gender stereotypes and gender-
managers’ values regarding the division of role ideologies.
roles between women and men than to their
beliefs about current practices in this area. This Women’s Participation in the
follows the overall pattern of stronger and Economy, Government, and Politics
more consistent relationships between Gender
Egalitarianism values and several key eco- In the conceptual model outlined in
nomic, social, and other indicators reported Figure 14.1 we proposed that attitudes regard-
earlier. As noted earlier, gender-role ideologies ing the appropriate roles of women and
reflect individuals’ “beliefs about proper men manifest themselves behaviorally in the
[emphasis added] role relationships between form of gender discrimination and equality.
women and men” (Williams & Best, 1990b, According to this argument, the more gender
p. 87). This emphasis on ideological pre- egalitarian a society, the greater women’s par-
ferences is echoed most closely in GLOBE’s ticipation in the economy, government, and
conceptualization of Gender Egalitarianism as politics. We examined this general thesis by
a value, or conviction, regarding the ideal roles computing the correlations between GLOBE’s
of women and men in a society. In contrast, the practices and values measures of Gender
conceptual link between gender-role ideologies Egalitarianism and four archival measures of
and Gender Egalitarianism practices may be women’s participation in various aspects of
more tenuous in that an individual may trumpet societies (see Table 14.19). The first archival
the virtues of a society in which women and measure is the rate of women’s economic
14-HOUSE.qxd 3/19/2004 4:47 PM Page 380
Table 14.19 Gender Egalitarianism and Women’s Participation in the Economy, Government, and Politics
activity, which is calculated as the percentage Gender Egalitarian a society’s values, the
of women in the labor force divided by the per- earlier a woman was appointed or elected to
centage of the society’s total population of its legislative body. Taken together, these
women age 15 years and older. Consistent with correlations between GLOBE measures of
the earlier hypothesized model, women’s eco- Gender Egalitarianism and various indicators
nomic activity is significantly and positively of women’s participation in key sectors of
correlated with managers’ perceptions of society provide support for the conceptual
the Gender Egalitarianism of their societies’ model proposed in Figure 14.1. In this model,
practices, but not their societies’ values. The individuals’ attitudes regarding what is possi-
second archival measure is the affluence or ble or appropriate for women and men delimit
purchasing power of women relative to men in their social roles, thereby affecting their partic-
a society. Not surprisingly, women’s purchas- ipation in the labor force, and in society over-
ing power is strongly and positively correlated all. The results from this section confirm that
with their rate of participation in the labor the more Gender Egalitarian a society’s values
force. Moreover, this power is positively in some instances, or practices in others, the
correlated with managers’ perceptions of both greater various indicators of gender equality
current Gender Egalitarianism practices and such as women’s participation and representa-
values regarding the division of roles between tion in the labor force, government, and poli-
women and men, although the latter correlation tics. These results provide support not only
does not achieve statistical significance. The for the conceptual model outlined earlier, but
third archival measure is the percentage of also for the validity of Gender Egalitarianism
women employed at all levels of government. as operationalized in the GLOBE project.
This measure is positively correlated with both
GLOBE measures of Gender Egalitarianism:
The more Gender Egalitarian a society’s prac- GENDER EGALITARIANISM AS A
tices and values, the greater the percentage of DIMENSION OF ORGANIZATIONAL CULTURE
women employed by the government. The
final archival measure is the year when the first Organizational culture is the glue that binds
woman was appointed or elected to a nation’s organization members together in pursuit of a
legislative body. Interestingly, this measure common goal. Scholars have argued that an
is negatively correlated with societies’ values, organization’s culture reflects the fundamental
but not practices, with regard to the division beliefs and values of its founder or, in some
of roles between women and men. The more cases, a transformational CEO who assumes
14-HOUSE.qxd 3/19/2004 4:47 PM Page 381
leadership at a critical point in the organization’s Egalitarianism practices have a significant and
history (Schein, 1983). Other scholars have strong positive relationship with organizational
noted, however, that an organization’s culture Gender Egalitarianism practices (p < .01). We
must also reflect the fundamental beliefs and found a similar significant and even stronger
values of the larger society in which it operates relationship between societal Gender Egalitari-
because, as Freud is often quoted, “that which anism values and organizational Gender Egalitari-
comes first is primary” (Adler, 1991; Schneider, anism values (p < .01). Both analyses support a
1988). Placed in the context of a discussion of principal proposition in the GLOBE theoretical
culture, Freud’s axiom acknowledges that model (i.e., Proposition 3, Figure 2.1, Chapter 2,
people typically identify more closely with their by House and Javidan): Societal cultural values
societies’ cultures than with their organizations’ and practices affect organizational cultural
cultures. For this reason, collisions between values and practices.3
societal and organizational cultures are often To summarize, societal Gender Egalitarianism
dramatic and noteworthy. A prime example is practices predicted organizational practices, and
the collision that occurred between the Disney societal Gender Egalitarianism values predicted
Corporation, with its taboo against alcoholic organizational values. Thus, the results from
beverages in its family-oriented theme parks, the two HLM analyses provide support for the
and the French culture in which wine is often general hypothesis that organizations’ cultures
served in family settings. Predictably, Disney, are, to some degree, a function of their societies’
not France, modified its culture to ensure the cultures regarding the division of roles between
survival of its then-floundering Euro Disney women and men.
theme park.
INTERPRETATION OF HLMS
THE RELATIONSHIP BETWEEN SOCIETAL USING GLOBE GENDER
AND ORGANIZATIONAL PRACTICES AND EGALITARIANISM TO PREDICT
SOCIETAL AND ORGANIZATIONAL VALUES CULTURALLY ENDORSED LEADERSHIP
In this section, we report the findings on the In the GLOBE conceptual model (House et al.,
relationship between societal and organiza- 1999), culturally endorsed leadership theories
tional Gender Egalitarianism practices and val- (CLTs) are determined, in part, by cultural
ues with regard to the division of roles between practices and values with regard to Gender
women and men. As House and Javidan indi- Egalitarianism, at both the societal and organiza-
cated in Chapter 2, GLOBE’s conceptual model tional levels. For this reason, we hypothesized
postulates that organizational practices and that managers’ implicit theories about the attrib-
values reflect, in part, the practices and values of utes associated with effective leadership would
the societies in which they operate, and so be related to the practices and values measures
we expected to find a link between the societal of Gender Egalitarianism at both the societal
and organizational measures of Gender Egalitar- and organizational levels. This chapter presents
ianism. We used hierarchical linear modeling relationships between one culture dimension
(HLM) to test for this link in the model. (Gender Egalitarianism) and the six CLTS.
Specifically, we conducted two HLMs to test Competitive tests of all culture dimensions and
these hypotheses for organizational Gender CLTs are presented in Chapter 21 by Dorfman
Egalitarianism practices and values. We tested and colleagues.4 In general, we expect that
the GLOBE hypothesis regarding the effect of societal and organizational values will be more
societal culture on organizational culture by con- strongly related to CLT leadership dimensions
ducting HLM analyses in which organizational than societal and organizational practices. As
Gender Egalitarianism was predicted by societal you read through the results discussed below, it
Gender Egalitarianism. These analyses sup- may be helpful to view Figure 14.2 for a visual
ported our hypotheses that societal Gender summary. The figure, however, only shows
14-HOUSE.qxd 3/19/2004 4:47 PM Page 382
Figure 14.2 Gender Egalitarianism Cultural Values as Drivers for CLT Leadership Dimensions
* Only statistically significant relationships are shown (p < .05; see Table 21.10, Chapter 21 by Dorfman et al.). The most
important leadership CLT relationships are bolded (i.e., relationship is significant at both society and organization levels of
analyses or highest HLM coefficient within each level of analysis).
O = Organizational level
S = Societal level
results regarding cultural values, not practices a total of 21.7% of the organizational and
(all HLM coefficients are presented in Table societal variance for this dimension. All of the
21.10 of Chapter 21 by Dorfman et al.) We explained variance was associated with forces
tested these hypotheses by using HLM analyses. operating at the societal level of analysis. The
Specific relationships are discussed below, but societal Gender Egalitarianism cultural values
as predicted, societal and organizational values scores were positively related (p < .01) to the
were more strongly related to most CLT Charismatic/Value-Based leadership dimension.
leadership dimensions than were societal and Charismatic/Value-Based leadership is more
organizational practices. The total amount of likely to be a part of the shared leadership belief
organizational and societal variance explained system in societies with high Gender Egalitari-
by Gender Egalitarianism ranges from 0 to anism values scores.
41.4%. Gender Egalitarianism explains the
greatest proportion of variance in Self-Protective • Team-Oriented leadership. Gender Egali-
and Participative leadership. It was not signifi- tarianism practices scores were significantly
cantly related to Autonomous leadership. related to the Team-Oriented leadership dimen-
More specifically, when organizational-level sion but explained a total of just 09% of the
and societal-level Gender Egalitarianism values organizational and societal variance for this
and practices were considered, significant rela- dimension. All of the explained variance was
tionships were found with the following: associated with forces operating at the organi-
zational level of analysis. The organizational
• Charismatic/Value-Based leadership. Gender Egalitarianism cultural practices scores
Gender Egalitarianism cultural values scores were positively related (p < .01) to the Team-
were significantly related to the Charismatic/ Oriented leadership dimension. Team-Oriented
Value-Based leadership dimension and explained leadership is more likely to be a part of the
14-HOUSE.qxd 3/19/2004 4:47 PM Page 383
was the values rather than the practices that supplemental analyses were limited to societal
predicted managers’ implicit theories of leader- culture samples in which both men and women
ship. Taken together, the findings from this were surveyed. Although the absence of female
section support the argument that societal and managers is perhaps telling in itself, it should be
organizational culture values influence the noted that gender bias is only one possible
desired level of Charismatic/Value-Based and explanation for why a particular sample does
Participative leadership in a positive manner, not include female respondents.
whereas they negatively influence the level of In addition to these considerations, a statisti-
Self-Protective leadership attributes. cal issue further limited the sample available for
this supplemental analysis. Because gender is a
dichotomous variable, relationships between
GENDER DIFFERENCES ON CLT gender and other variables such as leadership
LEADERSHIP RATINGS AND CULTURAL perceptions reflect not only actual mean rating
DIMENSION VALUES AND PRACTICES differences by men and women, but also differ-
ential proportions of men and women respon-
Given the discussion of gender differences dents in the samples across societies. As pointed
throughout the chapter, it is likely that males out by a number of statisticians (e.g., Cohen &
and females have different perceptions of what Cohen, 1983; Nunnally & Bernstein, 1993), the
is valued in a society. Indeed, the popular press maximum relationship possible between a
and the academic literature on gender and lead- dichotomous variable and a continuous variable
ership (cf. Eagly & Johannesen-Schmidt, 2001; decreases as the relative proportions of men and
Heilman, 2001; Schein 1983) indicate that male women vary from .50. If the different societies
and female managers view leadership roles varied dramatically in the percentage of women
quite differently, with men viewing hierarchical included in their samples, any significant gender-
and competitive interpersonal behaviors as by-culture interaction could simply reflect the
components of the leadership role and women differential proportions of men and women
viewing collaborative and consensus building across the GLOBE-sampled cultures. Because
behaviors as components of this role. we are interested in capturing only those gender
To test the possibility that gender differences differences that are conceptually meaningful,
would be found in GLOBE data, we conducted we included countries only if the percentage of
supplemental statistical analyses exploring female and male respondents was roughly equal.
differences between men and women in terms We identified 30 societies that met the afore-
of how they rated CLT attributes considered mentioned criteria. The full sample consisted of
important for outstanding leadership as well as 5,645 males (73.7%) and 2,012 females (26.3%).
how they rated societal Gender Egalitarianism Approximately one half of the societal samples
practices and values. These analyses involved could not be included in this analysis because
several steps. The first step was to identify they did not meet all the criteria.
the gender of GLOBE respondents; however, A series of hierarchical linear models
several countries did not provide demographic (see Chapter 11 by Hanges, Dickson, & Sipe)
information for any respondents in their were conducted to test whether men and
samples, or their samples consisted entirely of women rated the CLT leadership dimensions
men. In the first case, the lack of demographic differently. These analyses revealed signifi-
information was a conscious omission as the cant gender differences for four of the six
investigators had reason to believe that asking CLT dimensions. Specifically, women rated
demographic information would substantially Charismatic/Value-Based leadership (t (30) =
lessen the response rate, something akin to ask- 2.19, p < .05), Participative leadership (t (30) =
ing about personal income in questionnaires in 4.31, p < .01), and Team-Oriented leadership
the United States. Only those samples for which (t (30) = 3.78, p < .01) significantly higher than
the respondents’ gender could be identified did men. In contrast, men rated Self-Protective
were included in this analysis. Also, these leadership (t (30) = − 3.74, p < .01) significantly
14-HOUSE.qxd 3/19/2004 4:47 PM Page 385
higher than did women. It should be noted, CLT dimensions more similarly the more gender
however, that although these results are sta- egalitarian their societies.
tistically significant, the total variance accoun- In addition to gender differences on CLT
ted for by gender was small (i.e., 0.2% for dimensions, it is also possible that men and
Charismatic/Valued-Based; 0.9% for Partici- women might hold different perceptions of their
pative; 0.5% for Team-Oriented; and 0.7% societies’ cultural practices and values. The
for Self-Protective).5 Finally, there are no signi- gender role literature suggests that people’s
ficant gender differences for Humane-Oriented beliefs about the capabilities of males and
(t (30) = − 0.30, p > .05) or Autonomous (t (30) females are a function of how males and
= 1.16, p > .05) leadership. females are socialized in a society. The social-
Although the prior HLM analyses are infor- ization process emphasizes certain behaviors in
mative, it should be noted that these analyses children, and these gender-differentiated experi-
looked for a constant gender relationship on the ences will convey different “meta-messages”
CLT dimensions across countries. It may be about gender-appropriate behaviors. These
that gender differences interact with societal meta-messages are hypothesized to influence
culture such that certain societies would be the self-concept, personal goals, and cognitive
more prone to socialize men and women to dif- heuristics used by individuals in that society.
ferentiated leadership roles. In other words, it Thus, males and females will likely differen-
is possible that the gender effect differs across tially describe the practices and values of their
societal cultures. We therefore performed an societal cultures if they have experienced differ-
additional analysis to determine whether the ent socialization experiences in that society. We
gender effect on the CLT leadership dimen- explored this possibility with the GLOBE data.
sions differed across cultures. The HLM analy- Before discussing the results, however, it should
ses indicated that there were significant be noted that societal culture was measured only
interactions between gender and societal cul- on Form B of the GLOBE survey. Thus, the
ture for five of the six CLT dimensions. Only sample size for the gender differences on culture
the Humane-Oriented CLT dimension failed to is approximately one-half of the sample size
show a significant gender by culture interaction available for the prior analyses.
(t (30) = 27.32, p > .05). In general, the HLMs did not find gender
More specifically, gender differences on the differences when societal cultural practices
Charismatic/Value-Based CLT dimension signif- were examined. However, perhaps not surpris-
icantly differed across societal cultures (t (30) = ingly, societal Gender Egalitarianism cultural
45.39, p < .05). For example, the gender differ- practices were rated significantly differently, on
ences for this leadership dimension were minimal average, by men and women (t (16) = 2.98,
for Sweden but fairly substantial for Guatemala. p < .01). Women described their societies as
The gender effect also significantly differed for having slightly more gender egalitarian prac-
Participative (t (30) = 69.66, p < .01), Team- tices than did men. Although statistically signif-
Oriented (t (30) = 52.00, p < .01), Self-Protective icant, this effect did not account for a substantial
(R 2 (30) = 67.76, p < .05), and Autonomous (R 2 portion of societal Gender Egalitarianism prac-
(30) = 49.66, p < .05) CLT leadership dimen- tices (t = 0.2%). Interestingly, the HLMs also
sions. Although we could rank societies in terms revealed that gender differences did not differ
of the magnitude of their gender differences on across societies (t (16) = 0.00, p > .05). Overall,
these CLT scores, it would be more meaningful gender did not appear to affect ratings of societal
to identify one or more societal-level construct Gender Egalitarianism practices.
that might be causing these interactions. Similar results were obtained for societal
Although identifying such constructs is beyond Gender Egalitarianism values. Significant gen-
the scope of this chapter, interested readers der differences were found for societal Gender
should read Paris (2003) for a more complete Egalitarianism values (t (16) = 2.93, p < .05).
exploration of these cultural moderators. She Women described their societies as having
found that women and men rated the various more Gender Egalitarianism values than did
14-HOUSE.qxd 3/19/2004 4:47 PM Page 386
men. This gender effect accounted for 2.3% equality. The more gender egalitarian a society,
of the total variance in societal Gender the less gender discrimination and inequality
Egalitarianism.6 Moreover, this gender effect one would expect to find in key venues. The
differed as a function of societal culture relationship between the attitudinal domain and
(t (16) = 63.36, p < .05). the behavioral manifestations of gender egali-
The only other societal culture variable that tarianism comprises a recurrent feedback loop:
exhibited a significant gender effect was soci- Beliefs about what is possible or appropriate
etal Humane Orientation cultural dimension for women and men affect their treatment and
values (t (16) = −2.91, p < .01). Men described roles in homes, workplaces, and societies. Any
their societies as valuing Humane Orientation gender discrimination and inequality that
more than did women. Although statistically results serves to reinforce individuals’ stereo-
significant, this effect did not account for a types and ideologies, bringing us full circle.
substantial portion of the variance (R2 = 0.0%). We next reviewed prior research linking
In summary, we found few significant various concepts associated with gender egali-
gender differences when comparing men’s and tarianism and potential “cultural drivers.”
women’s ratings on all cultural practices and This review underscored the difficulty of
values, with the exception of one dimension. disentangling individual causes of gender
Not surprisingly, we found that societal Gender egalitarianism such as religion and political
Egalitarianism cultural practices and values systems and parental investment in children.
were sensitive to gender differences. The effect Past research has documented empirically a
of gender was stronger for cultural values than it strong positive relationship between men’s
was for cultural practices. parental investment and women’s status in
homes and societies. Similarly, parental invest-
ment has been identified as a key mediator of
SUMMARY AND CONCLUSIONS the relationship between ambient temperature
and gender equality: Colder climates require
We began this chapter by defining the construct men to invest more heavily in their offspring,
of gender egalitarianism as the way in which thereby promoting greater mixed-gender
societies divide roles between women and men. exchange and gender equality.
The more gender egalitarian a society, the less Past research presented a less straightfor-
it relies on biology to determine women’s and ward picture of the link between other cultural
men’s social roles. This construct varies along drivers and concepts related to gender egalitar-
a continuum in that societies can be rated as ianism. For example, the link between a soci-
more or less gender egalitarian relative to other ety’s religion and the status of women was
societies in the same study. More gender egali- equivocal. Monotheism did not have the same
tarian societies believe that men and women are impact on women’s status across cultures. In
suited for similar roles, whereas less gender addition, the overall religiosity of a culture
egalitarian societies believe that men and could not account for the division of roles
women should assume different roles. between the sexes. It may be that the differen-
We proposed a two-component model of tial status of men and women within religious
gender egalitarianism that included an attitudi- life may be reflected in their differential status
nal domain and behavioral manifestations. The in society overall. In the Catholic and Islamic
attitudinal domain included the concepts of faiths, religious leaders are almost always men,
gender stereotypes and gender-role ideology. whereas in other faiths leadership positions are
These concepts reflect individuals’ beliefs divided more equally between men and women.
about the qualities of women and men, and the It also has been argued, however, that beliefs
roles that they should assume, respectively. about appropriate roles for women and men
When individuals’ actions are consistent with might cause religious attitudes rather than the
their beliefs in these areas, we see the behavioral reverse.
manifestations of gender egalitarianism— The past literature on the link between
specifically, gender discrimination and gender nations’ economic health and women’s status
14-HOUSE.qxd 3/19/2004 4:47 PM Page 387
and roles is equally complex. The modernization members, despite the fact that most managers
perspective proposes that economic growth reported that this practice was not an ideal one.
creates additional work opportunities for If managers’ espoused values accurately reflect
women and other disadvantaged groups, whereas their deeply held values and they are in posi-
the economic competition perspective holds that tions to effect social change, then we may begin
gender inequality persists despite economic to see a movement in many countries toward
development because women do not possess increased gender egalitarianism.
sufficient power to capitalize on this develop- In addition, we explored the relationships
ment. Findings from past research are most between GLOBE’s measure of Gender Egalitari-
consistent with the economic competition per- anism and other social and economic indicators
spective: Although economic growth is asso- in these 62 societies and discovered several
ciated with a rise in women’s labor force significant correlations. Specifically, members of
participation, they tend to cluster in occupations societies whose practices are currently more gen-
traditionally deemed suitable for women— der egalitarian achieved greater longevity, knowl-
notably in service and in communication. Thus, edge, and standards of living. These members
economic growth seems simply to provide new also expressed greater disdain for democracy. A
venues—“pink-collar ghettos”—for women to potential explanation for this finding is that
pursue stereotypically female roles. members whose societies are currently more gen-
The ways in which resources are controlled der egalitarian may distrust democracy because
are also related to women’s status and power in the simple plurality system of what may be the
societies. Specifically, the intergenerational most visible and, hence, prototypical democracy
transfer of wealth is an institutionalized practice in the world, the United States, has been linked to
that affects the division of roles between men greater gender inequality in political representa-
and women. Women were ascribed much higher tion. This type of inequality would be incompati-
power and status in those few societies with ble with gender-egalitarian practices.
matrilineal descent and locality than in the Societies in which managers espoused more
many societies with patrilineal descent and gender-egalitarian values were more prosperous
locality. In one study, women fared better only economically. People living in these societies
if they had some control over property and had longer life expectancies and experienced
shared parental responsibility with men. greater overall satisfaction with their lives. They
Political systems were the last cultural dri- also acquired greater levels of knowledge and
ving force that we reviewed. Women were more enjoyed higher standards of living. Members of
likely to attain political offices in societies that societies that embraced more gender-egalitarian
favored a proportional representation system values expressed a desire for less government.
simply because the partisan nature of elections Although this finding seemed counterintuitive
may serve to decrease the salience of their bio- initially, it eventually wove together in our
logical sex. In addition, women made greater minds with two additional findings associated
gains in societies that tolerated the formation of with this measure. First, as already noted,
political opposition groups—a finding that was members of societies with stronger gender-
the first to pinpoint women’s solidarity as a egalitarian ideals enjoyed greater economic
powerful resource for women to gain greater prosperity, knowledge, longevity, satisfaction,
equality in their societies. and so forth. Consequently, these individuals
The GLOBE project surveyed 62 societies on may feel sufficiently empowered that they
nine cultural dimensions, including Gender desire less governmental “interference” in their
Egalitarianism. This dimension was measured lives. A sense of empowerment emerges
by surveying workplace managers regarding from the findings associated with various indi-
their perceptions of current practices and values cators of individuals’ political ideologies as
with regard to the division of roles between well. The more gender egalitarian a society’s
females and males. Overall findings suggested values, the more active, vocal, and open to
that societies currently rely to some extent on change were its members. If members feel
biological sex to allocate appropriate roles to that they can safeguard their own well-being
14-HOUSE.qxd 3/19/2004 4:47 PM Page 388
and effect change when necessary, they may We also examined Gender Egalitarianism as
have less desire for and tolerance of “big a dimension of organizational culture. An orga-
government.” nization’s culture can serve as the social glue
Our investigation of the relationship between that compels members of an organization to
gender egalitarianism and society demographics work collaboratively toward a single goal.
explored climate, ambient temperature, and Scholars have argued that organizational culture
geographic region. Findings indicated that soci- constitutes a microcosm of the larger society. A
eties with maritime climates such as Denmark comparison of societal and organizational levels
and New Zealand were most gender egalitarian of Gender Egalitarianism revealed that organi-
whereas those with desert climates such as Iran zational cultures reflect the culture (practices
and Qatar were least gender egalitarian. This is and values) in the society in which they are
consistent with the finding that the lower a soci- embedded. Societal values and practices pre-
ety’s average daytime temperature, the more dicted organizational values and practices
gender egalitarian its practices and values. With regarding the allocation of roles between
regard to geographic region, the Eastern Europe women and men.
and Nordic Europe clusters reported the highest An additional analysis focused on the
current levels of gender egalitarianism, whereas extent to which culturally endorsed implicit
the Middle East, Confucian Asia, and Germanic theories of leadership reflect, in part, organiza-
Europe clusters reported the lowest. In terms of tional and societal norms with regard to Gender
ideal levels of gender egalitarianism, however, Egalitarianism. The most consistent findings to
the Germanic Europe cluster reported the high- emerge from these analyses were that more gen-
est levels, followed closely by the Anglo clus- der egalitarian organizations and societies
ter. The Middle East cluster reported the lowest endorsed charismatic leader attributes such as
ideal gender egalitarianism. “foresight,” “enthusiastic,” and “self-sacrificial”
We devoted the next section to an examina- and participative leader attributes such as “egali-
tion of the relationships between gender egali- tarian,” “delegator,” and “collectively oriented.”
tarianism and various measures that were These same organizations and societies shunned
theorized to be conceptually related. Contrary to self-protective leader attributes such as “self-
our expectations, GLOBE’s measures of Gender centered,” “status-conscious,” “secretive,” “eva-
Egalitarianism were unrelated to Hofstede’s sive,” and “formal.” These results have important
(1980, 1998, 2001) measure of masculinity/ practical implications for leaders: Individuals
femininity. A potential explanation for this find- whose organizations and societies seek to mini-
ing is that masculinity/femininity confounds too mize gender-role differences must work to cul-
many societal norms to share sufficient, unique tivate charismatic and participative qualities and
variance with a measure designed specifically to extinguish self-protective qualities if they
to reflect beliefs about the division of roles wish to be perceived as effective leaders of their
between women and men. We did, however, organizations and societies.
find support for the notion that gender egalitari- For the final set of analyses, we found that
anism represents a powerful and overarching there were significant differences between male
cultural norm that grounds the feedback loop and female respondents regarding the rated
between individuals’ gender stereotypes and importance of CLT leadership dimensions,
ideologies on the one hand, and gender inequal- with men and women perceiving a greater
ity and discrimination on the other. Stereotypes contribution of Charismatic/Value-Based,
of women were more positive in societies that Participative, and Team-Oriented attributes to
embraced gender egalitarian values. Moreover, outstanding leadership. In contrast, women
greater gender equality was not only advocated perceived Self-Protective leadership attributes
in these societies in the form of more modern to be more detrimental to outstanding leadership
gender-role ideologies, but also achieved to a than did men. Finally, as might be expected,
greater degree in labor forces, governments, and gender effects differed significantly across
national politics. societal cultures. One plausible explanation lies
14-HOUSE.qxd 3/19/2004 4:47 PM Page 389
in our finding that gender differences in ratings court) brief with the Supreme Court. This APA brief
of various CLT dimensions increased with the outlined the current state of sex stereotyping and sex
level of gender inequality across societies. discrimination research as it pertained to the facts of
In conclusion, GLOBE’s measures proved the Price Waterhouse v. Hopkins case. The Supreme
useful for collecting and analyzing a wealth Court’s decision relied heavily on the brief, as evi-
of cross-cultural data. By gathering informa- denced by extensive references to the brief in the
tion regarding gender egalitarianism across 61 written opinion. The panel’s task did not end with the
societies and investigating its relationship Court’s ruling, however. Two scholars (Barrett &
with social, economic, and demographic factors, Morris) criticized the panel’s methods, conclusions,
we were able to identify the characteristics of and motives, thereby prompting a spirited and, at
societies that seek to minimize gender-role times, acrimonious debate in the scholarly literature
differences versus those that seek to maximize (Barrett & Morris, 1993a, 1993b; Fiske, Bersoff,
such differences. To the extent that managers Borgida, Deaux, & Heilman, 1991, 1993a, 1993b).
such as those who participated in this study are One wonders whether this debate—the fact that it
in positions to influence cultural norms, they occurred at all and was so contentious once begun—
may be able to serve as change agents by foster- stemmed, in part, from the taboo nature of issues
ing gender-egalitarian practices and values in surrounding gender egalitarianism.
their organizations and societies, thereby increas- 2. Our response bias correction procedure
ing the levels of prosperity and well-being in identified response bias in some countries for this
both realms. scale. We recomputed the predicted response bias
corrected-scale score for each country. Response bias
corrected scores are:
ENDNOTES Practices: Morocco, 3.08 (no change in band);
Qatar, 3.86 (no change in band); and Taiwan, 2.92
1. The case of Ann Hopkins v. Price Waterhouse (no change in band).
provides an interesting case in point of Hofstede’s Values: Finland, 4.47 (no change in band); France,
(1998) claim that masculinity/femininity (in Project 4.71 (moves from band B to band A); Morocco, 4.07
GLOBE, Assertiveness and Gender Egalitarianism) (no change in band).
constitutes the “taboo” dimension of societal culture. For a complete discussion of this procedure and
Very briefly, the case was one in which a woman, all response bias corrected scores, see Appendix B of
Ann Hopkins, charged Price Waterhouse with gender this volume.
discrimination after having been denied partnership 3. As reported in Chapter 20 by Brodbeck,
twice despite very positive work performance. Susan Hanges, Dickson, Gupta, and Dorfman, we found that
Fiske, a social psychologist, served as an expert all the cultural dimensions of organizational cultural
witness in this case. In her testimony, Fiske cited values and practices significantly differed across
empirical research on conditions that facilitate gender societies. Although important, this prior analysis did
stereotyping, indications and consequences of gender not identify the particular aspect of societal differ-
stereotyping, and ways to prevent it (Loftus, 1991). ences that was related to organizational culture. In the
As part of its appeal of a lower court ruling in Ms. present chapter, we found that societal and organiza-
Hopkins’s favor, Price Waterhouse challenged the tional Gender Egalitarianism practices were signifi-
scholarly integrity of Fiske, specifically, and of cantly related (R2 Total = 6.2%, R2 Societal = 34.4%,
social-scientific research, generally. An appellate p < .01). We found even stronger results for societal
court again ruled in Ms. Hopkins’s favor, and so and organizational Gender Egalitarianism values (R2
Price Waterhouse made one final appeal to the Total = 32.7%, R2 Societal = 90.8%, p < .01). As dis-
Supreme Court of the United States. In response to cussed in Chapter 11 by Hanges, Dickson, and Sipe, the
Price Waterhouse’s attack on social-scientific R2 Total considers all levels of analysis (i.e., individual,
research, the American Psychological Association organizational, and societal) whereas the R2 Societal
(APA) convened a panel of scholars (Fiske, Borgida, isolates the societal level portion of the dependent
Deaux, & Heilman) and one attorney (Bersoff) to variable and indicates the percentage of variance
construct and file an amicus curiae (friend of the accounted for by the predictor at only this level.
14-HOUSE.qxd 3/19/2004 4:47 PM Page 390
Although we have primarily taken the conservative Berk, R., & Berk, S. F. (1979). Labor and leisure at
approach and reported the R2 Total in GLOBE, home. Beverly Hills, CA: Sage.
a number of scholars suggest that R2 Societal Best, D. L., & Williams, J. E. (1993). A cross-cultural
provides a more accurate description of aggregated viewpoint. In A. E. Beall & R. J. Sternberg
relationships. For further discussion, see the paper by (Eds.), The psychology of gender (pp. 215–248).
Lance and James (1999). New York: Guilford Press.
4. Results between the single HLM and multiple Best, D. L., Williams, J. E., Cloud, J. M., Davis, S.
HLM tests will likely differ somewhat. The differ- W., Robertson, L. S., Edwards, J. R., Giles, H.,
ences between the results of the multiple HLMs & Fowles, J. (1977). Development of sex trait
and single HLMs are conceptually similar to the stereotypes among young children in the U.S.,
differences between a multiple regression analysis and England and Ireland. Child Development, 43,
a correlation coefficient. Table 21.10 in Chapter 21 by 1375–1384.
Dorfman et al. presents both single and multiple HLM Blackburn, R. M., Jarman, J., & Siltanen, J. (1993).
coefficients. In addition, the relationships for all cul- The analysis of occupational gender segregation
ture dimension values are summarized in Chapter 3. over time and place: Considerations of measure-
5. Gender differences accounted for 0.2%, 1.1%, ment and some new evidence. Work Employment
0.6%, 0.9% of variance at the individual level of and Society, 7, 335–362.
analysis in the Charismatic/Value-Based, Participative, Boserup, E. (1970). Woman’s role in economic devel-
Team-Oriented, and Self-Protective CLT leadership opment. London: Allen & Unwin.
dimensions, respectively. See Chapter 11 by Hanges, Braidotti, R. (1994). Nomadic subjects. New York:
Dickson, and Sipe for a discussion of these different Columbia University Press.
explained variances. Chafetz, J. S., & Hagan, J. (1996). The gender divi-
6. The individual level, explained variance for sion of labor and family change in industrial
gender on this variable was 3.1%. See Chapter 11 by societies: A theoretical accounting. Journal of
Hanges, Dickson, and Sipe for a discussion of this Comparative Family Studies, 27, 187–219.
explained variance term. Charles, M. (1992). Cross-national variation in occu-
pational sex segregation. American Sociological
Review, 57, 483–502.
REFERENCES Chinese Culture Connection. (1987). Chinese values
and the search for culture-free dimensions of
Adler, N. J. (1991). International dimensions of culture. Journal of Cross-Cultural Psychology,
organizational behavior (pp. 58–60). Boston: 18, 143–164.
PWS-Kent. Chodorow, N. (1978). The reproduction of mother-
Barrett, G. V., & Morris, S. B. (1993a). The ing. Berkeley: University of California Press.
American Psychological Association’s amicus Clark, R., & Carvalho, J. (1996). Female revolt revis-
curiae brief in Price Waterhouse v. Hopkins: The ited. International Review of Modern Sociology,
values of science versus the values of the law. 26, 27–42.
Law and Human Behavior, 17, 201–215. Clark, R. L., & Anker, R. (1993). Cross-national
Barrett, G. V., & Morris, S. B. (1993b). Sex stereo- analysis of labor force participation of older
typing in Price Waterhouse v. Hopkins. men and women. Economic Development and
American Psychologist, 48, 54–55. Cultural Change, 41, 489–512.
Barry, H., Bacon, M. K., & Child, I. L. (1957). A Cohen, J. (1988). Statistical power analysis for the
cross-cultural survey of some sex differences in behavioral sciences (2nd ed.). Hillsdale, NJ:
socialization. Journal of Abnormal and Social Lawrence Erlbaum.
Psychology, 55, 327–332. Cohen, J., & Cohen, P. (1983). Applied multiple
Becker, G. S., & Posner, R. A. (1993). Cross-cultural regression/correlation analysis for the behav-
differences in family and sexual life. Rationality ioral sciences (2nd ed.). Hillsdale, NJ: Lawrence
and Society, 5, 421–431. Erlbaum.
Bem, S. L. (1975). Sex role adaptability: One conse- Coltrane, S. (1988). Father–child relationships and
quence of psychological androgyny. Journal of the status of women: A cross-cultural study.
Personality and Social Psychology, 31, 634–643. American Journal of Sociology, 93, 1060–1095.
14-HOUSE.qxd 3/19/2004 4:47 PM Page 391
Coltrane, S. (1992). The micropolitics of gender in Girard, C. (1993). Age, gender, and suicide: A
nonindustrial societies. Gender and Society, 6, cross-national analysis. American Sociological
86–107. Review, 58, 553–574.
Coverman, S., & Sheley, J. F. (1986). Men’s house- Gray, J. P. (1987). Do women have higher social
work and child-care time, 1965–1975. Journal status in hunting societies without high gods?
of Marriage and Family, 48, 413–422. Social Forces, 65, 1121–1131.
Darwin, C. (1958). The origin of species by natural Heilman, M. E. (1995). Sex stereotypes and their
selection or the presentation of favoured races effects in the workplace: What we know and
in the struggle for life. New York: Mentnor. what we don’t know. Journal of Social Behavior
(Original work published in 1859) and Personality, 10, 3–26.
Denmark, F., Rabinowitz, V., & Sechzer, J. (2000). Heilman, M. E. (2001). Description and prescription:
Engendering psychology. Boston: Allyn & How gender stereotypes prevent women’s
Bacon. ascent up the organizational ladder. Journal of
Doney, P. M., Cannon, J. P., & Mullen, M. R. (1998). Social Issues, 57, 657–674.
Understanding the influence of national culture Heilman, M. E., Block, C. J., & Martell, R. F. (1995).
on the development of trust. Academy of Sex stereotypes: Do they influence perceptions
Management Review, 23, 601–620. of managers? Journal of Social Behavior and
Eagly, A. H., & Johannesen-Schmidt, M. C. (2001). Personality, 10, 237–252.
The leadership styles of women and men. Hendrix, L. (1994). What is sexual inequality? On the
Journal of Social Issues, 57, 781–787. definition and range of variation. Cross-Cultural
Erez, M. (1994). Toward a model of cross-cultural Research, 28, 287–307.
industrial and organizational psychology. In Hendrix, L., & Hossain, Z. (1988). Women’s status
H. C. Triandis, M. D. Dunnette, & L. M. Hough and mode of production: A cross-cultural test.
(Eds.), Handbook of industrial and organiza- Signs, 13, 437–453.
tional psychology (2nd ed., Vol. 4, pp. Hochschild, A. (1989). The second shift. New York:
559–607). Palo Alto, CA: Consulting Viking.
Psychologists Press. Hofstede, G. (1980). Culture’s consequences:
Fagenson, E. A. (1990). Perceived masculine and fem- International differences in work-related values.
inine attributes examined as a function of individ- Beverly Hills, CA: Sage.
uals’ sex and level in the organizational power Hofstede, G. (1991). Cultures and organizations:
hierarchy: A test of four theoretical perspectives. Software of the mind. London: McGraw-Hill.
Journal of Applied Psychology, 75, 204–211. Hofstede, G. (1998). Masculinity and femininity: The
Fiske, S. T., Bersoff, D. N., Borgida, E., Deaux, K., taboo dimension of national cultures. Thousand
& Heilman, M. E. (1991). Social science Oaks, CA: Sage.
research on trial: Use of sex stereotyping Hofstede, G. (2001). Culture’s consequences:
research in Price Waterhouse v. Hopkins. Comparing values, behaviors, institutions, and
American Psychologist, 46, 1049–1060. organizations across nations (2nd ed.).
Fiske, S. T., Bersoff, D. N., Borgida, E., Deaux, K., Thousand Oaks, CA: Sage.
& Heilman, M. E. (1993a). Accuracy and objec- Hofstede, G., & Vunderink, M. (1994). A case
tivity on behalf of the APA. American study in masculinity/femininity differences:
Psychologist, 48, 55–56. American students in the Netherlands vs. local
Fiske, S. T., Bersoff, D. N., Borgida, E., Deaux, K., students. In A. M. Bouvy, F. J. R. van de Vijver,
& Heilman, M. E. (1993b). What constitutes a P. Boski, & P. Schmitz (Eds.), Journeys into
scientific review? A majority retort to Barrett cross-cultural psychology (pp. 329–347). Lisse,
and Morris. Law and Human Behavior, 17, The Netherlands: Swets & Zeitlinger.
217–233. Hoppe, M. H. (1990). A comparative study of society
Fiske, S. T., & Taylor, S. E. (1991). Social cognition elite: International differences in work-related
(2nd ed.). New York: McGraw-Hill. values and learning and their implications
Gilmore, D. D. (1996). Above and below: Toward a for management training and development.
social geometry of gender. American Anthropo- Unpublished doctoral dissertation, University of
logist, 98, 54–67. North Carolina at Chapel Hill.
14-HOUSE.qxd 3/19/2004 4:47 PM Page 392
Hoppe, M. H. (1993). The effects of national culture designs. Organizational Research Methods, 2,
on the theory and practices of managing 395–418.
research-and-development professionals abroad. Linke, U. (1992). Manhood, femaleness, and power:
R & D Management, 23, 313–325. A cultural analysis of prehistoric images of
Hoppe, M. H. (1998). Validating the masculinity/ reproduction. Comparative Studies in Society
femininity dimension on elites from 19 coun- and History, 34, 579–620.
tries. In G. Hofstede (Ed.), Masculinity and fem- Loftus, E. F. (1991). Resolving legal questions with
ininity: The taboo dimension of national psychological data. American Psychologist, 46,
cultures (pp. 29–43). Thousand Oaks, CA: Sage. 1046–1048.
House, R. J., Hanges, P. J., Ruiz-Quintanilla, S. A., Lovenduski, J., & Hills, J. (1981). The politics of the
Dorfman, P. W., Javidan, M., Dickson, M. W., second electorate: Women and public partici-
et al. (1999). Cultural influences on leadership pation. London: Routledge & Kegan Paul.
and organizations: Project GLOBE. In W. H. Lovenduski, J., & Norris, P. (1989). Selecting women
Mobley, M. J. Gessner, & V. Arnold (Eds.), candidates: Obstacles to the feminization of the
Advances in global leadership (pp. 171–233). House of Commons. European Journal of
Greenwich, CT: JAI. Political Research, 17, 533–562.
Inkeles, A., & Levinson, D. J. (1969). National Martin, C. L. (1987). A ratio measure of sex stereo-
character: The study of modal personality and typing. Journal of Personality and Social
sociocultural systems. In G. Lindzey & Psychology, 52, 489–499.
E. Aronson (Eds.), Handbook of social psychol- McClelland, D. C. (1961). The achieving society.
ogy (Vol. 4, pp. 418–506). New York: McGraw- Princeton, NJ: Van Nostrand Rheinhold.
Hill. (Original work published 1954) McDonald, G. W. (1980). Family power: The
Inkeles, A., & Levinson, D. J. (1997). National assessment of a decade of theory and research.
character: The study of modal personality and Journal of Marriage and the Family, 42,
sociocultural systems. In A. Inkeles, National 841–854.
character: A psycho-social perspective. New McElvaine, R. S. (2000). Eve’s seed: Biology, the
Brunswick, NJ: Transaction. (Original work sexes and the course of history. New York:
published 1954) McGraw-Hill.
Jacobs, J. A., & Lim, S. T. (1992). Trends in occupa- Mead, M. (1949). Male and female. New York:
tional and industrial sex segregation in 56 coun- William Morrow.
tries, 1960–1980. Work and Occupations, 19, Menard, S. (1990). Cross-national models of
450–486. fertility, family planning, and development:
Jordan, J. O., & Patten, R. L. (1995). Nineteenth- Testing for reciprocal effects. Studies in
century British publishing and reading practices. Comparative International Development, 25,
Cambridge, UK: Cambridge University Press. 60–90.
Kattler, D., & Williamson, J. B. (1988). Welfare state Merritt, A. (2000). Culture in the cockpit: Do
development and life expectancy among the Hofstede’s dimensions replicate? Journal of
aged: A cross-national analysis. Journal of Cross-Cultural Psychology, 31, 283–301.
Aging Studies, 2, 13–24. Miller, E. M. (1994). Paternal provisioning versus
Katz, M. M., & Konnor, M. J. (1981). The role of the mate seeking in human populations. Personality
father: An anthropological perspective. In M. E. and Individual Differences, 17, 227–255.
Lamb (Ed.), The role of the father in child devel- Moore, G., & Shackman, G. (1996). Gender and
opment (2nd ed., pp. 155–185). New York: John authority: A cross-national study. Social Science
Wiley. Quarterly, 77, 274–288.
Kukreja, S. (1992). The political economy of the fem- Nunnally, J. C., & Bernsrein, I. H. (1994).
inization of the tertiary in developing countries: Psychometric theory (3rd ed.). New York:
Some cross-national findings. International McGraw-Hill.
Review of Modern Sociology, 22, 79–97. Nuss, S., & Majka, L. (1983). The economic integra-
Lance, C. E., & James, L. R. (1999). A propor- tion of women: A cross-national investigation.
tional variance accounted for index for some Work and Occupations: An International
cross-level and person-situation research Sociological Journal, 10, 29–48.
14-HOUSE.qxd 3/19/2004 4:47 PM Page 393
Pampel, F. C., & Park, S. (1986). Cross-national Schwartz, S. H. (1994). Beyond individualism
patterns and determinants of female retirement. and collectivism: New cultural dimensions of
American Journal of Sociology, 91, 932–955. values. In U. Kim, H. C. Triandis, C. Kagitçibasi,
Paris, L. (2003). The effect of gender and culture on S. C. Choi, & G. Yoon (Eds.), Individualism and
implicit leadership theories: A cross-cultural collectivism: Theory, method, and applications
study. Unpublished doctoral dissertation. New (pp. 85–199). Thousand Oaks, CA: Sage.
Mexico State University, Las Cruces, NM. Segall, M. H., Dasen, P. R., Berry, J. W., &
Parsons, T. (1964). Social structure and personality. Poortinga, Y. H. (1990). Human behavior in
New York: Free Press. global perspective: An introduction to cross-
Pattnayak, S. R., & Shai, D. (1995). Mortality rates cultural psychology. New York: Pergamon.
as indicators of cross-cultural development: Shen, C., & Williamson, J. B. (1999). Maternal
Regional variations in the Third World. Journal mortality, women’s status, and economic depen-
of Developing Societies, 11, 252–262. dency in less developed countries: A cross-
Paxton, P. (1997). Women in national legislatures: A national analysis. Social Science & Medicine,
cross-national analysis. Social Science 49, 197–214.
Research, 26, 442–464. Sigelman, L., & Tsai, Y. (1985). Urbanism and
Peterson, M. F., & Smith, P. B. (1997). Does women’s labor force status: A cross-national
national culture or ambient temperature explain study. International Journal of Comparative
cross-national differences in role stress? No Sociology, 26, 109–118.
sweat! Academy of Management Journal, 40, Slater, P., & Slater, D. (1965). Maternal ambivalence
930–946. and narcissism. Merrill-Palmer Quarterly, 11,
Piccinelli, M., & Simon, G. (1997). Gender and 241–259.
cross-cultural differences in somatic symptoms Smelser, N. (1970). Mechanisms of change and
associated with emotional distress: An interna- adjustment. In J. Finkle & R. Gable (Eds.),
tional study in primary care. Psychological Political development and social change. New
Medicine, 27, 433–444. York: John Wiley.
Price Waterhouse v. Hopkins, 490 U.S. 288 (1989). South, S. J., & Spitze, G. (1994). Housework in
Reskin, B., & Padavic, I. (1994). Women and men at marital and nonmarital households. American
work. Thousand Oaks, CA: Pine Forge Press. Sociological Review, 59, 327–347.
Roschrhomberg, I. (1994). Hierarchical opposition Stover, R. G., & Hope, C. A. (1984). Monotheism
and the concept of Um-Yang (Yin-Yang): A and gender status: A cross-societal study. Social
reevaluation of values in the light of the Forces, 63, 335–348.
symbolism of Korean rituals for the dead. Stroh, L. K., Brett, J. M., & Reilly, A. H. (1992). All
Anthropos, 89, 471–491. the right stuff: A comparison of female and
Rule, W. (1987). Electoral systems, contextual male managers’ career progression. Journal of
factors and women’s opportunity for election Applied Psychology, 77, 251–260.
to parliament in twenty-three democracies. The Szalai, A. (1972). The use of time: Daily activities
Western Political Quarterly, 40, 477–498. of urban and suburban populations in twelve
Sacks, K. (1979). Sisters and wives: The past and countries. The Hague: Mouton.
future of sexual inequality. Westport, CT: Triandis, H. C. (1994). Cross-cultural industrial and
Greenwood. organizational psychology. In H. C. Triandis,
Safilios-Rothschild, C. (1970). The study of family M. D. Dunnette, & L. M. Hough (Eds.),
power structure: A review, 1960–1969. Journal Handbook of industrial and organizational
of Marriage and the Family, 32, 345–352. psychology (2nd ed., Vol. 4, pp. 103–172). Palo
Schein, E. H. (1983). The role of the founder in Alto, CA: Consulting Psychologists Press.
creating organizational culture. Organizational Trivers, R. L. (1972). Parental investment and sexual
Dynamics, 12, 13–28. selection. In B. Campbell (Ed.), Sexual selection
Schneider, S. C. (1988, Summer). National vs. corpo- and the descent of man, 1871–1971. Chicago:
rate culture: Implications for human resource Aldine.
management. Human Resource Management, Van de Vliert, E. (1998). Gender role gaps, competi-
p. 239. tiveness, and temperature. In G. Hofstede
14-HOUSE.qxd 3/19/2004 4:47 PM Page 394
(Ed.), Masculinity and femininity: The taboo Welch, M. R. (1982). Female exclusion from religious
dimensions of national cultures (pp. 117–129). roles: A cross-cultural test of competing expla-
Thousand Oaks, CA: Sage. nations. Social Forces, 61, 79–98.
Van de Vliert, E., Schwartz, S. H., Huismans, S. E., Welch, M. R. (1983). Women and political leadership
Hofstede, G., & Daan, S. (1999). Temperature, roles: A cross-cultural study. International
cultural masculinity and domestic political Journal of Sociology of the Family, 13, 57–77.
violence: A cross-national study. Journal of Wells, S. (2001). Out of the dead house: Nineteenth-
Cross-Cultural Psychology, 30, 291–314. century women physicians and writing of medi-
Van de Vliert, E., & Van Yperen, N. W. (1996). Why cine. Madison: University of Wisconsin Press.
cross-national differences in role overload? Whyte, M. K. (1978). The status of women in prein-
Don’t overlook ambient temperature! Academy dustrial societies. Princeton, NJ: Princeton
of Management Journal, 39, 986–1004. University Press.
Verweij, J., Ester, P., & Nauta, R. (1997). Williams, J. E., & Best, D. L. (1982). Measuring sex
Secularization as an economic and cultural stereotypes: A thirty nation study. Beverly Hills,
phenomenon: A cross-national analysis. Journal CA: Sage.
for the Scientific Study of Religion, 36, 309–324. Williams, J. E., & Best, D. L. (1990a). Measuring sex
Walker, H. A., & Fennell, M. L. (1986). Gender stereotypes: A multination study. Newbury Park,
differences in role differentiation and organiza- CA: Sage.
tional task performance. Annual Review of Williams, J. E., & Best, D. L. (1990b). Sex and psy-
Sociology, 12, 255–275. che: Gender and self viewed cross-culturally.
Walton, W. (2000). Eve’s proud descendants: Four Newbury Park, CA: Sage.
women writers and republican politics in Wright, E. O., Baxter, J., & Birkelund, G. E. (1995).
nineteenth-century France. Stanford, CA: The gender gap in workplace authority: A cross-
Stanford University Press. national study. American Sociological Review,
Warner, R. L., Lee, G. R., & Lee, J. (1986). Social 60, 407–435.
organization, spousal resources, and marital
power: A cross-cultural study. Journal of
Marriage and the Family, 48, 121–128.
15-House.qxd 3/19/2004 4:51 PM Page 395
15
ASSERTIVENESS
DEANNE N. DEN HARTOG
“Blessed are the meek: for they shall inherit the earth” (Matt. 5:5).
his quote from the Bible reflects belief cultural set of GLOBE-defined practices and
395
15-House.qxd 3/19/2004 4:51 PM Page 396
Assertiveness– • –397
personality (e.g., Rathus, 1973). Without trying among a large segment of the United States
to be exhaustive, we will first discuss some of the population in the early to mid-70s. Books, arti-
literature on assertiveness as a style of respond- cles, and workshops aimed at the mass market
ing, and assertiveness as a personality trait. As claimed that a lack of assertiveness causes
assertiveness is often described as a primarily problems for people and offered techniques for
masculine characteristic, we will also very briefly becoming more assertive. As stated, assertiveness
review some findings from research on gender was seen as a style of (communicative) behavior,
roles and stereotyping in the workplace. Next, amenable to training. As Rakos (1991) noted, the
Assertiveness is examined as an attribute or dimen- self-improvement literature often simplified the
sion of national culture. After that the GLOBE concept of assertion to “saying no” and “getting
measures and findings concerning Assertiveness your own way” (Smith, 1975) or to “standing
as a dimension of culture are presented. up for your rights” and “get(ting) where and
what you want” (Baer, 1976). Many of the mass-
market books were aimed specifically at women
A PSYCHOLOGICAL PERSPECTIVE buyers. Virtually all such popular books claimed
women were especially in need of assertiveness
training and “blamed” gender socialization for
Assertiveness as a Style of Responding:
this. According to Crawford (1995), authors drew
The Assertiveness-Training Hype
heavily on stereotypes of female passivity. Only
As stated, in much of the psychological and rarely did they provide research evidence, other
behavioral literature, assertiveness became than paper-and-pencil inventories completed by
defined as a set of social and communicative college students, that women as a group have
skills amenable to training. The label “assertive” trouble asserting themselves.
for a specific category of interpersonal behaviors
and the focus of attention on the use of assertive- Assumptions Underlying
ness training to alleviate interpersonal problems
Assertiveness Training
originates in the work of Wolpe (1958). This
approach was based on Salter’s (1949) Condi- As Crawford (1995, p. 56) noted, underlying
tioned Reflex Therapy, the first approach to the behavioral literature on assertiveness was a
apply behaviorism to clinical problems involving set of values and assumptions about “healthy”
communication. Salter applied a conditioning and “adaptive” interpersonal behavior that
model to verbal and nonverbal behaviors, such remained implicit and unexamined. An individ-
as the ability to say what one feels, to contradict ualistic ethic made assertion seem like every-
and disagree, and to speak directly through the one’s right. In this body of literature on
frequent use of “I statements” (Crawford, 1995). assertiveness, individual perceptions, feelings,
An example of how assertiveness is defined in and beliefs are valued above the maintenance
the psychological literature is the following: and process of social relationships, and the ideal
“Assertion basically involves asking for what relationship is one of parallel self-fulfillment
one wants, refusing what one doesn’t want, and (Crawford, 1995). According to Shoemaker and
expressing positive and negative messages to Satterfield (1977) the implied philosophy of
others” (Booream & Flowers, 1978, p. 15). social relationships in assertiveness is one of
Assertive behavior is explicitly contrasted benign self-interest. The individualism of the
to passive behavior (Crawford, 1995). Such assertive philosophy is also evident in the lan-
passive behavior is attributed to those who guage of assertiveness techniques. Clients
fail to express their true thoughts and feelings, are taught to use “I–me” messages: “From this
allow themselves to be dominated or humiliated perspective, learning specific conversational
by others, and who comply with requests or techniques in order that one’s own opinions
demands of others even if they themselves do and decisions should prevail is a route to self-
not want to (Lange & Jakubowski, 1976). fulfillment” (Crawford, 1995, p. 57). The popu-
Assertiveness and assertiveness training lar assertiveness literature encourages people to
as a behavior therapy technique became popular dispense with reasons or justifications for their
15-House.qxd 3/19/2004 4:51 PM Page 398
feelings or behaviors, implying that individuals that can help alleviate specific problems, and
should be the sole judges of their own actions fewer studies are devoted to it. Rakos (1991)
(Crawford, 1995). holds the decline in interest in assertiveness train-
Besides individualism, the notion of assertive ing is probably due to the complex nature of
behavior also reflects rationality and pragma- assertiveness. For example, consideration of the
tism (Rakos, 1991). Rationality plays a role in social and cultural context is an essential founda-
determining the appropriate expression of tion of a functionally useful assertive response. In
emotion, a key element in assertive behavior. other words, the functionality of assertive behav-
For instance, Wolpe (1982) defines assertive ior to obtain desired outcomes, as well as whether
behavior as “the appropriate expression of any certain actions will be seen as assertive or
emotion other than anxiety toward another nonassertive, depends on the situation.
person” (p. 118). The ability to produce such an An example of such a situational constraint
appropriate response requires the accurate dis- has to do with gender of both “asserters” and
crimination of situational cues, decision-making observers. Research shows that the social con-
skills, and the emission of acceptable social sequences of assertive behavior differ for men
behavior, all of which are produced by rational and women. For instance, Kelly, Kern, Kirkley,
rather than emotional processes. Rationality is Patterson, and Keane (1980) showed that
involved in our deductive inference, informa- assertive women were seen as more competent
tion processing and decision making, the way in but less likable than nonassertive women, a dis-
which we judge and comprehend behavior of tinction that was not made for assertive men.
others, in explanations for events that happen to Following up on this, Crawford (1988) designed
us, and in the further understanding of our envi- a study testing the evaluation of assertive
ronment (Irani, 1986; Rakos, 1991). behavior. She developed scenarios in which a
Pragmatism—the notion that what works woman or man behaved assertively. Raters then
is much more important than dogmatism or rated the models. The results showed that raters
ideology—also underlies assertiveness training. judged the models on social competence and
Assertiveness training emphasizes that assertive likeability. Assertiveness was evaluated differ-
behavior is only one option for coping with diffi- ently depending on the sex of the assertive
cult or problematic circumstances, and in many model and the age of the research participant.
instances it may not be the preferred one. In Assertive women models received the lowest
other words, there is no ideological mandate likeability ratings of all from older male partic-
always to respond assertively. “Such pragmatism ipants. Such gender differences are the focus
is, of course, highly adaptive in our modern of Chapter 14 of this volume, which describes
world, with its shifting value systems, complex GLOBE’s measures and findings on Gender
and ambiguous situations, and increasing cultural Egalitarianism.
and social heterogeneity” (Rakos, 1991, p. 5). Obviously, culture is another important
Over the years, literally hundreds of studies situational influence that determines whether
were done on assertiveness and assertiveness responding assertively is valued and which
training. Some of these studies describe the types of responses are appropriate. “At least
definition and variety of assertive speech, for some of the differences in assertiveness among
instance, asking what would constitute an appro- cultural groups are fundamentally due to cogni-
priate assertive response. Others describe pro- tive variables stemming from cultural values
grams for assertiveness training, evaluate their and norms” (Rakos, 1991, p. 14). However, the
outcomes, or discuss which types of people can issue as to whether one culture is more or less
benefit most from such training. Yet other studies assertive than another may be complex. For
report the development, reliability, and validity instance, Sue, Ino, and Sue (1983) found that
of paper-and-pencil tests for measuring levels of whereas the self-reports of general assertiveness
assertiveness in prospective clients (Crawford, behavior for Chinese Americans were less than
1995). The true hype of assertiveness training assertive behavior reported by European
seems to have passed. It is now used in a far more Americans, actual assertiveness behaviors did
limited manner as a behavioral therapy technique not differ.
15-House.qxd 3/19/2004 4:51 PM Page 399
Assertiveness– • –399
As stated, the literature on assertiveness (McCrae & Costa, 1997). In this Big Five model,
training is strongly U.S.-dominated, and the assertiveness can be seen as part of extraversion.
concept seems linked to culture. As was Extroverts tend to be sociable and gregarious, but
described in earlier chapters, other cultures do also surgent, dominant, and ambitious as well as
not share the same assumptions. For instance, assertive, active, and adventurous.
Hofstede (1993, p. 81) noted that United States According to Watson and Clark (1997),
management theories contain several idiosyn- extroverts are more likely to take on leadership
crasies not necessarily shared by management roles. Extraversion has also been found to be
elsewhere; for example, “a stress on market associated with job performance and career
processes, a stress on the individual, and a focus success. For instance, in the United States,
on managers rather than workers.” The same dominance and sociability differentiated between
seems to hold in the area of assertiveness. As successful and unsuccessful executives, with
Furnham (1979) notes, dominant executives being more successful
(Rawls & Rawls, 1968). Also, extraversion pre-
the concept of assertiveness is culture bound, and dicted salary and job level in a United Kingdom
particularly North American. In many other cul- study (Melamed, 1996a, 1996b). In their meta-
tures, asserting oneself in the way that is normative analytical study, Barrick and Mount (1991)
in North America and parts of Europe is neither found that extraversion was a valid predictor of
encouraged nor tolerated. Humility, subservience, job performance for two occupational groups:
and tolerance are valued above assertiveness in managers and sales people. In these groups,
many other cultures, especially for women. (p. 522) traits such as being assertive, active, and socia-
ble were significantly (but only moderately)
For instance, an employee voicing disagree- associated with more effective performance.
ment to his or her boss is an accepted and posi- Several longitudinal studies also confirm this
tively evaluated manner of expressing oneself link between success and extraversion. For
in some cultures, but may be unacceptable instance, assessment center ratings of expres-
in others. Whether and how such assertive sive social skills predicted later managerial
responding is done and how it is valued depends promotions (Howard & Bray, 1994), and child-
on societal norms regarding such behavior. hood ratings of shyness were negatively associ-
Such norms describe the degree to which people ated with adult occupational status (Caspi,
in organizations or societies are (or should be) Elder, & Bem, 1988). Also, positive relation-
assertive in social relationships. ships of high extraversion with extrinsic career
success (a variable describing a combination of
income, promotions, and occupational status)
Assertiveness as a Personality Trait
were found (Judge, Higgins, Thoresen, &
In addition to the body of psychological Barrick, 1999). Interestingly in this study, both
assertiveness literature that focuses on asser- measures taken in childhood and adulthood
tiveness as a style of responding, there is also a predicted later career success.
body of literature that has examined assertive- Nonassertiveness (in terms of being tender
ness as a more stable personality trait. According and nondominant) can be linked to the Big Five
to this perspective, some people are inherently factor of agreeableness. Agreeable people are
more assertive than others. cooperative (trusting, caring) and likable (good
In the field of personality research, evidence is natured, gentle); they are also softhearted and
accumulating that most personality measures can tolerant. The meta-analysis by Barrick and
be reduced to a five-factor model, which is now Mount (1991) suggests agreeableness is not
labeled the “Big Five” (Goldberg, 1990). The an important predictor of job performance.
dimensions composing this model are neuroti- Moreover, in the previously mentioned longitu-
cism, extraversion or surgency, openness to dinal study by Judge and colleagues (1999)
experience, agreeableness, and conscientious- agreeableness negatively predicted extrinsic
ness. Interestingly, the five-factor structure has career success. Similarly, affability (the degree to
been found to generalize across many cultures which an individual is nurturing, nonaggressive,
15-House.qxd 3/19/2004 4:51 PM Page 400
supportive, and sympathetic) was negatively socially developed countries. In such countries,
related to management potential (Howard & higher proportions of women attend university
Bray, 1988). and are gainfully employed, the ideology
To summarize, literature shows that in regarding the status of women is more egalitar-
terms of personality, high rather than low ian, and women and men perceive themselves
levels of assertiveness are linked to effective- to be more similar. Interestingly, in most of the
ness and success. However, such studies are 25 countries Williams and Best studied, women
predominantly based on United States samples were more liberal than the men. Such differences
and, to a lesser extent, samples from the United have also been found in other studies (e.g.,
Kingdom. Whether these results are similar in Kalin, Heusser, & Edmonds, 1982; Scher, Nevo,
other cultures remains to be seen. Assertiveness & Beit-Hallahmi, 1979).
is also often seen as a “masculine” rather than Besides gender, assertiveness is also associ-
“feminine” trait. The next section briefly reviews ated with the stereotype of successful man-
some of the research on assertiveness and agers. Successful managers are seen as assertive,
gender. More detailed information on real and dominant, and frank. Different studies also
perceived gender differences in society is pre- found a strong concurrence between the ratings
sented in Chapter 14 by Emrich, Denmark, and of men and successful managers and a weak
Den Hartog. one between ratings of women and successful
managers. Successful managers were viewed as
more similar to men than to women on attrib-
Assertiveness and Stereotypes
utes considered critical to effective work
Hofstede (2001) notes that assertiveness performance, such as ambition, assertiveness,
seems to be a trait or behavior that people asso- leadership ability, self-confidence, and forceful-
ciate more with men than with women (see ness (e.g., Brenner, Tomkiewicz, & Schein,
chapter 14, this volume). Segall, Dasen, Berry, 1989; Heilman, Block, Martell, & Simon, 1989;
and Poortinga (1990) state that Schein, 1973). A recent international replication
by Schein (2001) found that this pattern still
risking oversimplification, we can summarize the held to a large extent, especially among male
picture of sex-differences in behavior that is pre- respondents in the five countries included in the
sented by anthropology and cross-cultural psychol- study (China, Japan, Great Britain, Germany,
ogy as showing males to be more self-assertive, and the United States). Fagenson (1990)
achieving, and dominant and females to be more assessed the effects of both gender and the posi-
socially responsive, passive, and submissive. (p. 250) tion one holds in the organizational hierarchy
and found that there was no relation between
Such characteristics refer to sex-trait stereo- individuals’ sex and possession of “masculine”
types and do not imply actual differences characteristics, such as assertiveness. Instead,
between the sexes. The real differences in behav- masculinity was found to be related to the
ior between genders seem to be less pronounced individuals’ perceived power and their position
than the beliefs about those behavioral differ- in the organizational hierarchy. Individuals in
ences (e.g., Segall et al., 1990). For instance, as upper levels were reported to be more masculine
seen in Chapter 14, this volume, Williams and than those in the lower levels, and men and women
Best (1982) found some general agreement in the organization’s upper tier saw themselves
across countries in differentially attributing traits similarly with respect to masculine attributes.
to men or women. Traits related to assertiveness Therefore, Fagenson holds labeling attributes such
such as dominance, aggression, autonomy, as assertiveness as masculine is a misnomer that
achievement, and endurance were ascribed to may cause confusion. More accurate labels
men and traits related to nonassertiveness such might be “powerful” or “successful” attributes.
as deference, nurturance, and abasement to Taken together, the findings from these and
women. Williams and Best (1989) also found many other studies in this area suggest that—
that the self-perceptions of men and women at least in the United States—assertiveness is
were less stereotypical in more economically and an important attribute associated with being
15-House.qxd 3/19/2004 4:51 PM Page 401
Assertiveness– • –401
successful in a managerial position. Also, it is were asked to indicate where someone from a
often associated more with men than women, certain nationality was likely to be located on
although the differences are in part stereotypical a pair of traits (e.g., an Englishman is likely
rather than true. The actual differences seem less to be [more or less] “passive” vs. [more or
pronounced. Men may indeed exhibit (slightly) less] “forceful”). Different groups of judges
more assertive behavior than women, although compared self-judgments (judging a target
this is likely be confounded with the fact that having one’s own nationality) and judgments
men more often hold positions of power in the by others (judging a target that has a different
organizational (and societal) hierarchy. nationality than one’s own).
Peabody found three factors: One he labeled
self-assertiveness versus unassertiveness, and
Assertiveness as a Dimension of Culture
the other two he labeled a general evaluative
The concept of assertiveness, as stated, dimension and a dimension describing tight ver-
originates (in part) from Hofstede’s culture sus loose control over impulse expression. Items
dimension of masculinity versus femininity. In in his final scale for assertiveness were aggres-
masculine societies men are supposed to be sive versus peaceful, passive versus forceful,
assertive and tough and women are expected conceited versus modest, self-confident versus
to be modest and tender. In contrast, femininity unassured, bold versus timid, and active versus
pertains to societies in which social gender roles inactive. The national characteristics were per-
overlap (Hofstede, 1980, 1998, 2001). As stated, ceived to be similar by both residents of a spe-
this dimension conceptually appears to encom- cific country and people with other nationalities.
pass several distinct aspects (including gender Thus, there were differences between nation-
differences, success striving, and assertiveness). ality targets on the assertive versus unassertive
See Chapter 14 by Emrich, Denmark, and Den dimension. For instance, judges in his study
Hartog for a review of the work on masculinity. agree that the English are relatively nonassertive,
According to Hofstede (2001), high MAS whereas the Germans and Americans are seen
means high assertiveness or ambitiousness as highly assertive. Thus, the stereotypical view
(p. 164). Although assertiveness is seen as part of Germans and Americans is that they are
of the masculinity dimension, the MAS index assertive. As stated, GLOBE sees Assertiveness
used to measure it does not include items that not only as a behavior, trait, or even a stereotyp-
specifically target assertiveness and aggres- ical national characteristic, but also as a relevant
siveness in relationships. Assertiveness as a dimension of national cultures that reflects
culture dimension originates partly in Hofstede’s shared societal beliefs about whether people are
masculinity dimension, but also differs from it or should be assertive and tough-minded, or
substantially. Below, we describe the Asserti- unassertive and tender in their social relation-
veness construct as used in the GLOBE study ships. To our knowledge, assertiveness has not
and review some of the other work on national been studied as a separate cultural dimension in
cultures that is reflected in the construct. this manner.
To our knowledge, there are no studies As noted earlier, culture can be seen as a set
focusing on assertiveness as a cultural dimen- of relatively stable, basic, and shared practices
sion. One study undertaken in several European and values that help human social groups
nations does investigate assertiveness as a so- or societies find solutions to two fundamental
called “national characteristic” that may differ problems. The first is how to survive, grow, and
across cultures (Peabody, 1985). Peabody was adapt to the environment (external adaptation).
interested in the question of whether differ- The second is the problem of internal integration
ent nationalities have different (stereotypical) that permits daily functioning and ensures the
perceived psychological characteristics. He capacity or ability to adapt and survive (Schein,
used pairs of trait-adjectives to describe such 1992). Assertiveness is an important dimension
“national characteristics,” with examples such of a community’s culture that relates both to the
as conceited versus modest, self-controlled ver- issues of external adaptation and especially to
sus impulsive, and severe versus lenient. Judges internal integration. It is an internally consistent
15-House.qxd 3/19/2004 4:51 PM Page 402
set of practices and values regarding the way in earth”; or they believe that man is part of nature
which people are seen to and ought to behave in and must go along with its laws, directions and
social relationships in a community. forces. (p. 141)
Assertiveness– • –403
the value of competition and competitiveness. to cooperative, peaceful relationships. Anger and
“Eat or be eaten,” “control your destiny,” “try to aggression are negatively valued, whereas, gen-
be a winner,” and “win-some, lose-some” are erally, nurturance and cooperation are positively
examples of statements associated with such a valued. Many such societies also de-emphasize
competitive mind-set. Most people from the individual achievement as this, to them, is closely
United States fervently believe in competition linked to competitiveness and hence aggressive-
according to Kohn (1986). It is believed to be ness. Bonta also noted that although these soci-
needed for economic prosperity and essential eties are cooperative and egalitarian, they are not
for achievement in sports, science, arts, and necessarily collectivistic. Bonta found substan-
other areas. In the U.S., competition is seen as tial variation in degrees of collective sharing of
“a fundamental aspect of human nature; people resources. He also noted that the cooperative
live in a dog-eat-dog world; people need to behavior and belief in peacefulness is not merely
compete to survive and prosper” (Bonta, 1997, a reflection of economic organization. Rather,
p. 121). Similarly, as noted in Chapter 12 by the peacefulness they achieve is due to their very
Javidan, Japanese executives have been criti- strong beliefs in their need to be peaceful, and
cized in the past for their extreme obsession the psychological strategies they use to reinforce
with destroying their competitors, even at the and strengthen those shared beliefs and attitudes
expense of their own companies. in daily practices (Bonta, 1997). For example, it
Shared views in society on how to deal with would be acceptable for a baseball game in
competition and cooperation are manifestations Japan to end in a tie; however, in the United
of human cultures. Bonta (1997) reviewed the States baseball games are extended until there
literature on nonviolent or peaceful societies are clear winners and losers regardless of the
and showed that, for most of these societies, the length of the game.
defining elements in their beliefs are a strong Assertiveness is also linked to the preferred
opposition to competition and strong support for use of language in society. As described above,
cooperation. For example, Bonta (1997) des- assertiveness can be seen as a style of respond-
cribes the differences in rituals in competition- ing that implies making one’s wants known
versus cooperation-oriented societies. In Western to others and in no uncertain terms. In many
societies, competitive activities such as ritual- Western cultures, this is expected or at least
ized sports and elections for leadership posi- asserted to be accomplished most efficiently
tions “promote social cohesion and civic pride by being direct and unambiguous. Indeed,
in the virtues of competition, ‘manliness’ and Holtgraves (1997) found a negative relationship
aggressiveness” (p. 313). In contrast, the rituals between assertiveness (as an individual-level
of peaceful societies (e.g., chanting and healing trait) and indirect language use in the United
ceremonies) “help integrate the society, protect States. Also, such conversational indirectness
them from hostile outside forces, and focus was found to correlate negatively with social
people on worldviews of opposition to com- desirability. Thus, in the United States, saying
petition, a need for harmony, and the virtue of what one means in a direct manner, even in
non-violence” (p. 313). terms of “brutal honesty” and “tough love,” is
The peaceful, or nonviolent, societies Bonta valued (Holtgraves, 1997). This is not the case
describes include the Amish and Balinese, as in many non-Western cultures. Many cultures
well as several aboriginal and native tribes from value a less-direct manner of responding. Thus,
different continents. In such peaceful societies, in assertive societies people will tend to use
competition is clearly linked to aggression and what is also referred to as low-context language,
violence. Children are not taught competitive which is speech that is direct, clear, and explicit
games. They learn by example, observing only (Hall, 1959; Schneider & Barsoux, 1997). In
cooperative and peaceful behavior and no aggres- contrast, less-assertive cultures tend to use high-
sive or competitive behavior. Bonta holds that in context language, which is less direct, more
many of these societies, people feel uncertain ambiguous, and more subtle (Schneider &
about the intentions of others, which forces them Barsoux, 1997). In such cultures, directions and
to constantly reaffirm their mutual commitments messages are implied rather than explicitly
15-House.qxd 3/19/2004 4:51 PM Page 404
expressed. It is expected that the person receiving Whereas some societies accord status to people
the implicit message will infer the meaning of on the basis of their achievements, others ascribe
the message or “read between the lines.” it to people on the basis of age, gender, social
High-context language or indirectness in class, profession, or other criteria (Trompenaars
communication can be linked to face manage- & Hampden-Turner, 1997). Achieved status
ment (e.g., Brown & Levinson, 1987; Holtgraves, refers to what one has done or personally
1992). As Holtgraves explained, people accomplished. Ascribed status refers to earning
status based on socioeconomic status, family
are motivated to collectively manage the face (or lineage, age, profession, or family connections.
public identity) of each other and they do this by Given the emphasis on competition, highly
phrasing their remarks politely. A primary linguis- assertive societies should theoretically be
tic mechanism for politeness and face management expected to accord status based on achieve-
is indirectness. Thus, the face threat of a speech ments rather than ascribe it based on attributes
act is lessened when it is performed indirectly. such as age, profession, or family connections.
(Holtgraves, 1997, p. 633) Doney, Cannon, and Mullen (1998) link
societal norms and values to trust-building
Although face management in some form processes within society. In more assertive soci-
or another is probably important in many eties, trust-building processes will mostly
cultures, people from collectivistic cultures are be based on either calculations of the intentions
generally more concerned with face manage- of others or on estimates of their capabilities to
ment than people from individualistic coun- fulfill their commitments. The perception that
tries (Ting-Toomey, 1988). Some empirical others are opportunistic and seek to maximize
support for the idea of cultural differences in their self-interests stimulates a calculative trust-
this area exists. In Holtgraves’s (1997) research, building process. Trust can be based on the cal-
for instance, Koreans were found to be more culated risks and costs involved in trusting or not
indirect than U.S. Americans. trusting others. Cultural assertiveness influences
Besides a preference for directness versus the likelihood that people will act opportunisti-
indirectness in communication, societal norms cally, as well as the costs associated with such
can also influence the amount of emotion one behavior. Doney and colleagues note that
typically shows in public within a certain soci- evidence from anthropology, psychology, and
ety. Trompenaars and Hampden-Turner (1997) political science confirms a cultural pattern of
contrast neutral and affective cultures. In affec- assertiveness and aggressiveness that is consis-
tive cultures, showing one’s emotions—in tent with a tendency toward opportunism. The
laughter, gesture, as well as heated debate—is so-called “tough” values in such assertive and
the norm. For instance, in Latin American and masculine societies (e.g., visible achievement
Southern European countries (such as Spain and and making money) suggest that in such soci-
Italy) people openly show emotions. In more eties the potential rewards for opportunistic
neutral cultures, people tend to keep their behavior may well exceed their costs.
emotions in check. In such cultures, keeping a The second trust-building process likely in
subdued manner, maintaining self-possessed societies stressing assertiveness involves others’
conduct, and not openly showing emotion is the capabilities. Individuals differ in their compe-
norm. For instance, in Scandinavian and many tence, ability, or expertise. Trustees’ capabili-
Asian countries, people tend to show less emo- ties will, in the eyes of a trustor, influence their
tion in public. In highly assertive and direct ability to keep their promises, and trust is based
societies, communication is also likely to be more on a trustee’s perceived capability to keep such
emotionally expressive. promises. Societal norms and values supporting
As described by Carl and Gupta in Chapter assertive behavior, individual initiative, and
17 on Power Distance, within all societies there wealth creation should influence the importance
are status and power differentials. Assertiveness placed on individual capabilities. For instance, in
is one of the aspects of culture that relates to highly assertive cultures, “super-achievers” are
how status is typically accorded in society. likely to be respected, and a norm for excelling
15-House.qxd 3/19/2004 4:51 PM Page 405
Assertiveness– • –405
• Value assertive, dominant, and tough behavior • View assertiveness as socially unacceptable
for everyone in society and value modesty and tenderness
• Have sympathy for the strong • Have sympathy for the weak
• Value competition • Value cooperation
• Believe that anyone can succeed if he or she • Associate competition with defeat and
tries hard enough punishment
• Value success and progress • Value people and warm relationships
• Value direct and unambiguous communication • Speak indirectly and emphasize “face-saving”
• Value being explicit and to the point in • Value ambiguity and subtlety in language and
communications communications
• Value expressiveness and revealing thoughts • Value detached and self-possessed conduct
and feelings • Have far more negative connotations with the
• Have relatively positive connotations for the term aggression (e.g., aggression leads only
term aggression (e.g., aggression helps to win) to negative outcomes)
• Have a just-world belief • Have an unjust-world belief
• Try to have control over the environment • Value harmony with the environment rather
• Stress equity, competition, and performance than control
• Have a “can-do” attitude • Stress equality, solidarity, and quality of life
• Emphasize results over relationships • Emphasize tradition, seniority, and experience
• Value taking initiative • Emphasize integrity, loyalty, and cooperative
spirit
• Reward performance
• View “merit pay” as potentially destructive to
• Expect demanding and challenging targets
harmony
• Believe that individuals are in control
• Value who you are more than what you do
• Value what you do more than who you are
• Build trust on the basis of predictability
• Build trust on the basis of capabilities or
• Think of others as inherently worthy of trust
calculation
• Act and think of others as opportunistic
is supported. This also holds in cultures that What Does It Mean to Be a High
Hofstede (1980) labeled “masculine.” Individual or Low Assertiveness-Oriented Society?
brilliance is admired and the “stars” are idolized
(Doney et al., 1998; Kale, 1991; Kale & Barnes, The above descriptions present a picture of
1992). In such circumstances, one’s capability what it means to be assertive. Table 15.1 pro-
seems a reasonable base on which to form trust. vides a summary comparison of a typical society
In contrast, in cultures emphasizing norms for that has a strongly assertive culture and a society
solidarity, service, and cooperation and that have that is low on assertiveness. However, one must
a high social pressure to honor moral obligations, keep in mind that societal culture is obviously
other forms of trust-building should prevail. For far too complex to be presented in terms of a
instance, in such societies people may be more single dimension or in simple black and white
likely to base trust on others’ predictability terms. The extreme cases presented here are
rather than their capabilities (Doney et al., 1998). meant to help explain the concept of the GLOBE
15-House.qxd 3/19/2004 4:51 PM Page 406
Assertiveness dimension and some expected to represent high versus low Assertiveness. In
correlates, but it should also be clear that cultures contrast, Hofstede places one concept (mas-
do not neatly fit into stereotypes. Although the culinity) at one extreme of the dimension and
table shows cultural attributes that conceptually another (femininity) at the other. As seen
should tend to cluster together, it does not rule earlier, at the individual level masculine and
out the fact that not all these attributes cohere at feminine characteristics are not necessarily at
all times. Societies usually have differing mixes opposite ends of a single dimension and people
of the extreme cases presented in this table (and can have both types of traits. Such masculine
in the other tables of this volume). and feminine traits are often found to be sepa-
rate and even unrelated dimensions (Fagenson,
1990). Although we are clearly talking about
THE GLOBE MEASURES different levels of analysis and different con-
OF ASSERTIVENESS structs, we feel the same may hold for “mascu-
line” versus “feminine” values in society. In
Project GLOBE, such feminine values are
Comparison of High Versus
reflected more in the Humane Orientation cul-
Low Assertiveness Societies
tural dimension (see Chapter 18 by Kabasakal &
As mentioned, Hofstede (1980, 2001) explic- Bodur). This separation is conceptually appeal-
itly links his tough–tender dimension in terms ing and also allows us to test the relationship
of values to sex roles and gender equality (or between the Assertiveness cultural dimension
inequality), even though his MAS index has no and the Humane Orientation cultural dimension.
items explicitly tapping assertive attributes or As stated, Hofstede’s MAS index items
behaviors. On the basis of the results of the two measure success drive and materialism. In Project
pilot studies reported in Chapter 8 by Hanges and GLOBE, such a drive to excel and perform highly
Dickson, the GLOBE research design separates is measured explicitly in the measure for Perfor-
the three underlying dimensions that comprise mance Orientation (see Chapter 12 by Javidan).
Hofstede’s MAS index and masculinity–feminin- Also, the GLOBE items measuring Assertiveness
ity dimension, namely: Performance Orientation, do not explicitly emphasize the individualistic
Assertiveness, and Gender Egalitarianism. This component of Hofstede’s masculinity dimension.
separation has intuitive appeal, as these three However, due to the shared emphasis of these
variables may or may not be associated with each three constructs on competition, dominance over
other in any particular culture. For instance, gen- the environment, and personal initiative we do
der role differences in a society may be minimal expect a positive relationship among Assertive-
(e.g., girls and boys are equally encouraged to ness, Performance Orientation, and Individualism.
attain higher education) and at the same time A caveat is in order. If indeed Hofstede’s
individuals in such a society may stress the observation holds that within virtually all soci-
necessity to be assertive, dominant, or aggressive eties men tend to hold more “masculine” values—
in relationships with others. On the other hand, including the subset of those that have to do
gender roles may be differentiated (e.g., it is eas- with Assertiveness—the sex distribution of
ier and more accepted for men to attain higher- respondents may influence results. As GLOBE
level positions than for women to do so) and at respondents are all middle managers, women
the same time the need to be submissive and are likely to be underrepresented. The GLOBE
nonassertive in business relationships within sample included 25.2% female and 74.8% male
organizations is stressed, such as in Japanese cul- middle managers. Whereas Hofstede (2001)
ture. This is a first major difference between the does report gender differences on the masculin-
GLOBE measures of Assertiveness and Gender ity dimension, Hoppe (1998) failed to find a sig-
Egalitarianism and Hofstede’s MAS index. nificant difference in the responses of men and
The GLOBE conceptualization of Assertive- women with respect to Hofstede’s dimensions.
ness also differs from Hofstede’s concept of The results may also be influenced by the fact
masculinity on several other grounds. The poles that middle managers have a certain position
of the GLOBE assertiveness scales are claimed in the organizational hierarchy. Their view on
15-House.qxd 3/19/2004 4:51 PM Page 407
Assertiveness– • –407
matters may differ from that of the shop floor across all the societies in the GLOBE study. As
or higher organizational levels. The views of these are new scales, no cross-cultural work
other groups in society may also differ. However, has been done on the concept of assertiveness as
as will be seen in Chapters 17 through 19 there a societal culture dimension in the way it was
is reason to believe that the middle manager conceptualized and operationalized in project
reports of the societal dimensions in Project GLOBE.
GLOBE are relatively representative of the soci- The two questionnaire items in Table 15.2a
ety in which they are embedded. Clearly, more are sample items that measure respondents’ As
research is needed to clarify these issues further. Is responses. These reflect the current societal
practices.
Table 15.2b shows sample items of the soci-
The GLOBE scales
etal scales for the managers’ expressed views of
Two constructs were created to measure a how assertive people in the society Should Be in
society’s level of Assertiveness: Society As Is social relationships. The responses to these
refers to the participants’ assessment of the extent items reflect the respondents’ values with
to which a society engages in Assertiveness prac- respect to Assertiveness practices.
tices, and society Should Be refers to the extent All scales range from 1 to 7. As was also
to which a society exhibits Assertiveness values. described in previous chapters, Project GLOBE
It is measured as a composite of quantitative measures both societal and organizational cul-
items directly asked from the participating man- tures. Responding managers filled out either
agers. Tables 15.2a an 15.2b and 15.3a and 15.3b the questionnaire measuring societal culture or
show sample questionnaire items that survived the questionnaire measuring organizational
rigorous statistical procedures (see Hanges & culture. Sample items for the organizational
Dickson, chapter 8). The scales were confirmed culture items measuring Assertiveness can be
found in Tables 15.3a (practices) and 15.3b detail below. As explained in earlier chapters, in
(values). By having separate samples of each country, data were collected from at least
managers independently complete these two two of three industries: food processing, finance,
parts of the questionnaire we eliminate the and telecommunications. Table 15.5 shows the
possibility of common source bias when relat- correlation between the societal Assertiveness
ing societal and organizational scores to each practices and values scales for each participating
other. A more detailed description of the method industry as well as overall. As seen in this table,
used in the study can be found in Chapter 8, this there is a significantly negative correlation
volume. (r = −.26, p < .05) between overall scores for
Assertiveness practices and Assertiveness val-
ues. Similar negative correlations are found in
THE FINDINGS: THE OVERALL the separate industries; however, for the subsam-
AND INDUSTRY SCORES ON ples in the food and telecommunications indus-
SOCIETAL ASSERTIVENESS try this correlation is not significant for the
telecommunication industry and only marginally
Table 15.4 shows the grand means of society significant for the food industry (see Table 15.5).
practices and society values scales for Asser- People’s aspirations regarding Assertiveness
tiveness across all countries represented in the in society are modestly (and negatively) related
GLOBE study. to their assessments of current levels of
Compared to the other cultural dimensions Assertiveness. Table 15.6 shows the individual
measured, the society practices score has a country scores on the practices scale and Table
midlevel average rating of 4.13. Scores range 15.7 shows the individual country scores on the
from 3.38 (Sweden) to 4.89 (Albania). The values scale.1
means for the Assertiveness values scale range In many cultures, people see their country as
from 2.66 (Turkey) to 5.56 (Japan). The individ- having a fair amount of assertiveness and they
ual country scores will be described in more want less of it. This was also concluded from
15-House.qxd 3/19/2004 4:51 PM Page 409
Assertiveness– • –409
Standard
Variable Mean Deviation Minimum Maximum Valid N
Telecommunication
Overall Food Industry Finance Industry Industry Score—
Industry Score Score—Practices Score—Practices Score—Practices Practices
Overall –.26*
score values N = 61
Telecommunications –.21
industry score values N = 32
the negative correlation between Assertiveness more Assertiveness than they currently have.
practices and values. For instance, Austria Many of these countries have relatively low
scores 4.62 out of a range of 1 to 7 on Assertive- levels of present Assertiveness. For instance,
ness practices (band A), whereas the score for China has a relatively low societal practices
Assertiveness values is only 2.81 (band C) (see score on Assertiveness (3.76) and a much
Tables 15.6 and 15.7). higher societal values score (5.44), as does
A similar pattern is found for many other Japan (3.59 on practices and 5.56 on values).
countries. For instance, Greece, former East This pattern of scoring lower on practices
Germany, and Turkey are also in the highest and higher on values is found for several
practices and lowest values bands (or group other Asian countries, including Indonesia,
of countries). A similar but somewhat less- Singapore, Hong Kong, the Philippines, India,
prominent pattern of wanting less Assertiveness and Malaysia. Other such countries include
is found for most other countries (respondents Zimbabwe, Zambia, Iran, and Slovenia. As can
in 40 of 61 societies studied want less Asserti- be seen in Tables 15.6 and 15.7, all but 3 (Hong
veness than is currently found in their soci- Kong, Georgia, and Singapore) of the 29
eties). However, in 21 cultures, including most cultures that are in the practices A band (this
Asian countries, people indicate they want band represents the highest scoring societies
15-House.qxd 3/19/2004 4:51 PM Page 410
Band
A B C
on current Assertiveness practices) express a societies scoring relatively low want more
desire for less Assertiveness than they currently Assertiveness.
perceive in practice. In contrast, all 12 cultures In summary, the pattern found in the
except Albania in the values A band (highest majority of countries shows the desire for less
scoring on desired Assertiveness) want more Assertiveness, perhaps reflecting the wish to
Assertiveness than they practice. In other belong to a relatively nonthreatening, nonag-
words, societies scoring relatively high on gressive society. However, as stated, a third of
current Assertiveness practices want less and the countries seem to want more Assertiveness,
15-House.qxd 3/19/2004 4:51 PM Page 411
Assertiveness– • –411
Band
A B C
and many Asian countries, especially, stress Clearly, additional research in this area needs to
assertiveness in this regard. Later in the chapter be conducted.
we will examine regional differences in asser- Also, as stated in the introduction to this
tiveness in more detail. It seems societies section, another problem in this type of research
scoring high on Assertiveness tend to strive for is that respondents in some cultures might be
less, and societies scoring low on Assertiveness reluctant to choose extreme scores on either side
express a desire for more. These results indicate of the scale when reporting their observations
that in contrast to some other culture dimen- about their societies’ level of Assertiveness.
sions, Assertiveness is not universally desirable As seen above, the means on the GLOBE
or undesirable. A possible explanation is that Assertiveness practices scales range from 3.38
too much Assertiveness may be felt to be a to 4.89. The same degree of reluctance may not
threat to internal integration. Too much compe- exist in reporting their aspirations and ambitions
tition, dominance, and aggression would create because the questionnaire items call for hypo-
an unsafe, untrusting environment, which thetical responses. The means for the assertive-
would hinder people from achieving required ness values scale range from 2.66 to 5.56. In
levels of cooperation to complete tasks. On the other words, the observed differences between
other hand, too little Assertiveness may be a practices and values may be a consequence of
threat to external adaptation and survival. As response propensities rather than real differ-
such, too little Assertiveness may be experi- ences. Also, response tendencies may be larger
enced as weakness. In most environments a in some cultures than in others. However, as
certain level of competition and dominance will be shown later, the GLOBE Assertiveness
is needed to survive. Such issues need to be dimension measures discriminate between
further explored in future studies. countries and geographic regions, thus indicat-
A few things need to be kept in mind. As ing that if there is a response tendency bias
pointed out by Hanges and Dickson in Chapter in some cultures, it is not so strong that it
8, sample characteristics and response biases overwhelms cultural differences. Also, as can
may influence the indicated levels of actual be seen in Appendix B (this volume), tests for
and desired Assertiveness. Respondents were all response bias for the Assertiveness values and
middle managers. This group may not represent practices measures reveal that such bias was
the society as a whole regarding Assertiveness. indicated in a small number of societies (includ-
On average, managers may well be more ing the Czech Republic, Morocco, Qatar, and
assertive and tough than nonmanagers in the Taiwan). With the exception of these societies
population. The items referring to Assertiveness there were no societal-level practices or values
may therefore be more favorably interpreted dimensions exhibiting response bias across
by managers than by other people in the man- several dimensions.
agers’ society. Also, in many societies man-
agers are predominantly male, and women in
managerial roles may need to exhibit more GLOBE ASSERTIVENESS AND
assertive leadership behaviors than women HOFSTEDE’S (1980) MASCULINITY SCALE
in general. Surprisingly, as discussed in the
chapter on Gender Egalitarianism by Emrich As we noted previously, Assertiveness origi-
and colleagues, we did not find significant nates in the masculinity–femininity culture
differences between men and women in terms dimension identified by Hofstede (1980).
of how they rated Assertiveness values and As was described above, Hofstede asked
practices. On the other hand, unobtrusive mea- respondents to rate the importance of work
sures presented in Chapter 9 (by Gupta, Sully de goals, such as earnings, recognition, advance-
Luque, & House) and other indicators of con- ment to higher-level jobs, and challenge (for the
struct validity throughout the book suggest that masculine pole) and a good working relation-
the responses of the middle managers reflect ship with one’s supervisor, cooperation, a desir-
the society in which they are embedded and able living area, and employment security
not the culture of middle managers exclusively. (for the feminine pole). Nothing in his MAS
15-House.qxd 3/19/2004 4:51 PM Page 413
Assertiveness– • –413
scale dealt explicitly or even implicitly with This tentative explanation supports the idea
assertiveness. In contrast, the GLOBE items that masculinity and assertiveness at the cul-
focus on whether people in a society are or tural level are indeed related and seems worthy
should be assertive, dominant, and tough or of future exploration.
nonassertive, nonaggressive, and tender. It is, of The relationships with the other GLOBE
course, interesting to test whether these new dimensions are of interest because assertive-
GLOBE Assertiveness scales have significant ness has not yet truly been studied as a societal
associations with the Hofstede (1980) masculin- or cultural attribute. Knowledge of these
ity scale. This was tested using the society-level relationships will also shed more light on the
scale scores reported in Hofstede’s book (1980, differences between the GLOBE Assertiveness
as well as later updates to these country rank- and Hofstede’s masculinity dimension. Given
ings). Those culture-level scores were corre- that both the GLOBE Assertiveness and
lated with the cultural-level scores from the Gender Egalitarianism cultural dimensions
GLOBE project. originate in Hofstede’s masculinity–femininity
The Hofstede masculinity scale significantly dimension (see Chapter 14 by Emrich,
correlated with the GLOBE Assertiveness Denmark, and Den Hartog), the relationship
practices scale (r = .37, p < .05). In other words, between these two is also of particular interest.
societies scoring high in Assertiveness practices A negative relationship between Assertiveness
are also more likely to be higher on the Hofstede and Gender Egalitarianism is expected. As
masculinity scale. However, Hofstede’s mas- described in Chapter 14, assertiveness has been
culinity scale scores did not correlate signifi- associated with men as well as “masculine”
cantly with the GLOBE Assertiveness values societies. According to Hofstede (1980), male-
scale (r = .17, p > .10). In other words, prefer- dominated societies should also be low on
ences for more assertiveness in societies (as gender egalitarianism.
measured by the GLOBE Assertiveness values Similarly, respondents who scored high on
scale) are not related to Hofstede’s measure of Assertiveness practices report that people are
masculinity. This raises a further question con- generally assertive, dominant, and tough rather
cerning what was being measured by Hofstede’s than nondominant and tender. Respondents
measure of masculinity, which purportedly scoring high on Assertiveness values means
measures values and not practices. they feel people in society should be encour-
Hanges and Dickson (in Chapter 8) point aged to be tougher and more dominant. It there-
out that a lack of correlations is not necessarily fore seems likely that high levels of both current
surprising given that GLOBE scales were not and aspired Assertiveness are related to valuing
constructed to build directly on Hofstede’s more Power Distance in society.
work. The measures are quite different, and Another relationship that is interesting to
Hofstede’s items do not explicitly address the test is the relationship between Assertiveness
question of Assertiveness in values. In fact, as and the Performance Orientation dimension,
stated earlier, there are no items in the especially when comparing our results to
Hofstede MAS scale that deal exclusively with Hofstede’s masculinity dimension. Hofstede
assertiveness. Rather, Hofstede’s MAS scale is explicitly links masculinity to striving for
similar to his individualism scale in that it success and valuing challenging work and
assesses the importance of the selected aspects material wealth. He holds that in masculine
of work. Therefore, the significant correlation societies, individuals strive for personal suc-
between the Hofstede MAS scale and the cess and material gain and in feminine soci-
GLOBE Assertiveness practices scale can be eties for a welfare-oriented society (Hofstede,
seen as surprising. The likely explanation 1991). Thus, as—according to Hofstede—
seems to be that GLOBE’s scale assesses assertiveness is one of the key features of such
assertive behavior patterns, and people in soci- masculine societies, a positive relationship
eties that are more assertive are indeed more between Assertiveness and success striving is
likely to value the types of work goals and expected. As stated earlier, on the basis of a
opportunities Hofstede identified as masculine. masculine society’s stress on competition and
15-House.qxd 3/19/2004 4:51 PM Page 414
individualism, a modest positive correlation practices and values correlate significantly and
with Individualism is also expected. positively (r = .32). In other words, in contrast
Also interesting in terms of the comparison to with the negative correlation between Assertive-
Hofstede’s masculinity dimension is, of course, ness practices and values, societies with higher
GLOBE’s Humane Orientation dimension (see levels of existing egalitarianism want even more.
Chapter 18 by Kabasakal & Bodur). Bajdo and The pattern of correlations between Assertive-
Dickson (2001) take humane orientation as an ness and Gender Egalitarianism supports treat-
indication of femininity. If femininity and ing the two as separate dimensions rather than
masculinity can be seen as “opposite ends of integrating them into a single dimension (which
a continuum,” a significant negative relationship is what Hofstede’s masculinity–femininity
between Assertiveness and Humane Orientation dimension suggests).
can be expected. However, the Chinese Culture As stated, Hofstede (1980) links masculinity
Connection (1987) reported a positive correla- to a performance-oriented society. Thus, posi-
tion between the MAS index and their measure tive relationships between the Assertiveness
of human-heartedness (which deals with social and Performance Orientation dimensions were
awareness, kindness, and being courteous to hypothesized. Only the relationship between
others; see also Chapter 14 by Emrich, Denmark, GLOBE Assertiveness values and Performance
& Den Hartog). Correlations of the societal-level Orientation practices was found to be significant
Assertiveness scales with the other GLOBE (r = .36). The relationship between the mea-
dimensions are reported in Table 15.8. (see also sures for Assertiveness and Power Distance also
Appendix A). presents partial support for our expectations. A
As Table 15.8 shows, the results support significantly positive relationship between
only some of the expectations. The hypothe- societal Assertiveness values and Power
sized relationship between Assertiveness and Distance values was found (r = .29). Striving for
Gender Egalitarianism is partially supported. more assertiveness seems to be accompanied by
Assertiveness practices are not correlated valuing power distance in society.
with Gender Egalitarianism (practices or As can be seen in Table 15.8, the rela-
values). However, the correlation between tionship between the GLOBE measures of
GLOBE’s Assertiveness values and Gender Assertiveness and Humane Orientation is not
Egalitarianism values (r = −.28) is negative. as straightforward as expected. In line with
This makes sense because striving for more expectations, the GLOBE Humane Orientation
assertiveness is related to striving for less practices scale correlates negatively with the
gender equality or, in other words, for more Assertiveness practices scale. Thus, in line
“male-oriented” values. with expectations, societies whose aggregated
This general pattern does not always hold, reports indicate a high current Humane
though. For example, according to its middle Orientation also report less assertiveness and
managers, behavior in social relationships in aggressiveness in social relationships (r =
Sweden tends to be tender and nonassertive −.42). However, the Assertiveness practices
(lowest rank). Swedish managers also indicate measure correlates significantly positively with
they want slightly more Assertiveness. Also, the Humane Orientation values measure (r =
Sweden is a country where equality between .35), and the Assertiveness values measure cor-
men and women is relatively high. Swedish relates positively with the Humane Orientation
managers indicate that they strongly feel that practices measure (.45).
such equality should be advanced even further GLOBE’s Assertiveness practices scale also
(Akerblom & Holmberg, in press). In other correlates significantly negatively with the
words, in this specific case, Assertiveness and measure for Institutional Collectivism prac-
Gender Egalitarianism are in the same direc- tices (r = −.42). This measure for Collectivism
tion regarding practices, but not necessarily refers to collective interests and work (see
values. also Chapter 16 by Gelfand, Bhawuk, Nishii,
Interestingly, as reported in Chapter 14 by & Bechtold). Thus, in line with expectations,
Emrich and colleagues, Gender Egalitarianism societies reported to be assertive, or “tough,”
15-House.qxd 3/19/2004 4:51 PM Page 415
Assertiveness– • –415
N = 61.
* p < .05.
** p < .01 (2-tailed).
also indicate lower current levels of social practices and values show different patterns of
solidarity. In other words, in such societies a relationships.
more individualistic approach is taken.
However, Assertiveness values are positively
correlated with the practices of both forms of
Collectivism (.37 for Institutional Collectivism GLOBE ASSERTIVENESS AND OTHER
and .28 for In-Group Collectivism). In other ECONOMIC AND SOCIAL INDICATORS
words, societies that scored high on these
forms of collectivism tend to strive for more In this section, we compare and contrast the
Assertiveness. findings on GLOBE Assertiveness measures
The patterns show that the relations with those of other major cross-cultural and
between the culture dimensions are compli- comparative studies focusing on different
cated and not as straightforward as one might countries’ social and economic performance.
expect. Some of the relationships we find As explained in chapter 7 by Javidan and
support Hofstede’s work on the masculinity Hauser, four major ongoing reports produce
dimension, others run against it. Often the the relevant data: the World Competitiveness
15-House.qxd 3/19/2004 4:51 PM Page 416
Yearbook, the United Nations Human health. Most of these indicators are available
Development Report, the World Development for only 40 of the GLOBE-sampled countries;
Indicators, and the World Values Survey. The therefore, some caution is needed in interpreting
findings of these studies are grouped into two these results.
categories: economic health and the human
condition.
Assertiveness and Human Condition
As stated in Chapter 7 by Javidan and
Assertiveness and Economic Health
Hauser, economic health is only one aspect of
Is there a relationship between Assertiveness a society’s well-being. Another dimension is
as a societal culture dimension and economic the general health of the populace and the soci-
health? Are societies with high Assertiveness ety. The term human condition used here refers
practices or values scores more competitive? Do to the general quality of life and the state of
such societies enjoy healthier economies? mind of the people in a particular country. In
Triandis (1994, p.129) holds that the rapid this section, we examine the relationship
increases in standard of living among the high between GLOBE findings on assertiveness and
masculinity countries (in Hofstede’s terms) sug- different dimensions of the human condition.
gests that masculinity may be relevant to eco- In many countries, our respondents indicate
nomic development. Similarly, this may be the that they strive for less assertiveness, aggres-
case for assertiveness. In other words, a positive siveness, and dominance in interpersonal rela-
relationship between Assertiveness practices tionships within their society than they
and economic health is expected. Accordingly, currently experience. In Chapter 7 by Javidan
in Chapter 7 of this volume, the following and Hauser, the following hypothesis concern-
hypotheses were proposed: Cultures that score ing the relationship between the GLOBE mea-
higher on Assertiveness tend to perform better sure of Assertiveness and the human condition
in terms of global competitiveness and be more was proposed: Societies that score higher on
economically prosperous. Assertiveness tend to exhibit lower levels of
Table 15.9 shows the correlation coeffi- human health. As stated, several variables asso-
cients between the two GLOBE measures of ciated with human health were examined (see
Assertiveness (practices and values), and the Chapter 7 for a detailed description of these
various elements of economic health and global measures). Table 15.10 reports the relation-
competitiveness. These measures are described ships between GLOBE Assertiveness and these
in more detail in Chapter 7 by Javidan and measures. Table 15.10 shows that none of
Hauser. the measures of human health are significantly
Table 15.9 shows that peoples’ aspirations in related with either the Assertiveness practices
terms of how much a society should value scale or values scale.
assertiveness, as well as their perceptions of Two of these indicators are available for
how much a society practices assertiveness, are only 38 or 40 of the GLOBE countries; there-
not substantially related to the country’s eco- fore, some caution is needed in interpreting
nomic performance. Of the 14 presented coeffi- these results. The specific subset of countries
cients, only one is significant (in the expected available may mask possible effects. Further
direction). A significantly positive correlation is research in this area is required. For instance,
found between GLOBE Assertiveness values the results of a study reported by Arrindell
and success in science and technology. Thus, and colleagues (1997) suggest that the rela-
although the one relationship that is found is in tionships between assertiveness and well-being
the expected direction, it does not present strong may be somewhat more complicated. They
evidence in support of our hypotheses. found a significant negative interaction between
Therefore, one can conclude that no (or only Hofstede’s masculinity and national wealth in
very limited) support was found for the predicting national subjective well-being
expected positive relationships between (SWB) levels: In the poorer countries, mas-
GLOBE Assertiveness and societal economic culinity correlated positively with SWB,
15-House.qxd 3/19/2004 4:51 PM Page 417
Assertiveness– • –417
Human
Societal Human Life General Development
Assertiveness Health Health Expectancy Satisfaction Index
Table 15.11 Relationship Between Assertiveness, Religious Values, and Political Ideology
Assertiveness– • –419
Such violence occurred more frequently and show less violent behavior (Van de Vliert
in warm (M = 24°C/76°F) than in cold et al., 1999).
(M = 17°C/62°F) or hot (M = 30°C/86°F) In contrast, in warmer climates, greater male
countries. In a replication of this study, Van de investment in mate seeking leads to immediate
Vliert and colleagues (1999) controlled for gratification and increases the number of off-
population size and density, socioeconomical spring. This more likely produces dominant and
development, and political democracy. The contentious attitudes and behaviors toward
resultsm of their study using 136 countries show women and rivals in men. Asymmetrical
that, indeed, countries with moderate climates parental roles are likely to emerge under such
suffer much more from domestic political vio- conditions, with men not needing to show much
lence than colder countries, and slightly more concern for offspring and women expected
than hotter countries. to invest in child-rearing and provisioning
Van de Vliert and colleagues (1999) intro- (e.g., Miller, 1994). Thus, it can be expected that
duce cultural masculinity as a possible mediator societies in warmer regions would tend to evolve
between temperature and aggression. Following toward greater cultural masculinity, with men
Hofstede, they describe high cultural masculin- socialized to deal with interpersonal problems
ity as characterizing societies in which men are through aggressive competition rather than inte-
expected to be dominant, assertive, tough, and grative cooperation (e.g., Van de Vliert, 1998).
focused on material success, whereas women
are expected to be subordinate, modest, tender, In sum, Paternal Investment Theory can explain
and concerned with quality of life. In contrast, why culturally more masculine societies, charac-
low masculinity (or cultural femininity) is found terized by more dominance and violence, evolve
in cultures in which both men and women are in warmer climates. In more masculine compared
expected to be subordinate, modest, tender, and with less masculine societies, issues of conflict
concerned with quality of life. The explanation are more likely to be managed through fight rather
they provide for the proposed relevance of mas- than flight and through aggression rather than
culinity derives from paternal investment theory negotiation. (Van de Vliert et al., 1999, p. 301)
(e.g., Bjorklund & Kipp, 1996; Coltrane, 1988;
Miller, 1994; Van de Vliert et al., 1999). This Van de Vliert and his coauthors also suggest
evolutionary social psychological theory that similar to cold countries, it is harder to
derives from the biological construct of differ- ensure survival in very hot countries than it is
ential parental care. According to this theory it in moderately warm countries, leading to an
is more arduous for families to meet basic increased need for paternal investment in
needs for safety, food, and security in colder cli- survival of the family. Thus, they propose an
mates. Therefore, more parental investment in inverted U-shaped association between temper-
the family is needed for survival of mother and ature and cultural masculinity. They also pro-
offspring. Both paternal and maternal invest- pose a linear relationship between masculinity
ment in the family is characterized by sacri- and domestic political violence, thus producing
fice, delay of gratification, and the evolution an overall curvilinear relationship between tem-
of cooperative attitudes and behaviors. Such perature and violence.
symmetrical parental roles produce normative In a GLOBE subsample of 53 countries
pressure on husbands to share responsibility for for which a MAS score was available, national
caring for children with their wives. differences on the cultural masculinity dimen-
The emphasis on cooperative behavior in sion were found to account for the curvilinear
the nuclear family would likely generalize to temperature–violence association, suggesting
other societal behavior, influencing people to that culture mediates this association (Van de
make cooperative rather than competitive Vliert et al., 1999). Assertiveness and MAS are
choices when dealing with others. Greater obviously different (as described and tested);
parental investment in the family in colder however, it is interesting to see whether similar
climates would then produce, according to this relationships are also found between climate and
theory, cultures that are lower in masculinity Assertiveness as measured in the GLOBE study.
15-House.qxd 3/19/2004 4:51 PM Page 420
As was indicated in Chapter 10 (Gupta & The climatic means of societal practices
Hanges), the GLOBE societal cultures were and values of Assertiveness are given in Table
categorized in terms of their climate. We created 15.12. The means for practices ranged from
seven categories of climate: tropical humid, 3.96 (tropical humid) to 4.30 (Mediterranean);
tropical wet and dry, desert, subtropical humid, however, these differences did not reach
subtropical wet and dry (Mediterranean), marine statistical significance. With respect to the
west coast (maritime), and continental. One-way values scale, tropical humid climate region
ANOVA tests were conducted to determine (mean = 4.37, contrast = 0.62, p < .01) had a
whether climatic differences are found in the significantly higher average; and maritime
practices and values of societal Assertiveness. climate region (mean = 3.43, contrast = −.43,
No significant differences are found for practices p < .05) had a significantly lower average. To
(F6, 54 = .66, p > .10) mean scores across differ- summarize, in our sample, climate has a very
ent climate types, but weak cross-climatic differ- weak influence on the values measure of soci-
ences exist in values (F6, 54 = 1.98, p < .10) etal Assertiveness. Specifically, respondents in
mean scores of societal Assertiveness. Further societies with a maritime climate tended to
analysis showed that climatic differences value softer, nonassertive approaches, whereas
account for 18% of the variation in societal respondents in the humid tropics seem to
Assertiveness values scores, whereas 82% repre- aspire for more assertiveness within their
sent idiosyncratic societal differences. society.
Standard Standard
Climate Clusters N Mean Deviation Mean Deviation
Assertiveness– • –421
Standard Standard
Climate Clusters N Mean Deviation Mean Deviation
Qatar
South Africaa
South Africab
Turkey
Subtropical Humid 5 3.60 0.74 4.16 0.34
Argentina
Bolivia
Brazil
Hong Kong
Taiwan
Mediterranean 7 3.83 0.56 4.30 0.42
Albania
Greece
Italy
Morocco
Portugal
Slovenia
Spain
Maritime 10 3.43 0.32 4.10 0.45
Denmark
France
Germanyc
Germanyd
Ireland
Netherlands
New Zealand
Switzerland
Switzerlande
United Kingdom
Continental 13 3.97 0.83 4.09 0.43
Australia
Austria
Canadaf
China
Finland
Georgia
Hungary
Japan
Poland
Russia
South Korea
Sweden
U.S.
In the climate analysis presented above, the temperature (r = .34) was also significant at
countries were classified into different climates the .05 level. Thus, Assertiveness values are
rather than examining temperature the way the related to DPV.
aforementioned study by Van de Vliert and col- Van de Vliert and colleagues (1999) also
leagues (1999) did. As they did find a signifi- investigated the possible mediating role of mas-
cant relationship between temperature and the culinity in the relationship between temperature
MAS index, further exploration of the relation- and DPV. They investigated curvilinear rela-
ship between temperature and Assertiveness tionships between the variables and controlled
was of interest. Fifty-four of the GLOBE coun- for several possible confounding variables
tries are also in the analyses presented by Van (e.g., population size and democracy). We did
de Vliert and coauthors (1999), excluding not perform similar analyses here as they are
Georgia, Kazakhstan, and Qatar. Using the beyond the scope of this chapter. However,
data on temperatures presented by Van de Vliert the positive correlation between GLOBE
and colleagues (pp. 308–311), we calculated a Assertiveness values and both rainfall and
correlation between temperature and the two DPV seem interesting topics for further explo-
societal Assertiveness scores. Assertiveness ration. Along with the negative correlation
practices did not correlate with temperature between Assertiveness values and Gender
(r = −.06). The correlation between societal Egalitarianism values, these data may help
Assertiveness values and average daytime further develop ideas described by Van De
temperature of the country’s capital city was Vliert and colleagues (1999).
significant at the 5% level (r = .28) for these
54 societies. However, for the full sample of
Geographical Regions
61 cultures, using data explained in Chapter 10
by Gupta and Hanges, the correlation of values As discussed in Chapter 8 by Hanges and
with temperature measure is not significant at Dickson, there are significant differences in
the .05 level, but only at the .10 level (r = .22, values and practices for every cultural dimension
p < 0.10). Also, the Assertiveness values score among the 62 societal cultures (recall that we
is significantly positively correlated with average use the term societal cultures rather than
rainfall (r = .36, p < .01, N = 61). Thus, although nations). To explore further the possible impact
the specific countries and data sets used in of country demographics, the relationship
the calculation seem to influence the strength of between the GLOBE Assertiveness measure
the relationships that are found, on the whole and geographic region was examined. As dis-
Assertiveness values are to some degree related to cussed in Chapter 10, GLOBE societies were
climate, whereas practices are not. grouped into 10 regional clusters. One-way ana-
Following the Van de Vliert and colleagues lyses of variance show that significant cross-
(1999) study, we also assessed the relationship regional differences exist in the societal practices
between Assertiveness and domestic political (F9, 51 = 2.41, p < 0.05) as well as in the
violence (DPV). They used the most recent societal values (F9, 51 = 5.40, p < 0.01) of
30-year period for which violence scores were societal Assertiveness. Further analysis showed
available (1948–1977). The index was based that regional differences account for 30% of
on the logarithm of an additive combination the variation in societal Assertiveness prac-
of two indicators, namely the number of polit- tices scores, whereas 70% represent idio-
ical riots and armed attacks (see Van de Vliert syncratic societal differences. With respect to
and colleagues for a more detailed description societal Assertiveness values scores, 49% of the
of the origins of this measure and the data on variation was explained by regional differences,
DPV). Assertiveness practices did not correlate whereas 51% represented unique societal differ-
significantly with DPV (r = .15); however, the ences. In other words, a substantial amount of
correlation between societal Assertiveness val- variance in both societal Assertiveness practices
ues and DPV was significant at the .05 level and values is accounted for by differences in
(r = .27). The correlation between DPV and geographical regions.
15-House.qxd 3/19/2004 4:51 PM Page 423
Assertiveness– • –423
Table 15.13 shows the mean scores on values. The lowest values score is found for
Assertiveness values and practices for each the Germanic Europe (3.07, contrast = −.81,
geographic region. The highest practices p < .01) and Middle East clusters (3.39,
scores are found for the Germanic Europe contrast = −.47, p < .05).
(4.55, contrast = .48, p < .01) and Eastern It is noteworthy that in all regions but two
Europe clusters (4.33, contrast = .24, p < .10). (and these two score highest on Assertiveness
The lowest average practices scores are found values), people tend to strive for less assertive-
for the Nordic Europe (3.66, contrast = −.50, ness. Only in the Southern Asia and Confucian
p < .05) and Southern Asia clusters (3.86, Asia clusters did people indicate they wanted
contrast = −.28, p < .10). The Southern Asia more assertiveness, dominance, and aggression
(4.65, contrast = .92, p < .01) and Confucian in their relationships with one another; these are
Asia (4.54, contrast = .80, p < .01) clusters two of the lowest-scoring regions with respect to
have the highest scores on Assertiveness current GLOBE Assertiveness practices.
Standard Standard
Geographic Region N Mean Deviation Mean Deviation
(Continued)
15-House.qxd 3/19/2004 4:51 PM Page 424
Standard Standard
Geographic Region N Mean Deviation Mean Deviation
a Switzerland (French-speaking)
b Germany (West): Former FRG
c Germany (East): Former GDR
d South Africa (Black Sample)
e Canada (English-speaking)
f South Africa (White Sample)
15-House.qxd 3/19/2004 4:51 PM Page 425
Assertiveness– • –425
THE RELATIONSHIP BETWEEN SOCIETAL Table 15.14 shows the grand means of
AND ORGANIZATIONAL PRACTICES AND organizational practices and organizational val-
ues scales for Assertiveness across all identifi-
SOCIETAL AND ORGANIZATIONAL VALUES
able responding organizations in the GLOBE
study.
As indicated in Chapter 2 by House and Javidan, The organizational practices score has a
the GLOBE theoretical model postulates that midlevel average rating of 4.11. Scores range
societal practices and values affect organiza- from 3.16 to 5.88. Compared with other cultural
tional practices and values. Two hierarchical lin- dimensions measured, the organizational values
ear models (HLMs) were conducted to test these scores have a low average rating of 3.96. Further,
hypotheses for organizational Assertiveness there is a huge interorganizational variation in
practices and values. We tested the GLOBE the preferred assertiveness levels, ranging from
hypothesis regarding the effect of societal culture 1.71 to 6.50.
on organizational culture by conducting HLM Table 15.15 shows the correlation between
analyses in which organizational Assertiveness the organizational Assertiveness practices
was predicted by societal Assertiveness. These and values scales. As seen in this Table,
analyses did not support our hypotheses that there is a significant positive correlation
societal Assertiveness practices have a signifi- (r = 0.38, p < 0.01) between overall scores
cant and strong positive relationship with for Assertiveness practices and Assertiveness
organizational Assertiveness practices (p > .05). values.
However, we found a significant relationship One-way ANOVA tests were conducted
between societal Assertiveness values and orga- to determine whether industry differences are
nizational Assertiveness values (p < .01). There- found in the practices and values of organi-
fore, for this culture dimension, only the analysis zational Assertiveness. No significant differ-
for Assertiveness values supports the principal ences are found for practices (F2, 271 = .07,
proposition in the GLOBE theoretical model p > .10) mean scores across three industries.
(i.e., Proposition 3, Figure 2.1, Chapter 2): soci- Similarly no significant differences are found
etal cultural values and practices affect organiza- for values (F2, 271 = .48, p > .10) mean scores
tional cultural values and practices.2 across industries.
Standard
Variable Mean Deviation Minimum Maximum Valid N
Additional one-way ANOVA tests were Assertiveness practices scores, and for 23% of
conducted to determine if regional differences the variation in organizational GLOBE assertive-
are found in organizational Assertiveness. ness values scores.
Significant differences are found for practices Table 15.16 reports mean practices and
(F9, 266 = 6.97, p < .01) mean scores across values scores of organizational Assertiveness
regional cultures. Similarly, significant differences in different regions. Assertiveness practices are
are found for values (F9, 266 = 9.01, p < .01) rated significantly higher in the Southern Asia
mean scores across regional cultures. Further (4.39; contrast = .33, p < .01) and Anglo (4.37;
analysis showed that regional differences account contrast = .31, p < .01) clusters. Assertiveness
for 19% of the variation in organizational practices are rated as lower in the Middle East
15-House.qxd 3/19/2004 4:51 PM Page 427
Assertiveness– • –427
N = 276.
** p < .01 (2-tailed).
Organizational Organizational
Assertivenesss Assertiveness
Values Practices
Standard Standard
Geographic Region N Mean Deviation Mean Deviation
(3.99) and Latin America clusters (3.99), but of Assertiveness in Anglo organizations appear
these are not significantly different from the to be consistent with a high emphasis on
rest of the regional clusters. Assertiveness assertiveness found in the Anglo business
values are rated significantly higher in the literature.
Anglo (4.36; contrast = .49, p < .01), Confucian
Asia (4.35; contrast = .47, p < .01), and Latin
Europe (4.31; contrast = .43, p < .05) clusters, ASSERTIVENESS AS A
and lower in the Middle East (3.24; contrast PREDICTOR OF CULTURALLY
= −.75, p < .01) and Germanic Europe (3.43; ENDORSED LEADERSHIP THEORIES
contrast = −.54, p < .01) clusters.
In summary, Anglo organizations stand out One of the most interesting questions that
for their reported positive proclivity toward the GLOBE project can answer concerns the
assertiveness. The higher practices and values relationship between perceived leader attributes
15-House.qxd 3/19/2004 4:51 PM Page 428
and culture. Culture dimensions such as those organizational practices. As indicated previously,
studied in GLOBE yield many hypotheses our notions of values and CLT leadership dimen-
regarding cross-cultural differences in leadership sions represent idealized concepts of how the
behaviors and attributes. For instance, cultures world “Should Be” in contrast to practices that
rated high in Assertiveness are probably more represent the world “As Is.” As you read through
tolerant of strong, directive leaders than cultures the results discussed below, it may be helpful
rated low, in which a preference for more consul- to view Figure 15.1 for a visual summary. The
tative, considerate leaders seems likely. figure, however, only shows results regarding
Hofstede (1991) also speculates that masculine cultural values, not practices. (All HLM coeffi-
and feminine societies create different manage- cients are presented in Table 21.10 of Chapter 21
ment hero types, the masculine manager being by Dorfman, Hanges, & Brodbeck.)
assertive, decisive, and aggressive: As described earlier in this volume, culture
is believed to have its effect on the content of
He is a lonely decision-maker looking for facts CLTs at multiple levels of analysis. Leader
rather than a group discussion leader. It does attributes perceived to be effective might be
not hurt if he is slightly macho. . . . The manager a function of (a) societal cultural practices as
in a feminine country is less visible, intuitive measured by the societal practices culture
rather than decisive, and accustomed to seeking scales, (b) societal cultural values as measured
consensus. (p. 94) by the societal values scales, (c) organizational
cultural practices as measured by the organiza-
One would expect that leaders would be tional practices scales, and (d) organizational
expected to be tough, autonomous, and strong cultural values as measured by the organiza-
in cultures emphasizing assertiveness and tional values scales. In the present chapter we
dominance in social relationships. discuss the results of statistical analyses exam-
As stated earlier in this volume, we identi- ining the extent to which the cultural dimen-
fied 21 basic leadership factors (created from sion of Assertiveness has an effect on CLTs.
112 items) that were later grouped into 6 global Specifically, we examine the extent to which
leadership dimensions by conducting first-order the content of CLTs varies as a function of
and second-order factor analyses, respectively. Assertiveness values and practices in societies
We identified six underlying dimensions of and organizations within societies.
global leadership patterns that are viewed by We tested for the relationship between
managers as contributors or impediments to out- culture and the CLT leadership dimensions by
standing leadership (see Chapter 8 by Hanges & using hierarchical linear modeling (HLM).
Dickson). These six leadership dimensions are: Although this statistical technique is some-
Charismatic/Value-Based leadership, Team- what new to the organizational literature, it
Oriented leadership, Participative leadership, actually has been used in other fields for quite
Humane-Oriented leadership, Autonomous lead- some time. HLM can be thought of as a
ership, and Self-Protective leadership. multistep process designed to test the signifi-
cance of relationships between independent
and dependent variables at multiple levels of
Interpretation of HLM Using
analyses. An overview of HLM analyses and a
GLOBE Assertiveness to Predict
detailed discussion of how these analyses were
Culturally Endorsed Leadership conducted is provided in Chapter 11 by Hanges,
In this section of this chapter we present Dickson, and Sipe, and in Appendix C by
relationships between one culture dimension Hanges, Sipe, and Godfrey.
(Assertiveness) and the six CLTs. Competitive
tests of all culture dimensions and CLTs are pre- HLM Analysis: Organizational
sented in Chapter 21 by Dorfman and coauthors.3
and Societal Variation
In general, we expect that societal and organiza-
tional values will be more strongly related to In this analysis we examine the simultaneous
CLT leadership dimensions than societal and associative relationships of organizational and
15-House.qxd 3/19/2004 4:51 PM Page 429
Assertiveness– • –429
Figure 15.1 Assertiveness Cultural Values as Drivers for CLT Leadership Dimensions
* Only statistically significant relationships are shown ( p < .05; see Table 21.10, Chapter 21 by Dorfman et al.). The most
important leadership CLT relationships are bolded (i.e., relationship is significant at both society and organization levels of
analyses or highest HLM coefficient within each level of analysis).
O = Organizational level
S = Societal level
leadership dimension, as were the organizational leadership belief system in organizations reported
Assertiveness cultural values scores ( p < .05). to have Assertiveness practices and values.
Participative leadership is less likely to be a part
of the shared leadership belief system in organi- • Self-Protective leadership. Assertiveness
zations reported to have Assertiveness prac- cultural practices and values scores were not
tices and it is less likely to be a part of the significantly related to the Self-Protective lead-
shared leadership belief system in organizations ership dimension.
reported to espouse Assertiveness values.
Assertiveness– • –431
the relationship among masculine and feminine Assertiveness practices scale. However, it did
practices, values, and leadership dimensions not correlate significantly with the GLOBE
is not as straightforward as suggested by Assertiveness values scale. Thus, the Hofstede
Hofstede’s (1980) masculinity–femininity masculinity index seems to reflect assertive
dimension. This was also confirmed by the practices rather than values. Hence, the GLOBE
pattern of correlations between the Assertive- Assertiveness dimension differs from the
ness and Humane Orientation societal culture Hofstede masculinity dimension. One crucial
measures presented earlier. difference is that Hofstede’s masculinity dimen-
If you look at the overall picture (including sion confounds gender inequality with success
the other culture dimensions presented in the striving—constructs that in our view are not
other chapters), values seem to predict more necessarily correlated. Further, his MAS scale
variance in CLTs than practices. What people included no items relevant to assertiveness. For
think about their effective leaders seems to instance, a society could at the same time be
be somewhat more influenced by their values highly assertive but not very achievement or
than by the practices they perceive. Given the performance oriented. The GLOBE study
fact that people were asked to think of effective measures these separately. In our view, the
leaders, this is not surprising. Both the values pattern of correlations between these constructs
and the CLTs have a “should be” component. supports separating these constructs as well as
They are a reflection of a desired rather than an separating values and practices.
actual reality. It would be interesting to test We found that the GLOBE Assertiveness
whether actual leader behavior is predicted practices measure was not correlated with
more by the practices than the values. Gender Egalitarianism (practices or values).
Finally, we found that the coefficients for However, Assertiveness values and Gender
Assertiveness were stronger at the organizational Egalitarianism values are negatively correlated,
level than at the societal level of analysis. meaning that preferring more assertiveness is
Organizational Assertiveness had a stronger rela- related to preferring less gender equality or more
tionship with Team-Oriented and Participative male-oriented values. The significant relation-
leadership (all negative), and Humane-Oriented ship of Assertiveness values we found with the
and Autonomous leadership (both positive) as current practices of Performance Orientation is
compared to societal Assertiveness. Again, given also of interest. Hofstede (1980, 1991, 2001)
the fact we were asking respondents to make explicitly links masculinity to striving for suc-
judgments about organizational leadership, this cess, progress, and material wealth. Thus, a posi-
is not surprising. tive relationship between Assertiveness and the
Performance Orientation dimension would have
been plausible but was found only for Assertive-
SUMMARY AND CONCLUSIONS ness values in relation to practices, not aspirations
of Performance Orientation.
The GLOBE culture dimension questionnaire on The Assertiveness values scale was positively
Assertiveness asks whether there are differences related to the Power Distance values measure.
among societies in the extent to which people In addition, Assertiveness practices correlated
in those societies generally practice or value negatively with Institutional Collectivism prac-
assertiveness, aggressiveness, dominance, and tices, and Assertiveness values were positively
toughness or nonassertiveness and tenderness. related with both Institutional and In-Group
In this chapter we have demonstrated that orga- Collectivism practices scales.
nizations reflect Assertiveness cultural values, The most intriguing results may be those
but not practices, in the society in which they related to the GLOBE Humane Orientation cul-
are embedded. ture dimension. If this orientation can be taken
As stated, this dimension originates in as an indication of femininity, as measured by
Hofstede’s (1980) masculinity construct, and Hofstede, one might expect negative relation-
Hofstede’s MAS scale indeed correlated sig- ships of this measure with GLOBE Assertiveness
nificantly and positively with the GLOBE and Hofstede’s masculinity. Humane Orientation
15-House.qxd 3/19/2004 4:51 PM Page 432
and Assertiveness practices, as measured in and colleagues and only included the countries
project GLOBE, were indeed negatively corre- they included. Clearly, more research is needed
lated at the societal level. However, Humane on this issue.
Orientation values were positively related to The findings presented in this chapter also
Assertiveness practices. Furthermore, when soci- point to the conclusion that people in different
eties are reported to value Assertiveness, the CLT geographic areas rate their societies differently
endorsed at the societal level of analysis more in terms of how assertive they should be, in con-
likely includes Humane-Oriented leadership. In trast to the current state of assertiveness within
contrast, at the organizational level, practicing their society. That is, there are regional differ-
and valuing Assertiveness the CLT less likely ences as to the actual difference between these
included Humane-Oriented leadership. This rela- two measures. In all regions but two (and these
tionship again indicates that the relationship two score highest on Assertiveness values),
among masculine and feminine practices, values, people tend to strive for less assertiveness. Only
and leadership dimensions is not as straightfor- the Southern Asia and Confucian Asia clusters’
ward as suggested by Hofstede’s (1980) mas- respondents indicate they want more assertive-
culinity–femininity dimension. Further research ness, dominance, and aggression in their rela-
in this area is needed. tionships with others. It is, therefore, not surprising
Few relationships were found with external that in the sample as a whole Assertiveness
data. For instance, Assertiveness values were pos- practices and Assertiveness values correlate
itively related to “success in science and tech- negatively. Overall, in most regions (with the
nology” and “respect for family and friends.” notable exception of Asia), less Assertiveness is
As presented in Chapter 8 by Hanges and reported as desired.
Dickson, the GLOBE Assertiveness values We also reported on Assertiveness values
scale was also expected and found to relate and practices at the organizational level. Whereas
significantly negatively to Schwartz’s (1994) a negative correlation between Assertiveness
egalitarian commitment. The more societies practices and values is found at the societal
value assertiveness, the less such a commitment level, a positive correlation is found at the organi-
to voluntarily helping others and transcending zational level. Although no industry differences
selfish interests is stressed. were found, there is a large variation between
Van de Vliert and colleagues (1999) found organizations, and, again, regional differences
significant relationship between Hofstede’s mas- are found. Aggregated responses from Anglo
culinity index and domestic political violence organizations, especially, tend to score high on
(DPV) over the most recent 30-year period Assertiveness, which seems consistent with the
for which violence scores were available emphasis in Anglo literature and media on indi-
(1948–1977). We also assessed the relationship vidualism and aggressive attitudes needed to
between Assertiveness and the DPV data over succeed in business.
this period. Assertiveness practices did not corre- Another interesting part of GLOBE Project
late significantly with DPV; however, the corre- concerns the relationship between perceived
lation between societal Assertiveness values and leader attributes and culture. The relationship
DPV was positive and significant. Comparing between societal level Assertiveness and
Assertiveness in climate types did not yield sig- Humane-Oriented leadership was mentioned
nificant differences; however, Van de Vliert and above. The relationships at the organizational
colleagues (1999) found a significant relationship level are stronger and in line with expectations.
between Hofstede’s MAS index and temperature. For instance, we found that the less that organi-
We calculated a correlation between temperature zations practice and value Assertiveness, the
and the two GLOBE societal Assertiveness more likely it is that the endorsed societal-level
scores and found that GLOBE Assertiveness CLT includes Participative leadership. That
practices did not correlate with temperature, is, organizations whose managers scored rela-
whereas Assertiveness values were weakly corre- tively high on actual or desired Assertiveness
lated with temperature. These findings held when (dominance and toughness) prefer less Partici-
we used the temperature data of Van de Vliert pative leaders. This also holds for Team-Oriented
15-House.qxd 3/19/2004 4:51 PM Page 433
Assertiveness– • –433
leadership. In contrast, in organizations valuing accounted for by the predictor at only this level.
and practicing Assertiveness, it was more likely Whereas we have primarily taken the conservative
that the Humane-Oriented and Autonomous approach and reported the R2 Total in GLOBE, several
leadership styles were part of the CLT. scholars suggest that R2 Societal provides a more accu-
In this chapter we presented our ideas and rate description of aggregated relationships. For further
the picture of findings regarding Assertiveness discussion, see the paper by Lance and James (1999).
as a culture dimension. The pattern we found is 3. Results between the single HLM and multiple
intriguing and in some cases puzzling. More HLM tests will likely differ somewhat. The differ-
research is needed to fully understand the pattern ences between the results of the multiple HLMs and
of relationships unearthed here. However, in our single HLMs are conceptually similar to the differ-
view, the findings do support GLOBE’s Assertive- ences between a multiple regression analysis and a
ness cultural dimension as an interesting and correlation coefficient. Table 21.10 in Chapter 21 by
valuable, but also truly complex, dimension of Dorfman et al. presents both single and multiple HLM
national and organizational cultures. coefficients. In addition, the relationships for all cul-
ture dimension values are summarized in Chapter 3.
ENDNOTES
REFERENCES
1. Our response bias correction procedure identi-
fied response bias in some countries for this scale. Akerblom, S. & Holmberg, I. (in press). “Primus inter
We recomputed the predicted response bias corrected pares”: Leadership and culture in Sweden. In
scale score for each country. Response bias corrected R. J. House & J. Chhokar (Eds.), Cultures of the
scores are: world: A GLOBE anthology of in-depth descrip-
Practices: Denmark, 4.04 (no change in band); tions of the cultures of 14 countries (Vol. 1).
Finland, 4.05 (no change in band); France, 4.44 Thousand Oaks, CA: Sage.
(moves from band B to band A); Qatar, 4.39 (moves Arrindell, W. A., Hatzichristou, C., Wensink, J.,
from band B to band A); and Nigeria, 4.53 (no change Rosenberg, E., van Twillert, B., Stedema, J., &
in band). Meijer, D. (1997). Dimensions of national culture
Values: Finland, 3.91 (no change in band); Japan, as predictors of cross-national differences in
5.84 (no change in band); Morocco, 3.68 (no change subjective well-being. Personality and Individual
in band); and Taiwan, 2.91 (moves from band B to Differences, 23(1), 37–53.
band C). Baer, J. (1976). How to be an assertive (not
For a complete discussion of this procedure and aggressive) woman in life, in love and the job.
all response bias corrected scores, see Appendix B. New York: Signet.
2. As reported in Chapter 20 by Brodbeck, Bajdo, L. M., & Dickson, M. W. (2001). Perceptions
Hanges, Dickson, Gupta, and Dorfman, we found that of organizational culture and women’s advance-
all the cultural dimensions of organizational cultural ment in organizations: A cross-cultural exami-
values and practices significantly differed across soci- nation. Sex Roles, 45(5–6), 399–414.
eties. Although important, this prior analysis did not Barrick, M. R., & Mount, M. K. (1991). The Big
identify the particular aspect of societal differences that Five personality and job performance: A meta-
was related to organizational culture. In the present analysis. Personnel Psychology, 44, 1–26.
chapter, we found that societal and organizational Bjorklund, D. F., & Kipp, K. (1996). Parental investment
Assertiveness practices were not significantly related theory and gender differences in the evolution of
( p > .05). However, we found strong results for inhibition mechanisms. Psychological Bulletin,
societal and organizational Assertiveness values (R2 120, 163–188.
Total = 29.6%, R2 Societal = 92.6%, p < .01). As Bonta, B. D. (1997). Cooperation and competition in
discussed in Chapter 11 by Hanges, Dickson, and Sipe, peaceful societies. Psychological Bulletin, 121,
the R2 Total considers all levels of analysis (i.e., indi- 299–320.
vidual, organizational, and societal), whereas the R2 Booream, C. D., & Flowers, J. V. (1978). A procedural
Societal isolates the societal level portion of the depen- model for training of assertive behavior. In
dent variable and indicates the percentage of variance J. M. Whitely & J. V. Flowers (Eds.), Approaches
15-House.qxd 3/19/2004 4:51 PM Page 434
to assertion training (pp. 15–46). Monterey, CA: Hall, E. T. (1959). The silent language. Garden City,
Brooks/Cole. NY: Anchor Press.
Brenner, O. C., Tomkiewicz, J., & Schein, V. E. (1989). Heilman, M. E., Block, C. J., Martell, R. F., & Simon,
The relationship between sex role stereotypes M. C. (1989). Has anything changed? Current
and requisite management characteristics characterizations of men, women, and managers.
revisited. Academy of Management Journal, 32, Journal of Applied Psychology, 74, 935–942.
662–669. Hofstede, G. (1980). Culture’s consequences: Inter-
Brown, P., & Levinson, S. (1987). Politeness: Some national differences in work related values.
universals in language usage. Cambridge, UK: Beverly Hills, CA: Sage.
Cambridge University Press. Hofstede, G. (1991). Cultures and organizations:
Caspi, A., Elder, G. H., & Bem, D. J. (1988). Moving Software of the mind. London: HarperCollins.
away from the world: Life-course patterns of Hofstede, G. (1993). Cultural constraints in management
shy children. Development Psychology, 24, theories. Academy of Management Executive,
824–831. 7(1), 81–94.
Chinese Culture Connection. (1987). Chinese values Hofstede, G. (1998). Masculinity and femininity: The
and the search for culture-free dimensions of taboo dimension of national cultures. Thousand
culture. Journal of Cross-Cultural Psychology, Oaks, CA: Sage.
18(2), 143–164. Hofstede, G. (2001). Culture’s consequences:
Coltrane, S. (1988). Father-child relationships and the Comparing values, behaviors, institutions and
status of women: A cross-cultural study. American organizations across nations (2nd ed., rev.).
Journal of Sociology, 93, 1060–1095. Thousand Oaks, CA: Sage.
Crawford, M. (1988). Gender, age and the social Hollandsworth, J. G. (1977). Differentiating assertion
evaluation of assertion. Behavior Modification, and aggression: Some behavioral guidelines.
12, 549–564. Behavior Therapy, 8, 347–352.
Crawford, M. (1995). Talking difference: On gender Holtgraves, T. (1992). The linguistic realization of
and language. London: Sage face management: Implications for language pro-
Deal, T. E., & Kennedy, A. A. (1982). Corporate duction and comprehension, person perception
cultures: The rites and rituals of corporate life. and cross-cultural communication. Social
Reading, MA: Addison-Wesley. Psychology Quarterly, 55, 141–159.
Denison, D. R. (1996). What is the difference between Holtgraves, T. (1997). Styles of language use:
organizational culture and organizational cli- Individual and cultural variability in conversa-
mate? A native’s point of view on a decade of tional indirectness. Journal of Personality and
paradigm wars. Academy of Management Review, Social Psychology, 73, 624–637.
21(3), 619–654. Hoppe, M. H. (1998). Validating the masculinity/
Doney, P. M., Cannon, J. P., & Mullen, M. R. (1998). femininity dimension on elites from 19 countries.
Understanding the influence of national culture In G. H. Hofstede (Ed.), Masculinity and femi-
on the development of trust. Academy of Manage- ninity: The taboo dimension of national cultures
ment Review, 23(3), 601–620. (pp. 29–43). Thousand Oaks, CA: Sage.
Fagenson, E. A. (1990). Perceived masculine and House, R. J. (1971). Path-goal theory of leadership
feminine attributes examined as a function of indi- effectiveness. Administrative Science Quarterly,
viduals’ sex and level in the organizational power 16, 321–338.
hierarchy: A test of four theoretical perspectives. House, R. J. (1996). Path-goal theory of leadership:
Journal of Applied Psychology, 75, 204–211. Lessons, legacy and a reformulated theory.
Furnham, A. (1979). Assertiveness in three cultures: Leadership Quarterly 7(3), 323–352.
Multidimensionality and cultural differences. House, R. J., Hanges, P. J., Ruiz-Quintanilla,
Journal of Clinical Psychology, 35, 522–527. S. A., Dorfman, P. W., Javidan, M., Dickson, M.,
Goldberg, L. R. (1990). An alternative “description et al. (1999). Cultural influences on leadership
of personality”: The Big Five factor structure. and organizations: Project GLOBE. In W. Mobley
Journal of Personality and Social Psychology, (Ed.), Advances in global leadership (Vol. 1,
59, 1216–1229. pp. 171–233). Greenwich, CT: JAI.
15-House.qxd 3/19/2004 4:51 PM Page 435
Assertiveness– • –435
Howard, A., & Bray, D. W. (1988). Managerial lives McCrae, R. R., & Costa, P. T., Jr. (1997). Personality
in transition: Advancing age and changing trait structure as a human universal. American
times. New York: Guilford. Psychologist, 52, 509–516.
Howard, A., & Bray, D. W. (1994). Predictions of Melamed, T. (1996a). Career success: An assessment of
managerial success over time: Lessons from a gender specific model. Journal of Occupational
the management progress study. In K. E. Clark and Organizational Psychology, 69, 217–242.
& M. B. Clark (Eds.), Measures of leadership Melamed, T. (1996b). Validation of a stage model of
(pp. 113–130). West Orange, NJ: Leadership career success. Applied Psychology: An Inter-
Library of America. national Review, 45, 35–65.
Irani, K. D. (1986). Introduction: Modes of ratio- Miller, E. M. (1994). Paternal provisioning versus
nality. In M. Tamny & K. D. Irani (Eds.), mate seeking in human populations. Personality
Rationality in thought and action (pp. 1–18). and Individual Differences, 17, 227–255.
New York: Greenwood. Peabody, D. (1985). National characteristics.
Judge, T. A., Higgins, C. A., Thoresen, C. J., & Cambridge, UK: Cambridge University Press.
Barrick, M. R. (1999). The Big Five personality Rakos, R. F. (1991). Assertive behavior: Theory,
traits, general mental ability, and career success research and training. London: Routledge.
across the life span. Personnel Psychology, 53, Rathus, S. A. (1973). A 30-item schedule for assess-
621–652. ing assertive behavior. Behavior Therapy, 4,
Kale, S. H. (1991). Culture-specific marketing commu- 398–406.
nications: An analytical approach. International Rawls, D. J., & Rawls, J. R. (1968). Personality char-
Marketing Review, 8, 18–30. acteristics and personal history data of successful
Kale, S. H., & Barnes, J. W. (1992). Understanding and less successful executives. Psychological
the domain of cross-national buyer-seller interac- Reports, 23, 1032–1034.
tions. Journal of International Business Studies, Salter, A. (1949). Conditioned reflex therapy. New York:
23, 101–132. Farrar, Straus, & Giroux.
Kalin, R., Heusser, C., & Edmonds, J. (1982). Schein, E. H. (1992). Organizational culture and lead-
Cross-national equivalence of a sex-role ideol- ership (2nd ed.). San Francisco: Jossey-Bass.
ogy scale. Journal of Social Psychology, 116, Schein, V. E. (1973). The relationship between sex
141–142. role stereotypes and requisite management
Kelly, J. A., Kern, J. M., Kirkley, B. G., Patterson, characteristics. Journal of Applied Psychology,
J. N., & Keane, T. M. (1980). Reactions to 57, 330–335.
assertive versus unassertive behavior: Schein, V. E. (2001). A global look at psychological
Differential effects for males and females and barriers to women’s progress in management.
implications for assertiveness training. Behavior Journal of Social Issues, 57, 675–688.
Therapy, 11, 670–682. Scher, D., Nevo, B., & Beit-Hallahmi, B. (1979).
Kluckhohn, F., & Strodtbeck, F. L. (1961). Variations Beliefs about equal rights for men and women
in value orientations. Westport, CT: Greenwood. among Israeli and American students. Journal of
Kohn, A. (1986). No-contest: The case against com- Social Psychology, 109, 11–15.
petition. Boston: Houghton Mifflin. Schneider, S. C., & Barsoux, J. L. (1997). Managing
Lance, C. E., & James, L. R. (1999). A proportional across cultures. London: Prentice Hall Europe.
variance accounted for index for some cross- Schwartz, D. C. (1968). On the ecology of political
level and person-situation research designs. violence: “The long hot summer” as a hypothesis.
Organizational Research Methods, 2, 395–418. The American Behavioral Scientist, 11, 24–28.
Lange, A. J., & Jabuwkoski, P. (1976). Responsible Schwartz, S. H. (1994). Cultural dimensions of
assertion: A model for personal growth. values: Towards an understanding of national
Champaign, IL: Research Press. differences. In U. Kim, H. C. Triandis, C.
Loeber, R., & Hay, D. (1997). Key issues in the Kagitçibasi, S. C. Choi, & G. Yoon (Eds.),
development of aggression and violence from Individualism and collectivism: Theoretical and
childhood to early adulthood. Annual Review of methodological issues (pp. 85–119). Thousand
Psychology, 48, 371–410. Oaks, CA: Sage.
15-House.qxd 3/19/2004 4:51 PM Page 436
Segall, M. H., Dasen, P. R., Berry, J. W., & Van de Vliert, E. (1998). Gender role gaps, competi-
Poortinga, Y. H. (1990). Human behavior in global tiveness, and masculinity. In G. Hofstede (Ed.),
perspective: An introduction to cross-cultural Masculinity and femininity: The taboo dimension
psychology. New York: Pergamon. of national cultures (pp. 117–129). Thousand
Shoemaker, M., & Satterfield, D. O. (1977). Assertion Oaks, CA: Sage.
training: An identity crisis that’s coming on Van de Vliert, E., Schwartz, S. H., Huismans, S. E.,
strong. In R. E. Alberti (Ed.), Assertiveness: Hofstede, G., & Daan, S. (1999). Temperature,
Innovations, applications, issues (pp. 49–58). San cultural masculinity, and domestic political vio-
Luis Obispo, CA: Impact. lence: A cross-national study. Journal of Cross-
Smith, M. J. (1975). When I say no, I feel guilty. Cultural Psychology, 30, 291–314.
New York: Bantam. Watson, D., & Clark, L. A. (1997). Extraversion
Sue, D., Ino, S., & Sue, D. M. (1983). Nonassertiveness and its positive emotional core. In R. Hogan,
of Asian Americans: An inaccurate assumption? J. Johnson, & S. Briggs (Eds.), Handbook of
Journal of Counseling Psychology, 30, 581–588. personality psychology (pp. 767–793). San
Ting-Toomey, S. (1988). Intercultural conflict styles. Diego, CA: Academic Press.
In Y. Kim & W. Gudykunst (Eds.), Theories Williams, J. E., & Best, D. L. (1982). Measuring sex
in intercultural communication. Newbury Park, stereotypes: A thirty-nation study. Beverly Hills,
CA: Sage. CA: Sage.
Triandis, H. C. (1994). Cross-cultural industrial Williams, J. E., & Best, D. L. (1989). Sex and psyche:
and organizational psychology. In H. C. Triandis, Self-concept viewed cross-culturally. Newbury
Dunnette & Hough (Eds.), Handbook of indus- Park, CA: Sage.
trial and organizational psychology (2nd ed., Wolpe, J. (1958). Psychotherapy by reciprocal inhi-
Vol. 4, pp. 103–172). Palo Alto, CA: Consulting bition. Stanford, CA: Stanford University Press.
Psychologists Press. Wolpe, J. (1982). The practice of behavior therapy
Trompenaars, F., & Hampden-Turner, C. (1997). (3rd ed.). New York : Pergamon.
Riding the waves of culture: Understanding
cultural diversity in business (2nd ed.). London:
Nicholas-Brealey.
16-House.qxd 3/19/2004 5:00 PM Page 437
16
INDIVIDUALISM AND
COLLECTIVISM
MICHELE J. GELFAND
DHARM P. S. BHAWUK
DAVID J. BECHTOLD
AUTHORS’ NOTE: The first two authors thank their mentor, Harry Triandis, for introducing us to the con-
structs of individualism and collectivism and for his friendship, tutelage, and support throughout the years.
437
16-House.qxd 3/19/2004 5:00 PM Page 438
Lastly, we discuss the relationship of GLOBE’s Code of Hammurabi identified the need for
measures of individualism and collectivism with individuals to maintain positive relations with
perceptions of effective leadership at multiple others or else they would have to face heavy
levels of analysis. sanctions (Durant, 1935). The recognition of
individuals as being interdependent and as
having duties and obligations to other group
HISTORICAL OVERVIEW members are defining attributes of the cultural
construct that we now call collectivism.
Early Philosophers and It is important to note that the Code of
Hammurabi was not the only legal expression of
Individualism and Collectivism
the collectivist cultural construct in the ancient
Although the constructs of individualism and Middle East. Written codes of conduct that cen-
collectivism received much empirical attention tered on creating group standards of behavior
in the 20th century, they have been manifested were also part of the law of the Hebrews, most
in cultural institutions for thousands of years. In notably in the book of the Law of Moses
this section, we briefly review how the con- (Kagan, 1966). Unlike the Code of Hammurabi,
structs were represented in legal and religious the Ten Commandments were clerical laws.
institutions in ancient civilizations. Nevertheless, their purpose was also to establish
standards for individual behavior to protect the
Legal Institutions. In the ancient civilizations group, rather than to allow individual prefer-
of the Middle East and the Mediterranean, a ences to determine what is right and what is
review of the evolution of laws illustrates the wrong (Durant, 1935).
shifting emphases of individualism and collec- As cultures later developed, the more
tivism throughout the ages. Specifically, as individualistic notion of rational principles and
discussed below, attributes of cultures first individual rights became more prevalent within
changed from being highly individualistic to legal systems. In Athens, Greece, and in Rome
becoming collectivist, as individuals became the practice of presenting individual cases
members of tribes and, ultimately, nations. before other individuals or formally appointed
Further, as these nations became more complex judges was part of the procedural codes. In the
the system of laws evolved, emphasizing the Law of Cincius (204 BC), legal representation
need for rational systems of justice and individ- was viewed as the most effective way to present
ual rights, both of which are attributes of the facts of a particular case. Importantly, this
modern-day individualism (Triandis, 1995). system was seen as superior to interpretations of
The first stage in the evolution of law in the right and wrong based on codes of normative
Middle East was based on the notion of personal behavior alone, the latter of which was common
revenge. However, as these ancient tribes grew in earlier centuries and was more akin to
into nations, the need to protect the rights collectivism (Durant, 1939).
and safety of the group soon took precedence
over the preference of individuals to act on their Religious Institutions. The constructs of individ-
own behalf. This was codified in the laws of ualism and collectivism were also manifested in
Hammurabi, the king of Babylonia (1792–1750 religious institutions throughout the centuries.
BC), who is credited with establishing some of In the West, concerns with group identity
the world’s first written laws, wherein collective and in-group and out-group distinctions, both
concerns replaced preferences of individuals attributes of collectivism, can be seen in reli-
Unlike laws that exist in most countries gious philosophies and practices. The ancient
today, these laws did not protect the rights of the Hebrews’ religion was based on the strong eth-
individual, but rather described universal codes nic identity of the Jews (Durant, 1935) and was
of behavior and punishment. Among other predicated on the belief that the Jews were the
changes, the codes replaced the more individu- “chosen ones” of God. Other religious groups
alistic notion of equivalent retaliation with a also viewed their religions as a form of group
system of monetary fines. More generally, the identity as contrasted to other groups. For
16-House.qxd 3/19/2004 5:00 PM Page 439
Dewey (1930) distinguished what he referred to a theme that contrasts the extent to which
to as “old” individualism (which included people are autonomous individuals or embed-
the liberation from legal and religious restric- ded in their groups (Hofstede, 1980; Markus &
tions) from the “new” individualism (which Kitayama, 1991; Schwartz, 1994; Triandis,
focused on self-cultivation). Dumont (1986) 1989). Below, we review major empirical pro-
argued that individualism was a consequence jects on individualism and collectivism.
of Protestantism (i.e., humans do not have to go
to church to communicate with God), political Kluckhohn and Strodtbeck. Kluckhohn and
developments (emphasis on equality and lib- Strodtbeck (1961) conducted one of the first
erty), and economic developments (e.g., afflu- ethnographic comparative studies on value ori-
ence). In more recent years, philosophers have entations in five small communities with dis-
explored the possibility that there are positive tinct cultures in the Southwestern United States
attributes of both individualism and collec- of America. They proposed a five-part typology
tivism (Taylor, 1989). of universal human value orientations. Most
In sum, the constructs of individualism pertinent to our discussion, they argued that
and collectivism have been central in discus- social relationships might be lineal (people do
sions of political philosophy in the past few what authorities say), collateral (people do what
centuries. Next, we turn to the empirical assess- peers say), or individualistic (people do what
ment of individualism and collectivism at the they themselves think is right). Other dimen-
societal, organizational, and individual levels sions discussed by Kluckhohn and Strodtbeck
of analysis. included the relationship between humans and
nature, the nature of time, the nature of activi-
ties, and the nature of human beings.
OVERVIEW OF EMPIRICAL Importantly, in their view, these five princi-
RESEARCH ON INDIVIDUALISM ples varied from culture to culture, but only in
AND COLLECTIVISM their pattern; that is, the principles consisted of
component parts that were themselves univer-
sals. In support of their value theory, Kluckhohn
Individualism and
and Strodtbeck provided an analysis of these
Collectivism at the Societal Level
value orientations in five communities in the
Within the 20th century there was extensive United States, including a Texan homestead
discussion of the constructs of individualism community, a Mormon village, a Spanish-
and collectivism in sociology (Durkheim, 1933; American village, a decentralized Navaho
Parsons, 1949; Riesman, Denny, & Glazer, Indian band, and a highly centralized pueblo of
1950), anthropology (Kluckhohn, 1956; Mead, Zuni. As such, they were of the first to advance
1961; Redfield, 1956), and psychology (Chinese theory and analysis of the individualism and
Culture Connection, 1987; Hofstede, 1980, collectivism constructs in the social sciences.
2001; Markus & Kitayama, 1991; Schwartz,
1994; Triandis, 1995). Across each of these dis- Hofstede. Hofstede (1980) used data from a
ciplines, scholars have been concerned with the morale survey of IBM employees from 40
nature of the relationship between the individual countries to uncover cultural dimensions. On
and the group, which is broadly referred to as the basis of a factor analysis of the sum of all of
individualism and collectivism. This theme has the responses in each culture, Hofstede (1980)
also been referred to as self-emphasis and col- named one of the factors individualism versus
lectivity (Parsons, 1949), Gesellschaft and collectivism, which he defined as follows:
Gemeinschaft (Toennies, 1957), mechanical and
organic solidarity (Durkheim, 1933), individu- Individualism pertains to societies in which the
alism and collaterality (Kluckhohn & Strodt- ties between individuals are loose; everyone is
beck, 1961), and agency and community (Bakan, expected to look after himself or herself and his or
1966). Although there are subtle differences her immediate family. Collectivism as its opposite
in the meanings of these terms, they all relate pertains to societies in which people from birth
16-House.qxd 3/19/2004 5:00 PM Page 441
onwards are integrated into strong, cohesive literature in management. Further, it also
in-groups, which throughout people’s lifetime provided a ranking of nations in many regions
continue to protect them in exchange for unques- of the world on the constructs, which have
tioning loyalty. (51) since been used for descriptive and prescriptive
purposes.
This bipolar factor was derived from a post Nevertheless, numerous questions were
hoc analysis and consisted of six items that raised regarding the reliability, validity, gener-
related to differences in preferences for work alizability, and robustness of Hofstede’s find-
goals. The individualistic pole of the dimension ings (Erez & Earley, 1993; Schwartz, 1994;
was associated with preferences for sufficient Smith & Schwartz, 1997). Scholars have also
time for personal or family life, considerable questioned whether Hofstede’s (1980) findings
freedom on the job, and having challenging were generalizable and robust. First, questions
work. By contrast, the collectivist pole of the were raised regarding the validity of the specific
dimension was associated with preferences for items that were derived post hoc in Hofstede’s
training opportunities, having good physical (1980) culture-level factor analysis. For example,
work conditions, and being able to use skills on it is not clear why concerns with training oppor-
the job. Hofstede (2001) argued that the positive tunities and good physical work conditions were
loading on Personal Time (.86), Freedom (.49), core aspects of collectivism. In this respect, the
and Challenge (.46) “stress the actors’ indepen- collectivistic pole of this scale lacks face valid-
dence from the organization,” whereas the neg- ity. For example, on the basis of the content of
ative loading on Use of Skills (−.63), Physical the items, the factor could well have been
Conditions (−.69), and Training (−.82) “stress labeled “individual freedom versus individual
what the organization does for the individual” development.” Alternatively, the factors could
(p. 214). Thus, for Hofstede, the positive load- be seen as similar to Herzberg and colleagues’
ing represented individualism and the negative (Herzberg, Mausner, & Snyderman, 1959)
loading reflected collectivism, a point to which intrinsic (work-related) and extrinsic (nonwork-
we will return below. It is also important to note related) factors (see Hofstede, 2001). Second,
that his analysis reflects factor loadings across questions regarding measurement also abounded:
societies. Indeed, this scale did not have high Would the same dimensions be found using a
internal consistency at the individual level of sample that was not from IBM? Are individual-
analysis. ism and power distance conceptually and empir-
Hofstede’s results (1980) showed that the ically distinct, especially given the fact that they
United States, Australia, England, and the are so highly negatively related (see Hofstede,
Netherlands were the most individualistic cul- 1980)? Do the measurements on which the
tures he studied, whereas most East Asian, dimensions were based have different meanings
African, and Latin American cultures were col- across cultures? Would the results be the same
lectivist. In Hofstede’s analysis, individualism if the measures were used in non-Western
was highly inversely correlated with power dis- contexts? And finally, would the results repli-
tance (r = .67; Hofstede, 1980, p. 221). In addi- cate across time and with different samples?
tion, Hofstede correlated scores on individualism Studies that have addressed these issues are
with a host of other social indicators (some discussed below.
of which are discussed below), providing fur- Furthermore, questions regarding the appro-
ther post hoc convergent validity for this factor. priate levels of analysis for the constructs have
The importance of Hofstede’s (1980) study been raised since the original publication. As
for cross-cultural research cannot be underesti- later noted by Hofstede (1994, 2001) and others
mated, as it was the first large-scale empirical (Triandis, 2002), despite an explicit discussion
project to put these abstract constructs on the of levels of analysis in his 1980 study, confu-
empirical map. As such, it provided a “concep- sion about the constructs has abounded since his
tual roof” under which existing studies could original work, as some authors have applied his
fall, as compared to the atheoretical stance that bipolar dimensions to distinguish the personali-
had previously characterized the cross-cultural ties of individuals within societies. Indeed, to
16-House.qxd 3/19/2004 5:00 PM Page 442
clarify this confusion, Hofstede (1994) later the integration factor related to Individualism
reiterated that his analysis was, in his words, included tolerance and trustworthiness, whereas
the pole related to Collectivism included filial
sociological and not psychological. It does not piety, chastity in women, and patriotism.
compare different personalities, but different soci- Nevertheless, some items loading on the posi-
etal contexts within which children grow up and tive pole of Integration (e.g., noncompetitive-
develop their personalities. It is not about individ- ness, contentedness with one’s position in life,
uals, but about the constraints within which, in being conservative)—and thus correlated with
different countries, a psychology of relatedness Individualism—are hard to explain. Scores of
should be developed. (Hofstede, 1994, p. x) another factor, namely Human Heartedness,
were not correlated with Individualism, but
As discussed in later sections below, the were negatively related to Hofstede’s masculin-
nature of the construct may vary at the individ- ity index. Again, the pattern of this correlation is
ual level of analysis, and Hofstede (1997) and counterintuitive, as the pole of this dimension
others have warned against confusing such that was related to high masculinity (a focus on
issues in theory and research on individualism success striving and materialism) included
and collectivism, as well as other dimensions of items such as patience and kindness, which one
cultures. would have expected to correlate with the femi-
nine pole of Hofstede’s dimension.
The Chinese Culture Connection. As discussed in In sum, data from the Chinese Culture
Chapter 4 of this volume, to address the possibil- Connection provide mixed results regarding
ity that Hofstede’s (1980) questionnaire items the generalizability of Hofstede’s (1980) indi-
were biased by Western values and assumptions, vidualism to the Eastern value survey that was
Bond and members of the Chinese Culture constructed. Although moral discipline was cor-
Connection (1987) developed a Chinese Value related with individualism in the expected direc-
Survey that was based on Eastern, rather than tion, other findings do not provide evidence
Western, values. Specifically, they used a research of convergent validity for the individualism
strategy that consisted of administering a ques- construct. It is important to note, however, that
tionnaire developed by Chinese scholars to sam- the Chinese Culture Connection survey was
ples of 50 male and 50 female university students composed of very different items and was
in 23 national cultures. A culture-level analysis conducted within only 23 countries around the
revealed four dimensions of values: Integration, world. For more information about the mas-
Human-Heartedness, Moral Discipline, and culinity scale see Chapter 14 by Emrich,
Confucian Work Dynamism. Chinese Culture Denmark, and Den Hartog.
Connection (1987) correlated nation-level
scores on these dimensions with scores from Triandis and Colleagues. Triandis and his col-
Hofstede’s 1980 study. Not surprisingly, scores leagues also provided converging evidence for
on moral discipline, which includes a focus on the existence of the individualistic–collectivistic
moderation, purity, and having few desires, was (IC) constructs at the culture level and have
negatively correlated with Hofstede’s individu- greatly elaborated on the meaning of the con-
alism index (r = −.54). structs. First, to probe the meaning of the con-
Somewhat more surprisingly, the Individ- structs further, Triandis and colleagues polled a
ualism scores positively correlated with Power sample of social scientists in many nations as
Distance (r = .55) in the Chinese Culture to their understanding of individualism and
Connection results. Scores on Integration were collectivism (Hui & Triandis, 1986). They
also positively associated with Individualism found general consensus and, based on their
(r = .65) and negatively correlated with Power results, developed more items to further investi-
Distance (r = −.58). At first, this seems counter- gate the constructs. In a later series of studies,
intuitive, yet an examination of the poles of this Triandis and colleagues (1986) examined the
dimension helps to elucidate why this pattern structure of these items at the culture level in
may have been found. Specifically, the pole of nine countries. Their culture-level analysis
16-House.qxd 3/19/2004 5:00 PM Page 443
revealed four factors, two of which were cultures, they are primarily attitudes, values,
reflective of individualism (Self-Reliance with beliefs, personal needs, perceived rights, and
Hedonism and Separation from In-Groups) and contracts (Bontempo & Rivero, 1992; Davidson,
two of which were reflective of collectivism Jaccard, Triandis, Morales, & Diaz-Guerrero,
(Family Integrity and Interdependence with 1976; Miller, 1994).
Sociability). They found Hofstede’s (1980)
4. Emphasis on relatedness versus rationality:
nation scores on individualism and collectivism
In collectivist cultures, people tend to empha-
were correlated only with scores on Family
size unconditional relatedness within groups,
Integrity, which is to be expected as a manifes-
whereas in individualistic cultures people tend
tation of collectivism (r = .78). More recently,
to emphasize rationality. Relatedness refers
Triandis, McCusker, and colleagues (1993)
to giving priority to relationships and taking
used the Leung-Bond procedure and extracted
into account the needs of others, even if such
multiple universal and culture-specific indepen-
relationships are not advantageous. Rationality
dent dimensions of individualism and collec-
refers to the careful computation of the costs
tivism across cultures. Thus, unlike previous
and benefits of relationships (Kim et al.,
analyses, Triandis and colleagues have found
1994).1
evidence of the multidimensionality of the con-
structs at the culture level while at the same time Apart from similarity on these core attrib-
confirming the overlap of some of the dimen- utes, Triandis suggested that individualistic
sions with Hofstede’s (1980) original work. and collectivist cultures also differ on additional
Consistent with this perspective, Triandis culture-specific elements of the constructs. In
argued that there are a large number of collec- this respect, he suggested that the constructs
tivist and individualist patterns (Triandis, 1994, must be defined polythetically as is done in
1995) and that not all collectivist cultures are other sciences (Triandis, 1994). For instance, in
identical, nor are all individualist cultures zoology each phylum contains many combina-
identical. Within this complexity, however, he tions of attributes, but only a few attributes are
argued that there are at least four defining fea- characteristic of all of the species within that
tures of the constructs, and that cultures are phylum. Thus, the defining feature of the cate-
probably similar on the main defining features gory birds may be wings and feathers, whereas
of the constructs. The four defining features are yellow beaks and carnivorous are attributes that
differentiate among various species of birds.
1. Definition of the self: In collectivist cultures, Accordingly, collectivism is specified by some
the self is generally viewed as interdependent common defining attributes, but collectivism in
with others, which is accompanied by the Japan, for example, would require several more
sharing of resources. In individualistic cultures, culture-specific attributes, and collectivism in
the self is generally viewed as autonomous an Israeli kibbutz would require some other
and independent of groups, and decisions combination of culture-specific attributes. Thus,
regarding whether or not to share resources are Triandis was one of the first researchers to argue
made individually (Markus & Kitayama, 1991; that individualism and collectivism should not
Reykowski, 1994). be seen in pure dichotomies, but rather require
2. The structure of goals: In collectivist cultures, additional theoretical refinement to understand
goals tend to be compatible with in-group goals. specific cultures. We will return to these differ-
In individualistic cultures, individual goals tend ences later when discussing the influence of
not to be correlated with in-group goals individualism and collectivism within organiza-
(Schwartz, 1992, 1994; Triandis, 1988, 1990; tional contexts.
Wagner & Moch, 1986).
Horizontal and Vertical Collectivism. Triandis
3. Emphasis on duties and obligations versus (1994, 1995) proposed that the most important
personal preferences: In collectivist cultures, the attributes that distinguish among different kinds
determinants of social behavior are primarily of individualism and collectivism are the rela-
duties and obligations, whereas in individualist tive emphases on horizontal and vertical social
16-House.qxd 3/19/2004 5:00 PM Page 444
relationships. Horizontal patterns assume that individual level (referred to as idiocentrism and
one self is more or less like every other self. By allocentrism, respectively) that will be described
contrast, vertical patterns consist of hierarchies, at length in sections below.
and one self is rather different from other selves.
The ways in which these relative emphases Schwartz and Colleagues. In another landmark
combine with individualism and collectivism study, Schwartz and colleagues (Schwartz,
produce four distinct patterns, namely horizon- 1992, 1994; Schwartz & Bilsky, 1990) exam-
tal individualism (HI), vertical individualism ined the structure of values among more than
(VI), horizontal collectivism (HC), and vertical 44,000 teachers and students in 54 countries.
collectivism (VC; Triandis & Gelfand, 1998). Participants in each of the countries were asked
More specifically, he argued that in horizon- to respond to a 56-item value instrument.
tal individualistic (HI) cultures, such as Sweden Notably, Schwartz and his colleagues developed
and Australia, people want to be unique and dis- an a priori theory regarding the dynamics of
tinct from groups and are high in self-reliance, value differences across nations. Specifically,
but they are not especially interested in becom- they argued, and confirmed, that cultural value
ing distinguished or in having high status. In dimensions would reflect core solutions that
vertical individualistic (VI) cultures, such as the emerge as nations attempt to cope with societal
United States and France, people often want problems. Such problems include the following:
to become distinguished and acquire status, and
they do this in individual competitions with 1. Relations between individual and groups: The
others. In horizontal collectivist (HC) cultures, extent to which people are autonomous
such as the Israeli kibbutz and Eskimo cultures, as opposed to being embedded in groups,
people see themselves as similar to others and reflecting an emphasis on autonomy versus
emphasize common goals with others, interde- conservation.
pendence, and sociability, but they do not sub-
2. Assuring responsible social behavior: How to
mit easily to authorities. In vertical collectivist
motivate people to consider others’ welfare
cultures (VC), such as India and China, people
and coordinate with them, reflecting an empha-
emphasize the integrity of the in-group, are will-
sis on hierarchy versus egalitarianism.
ing to sacrifice their personal goals for the sake
of in-group goals, and support competitions 3. The role of humankind in the natural and
between their in-groups and out-groups. If in- social world: Is it more important to submit, to
group authorities want them to act in ways that fit, or to exploit the environment, reflecting an
benefit the in-group but are extremely distaste- emphasis on mastery versus harmony.
ful to them, they submit to the will of these
authorities.2 Although this theory has not yet The results of a multidimensional scaling
been tested at the culture level, it has received study provided evidence of these cultural value
support at the individual level (Chen, Meindl, & dimensions, which were based on a circumplex
Hunt, 1997; Singelis, Triandis, Bhawuk, & of seven basic value types. Most pertinent to
Gelfand, 1995; Triandis & Bhawuk, 1997; this discussion, the first dimension in the cir-
Triandis, Chen, & Chan, 1998; Triandis & cumplex was closely related to individualism
Gelfand, 1998). and collectivism. At one pole were cultural
In sum, Triandis and his colleagues have fur- values that Schwartz (1994) labeled conservatism,
ther refined the constructs of individualism and which he defined as
collectivism and have elaborated on the notion
that cultures can be further differentiated within those values likely to be important in societies
this broad dimension. In addition, Triandis and based on close-knit harmonious relations, in
colleagues (Hui, 1988; Triandis, Bhawuk, Iwao, which the interests of the person are not viewed as
& Sinha, 1995; Triandis, Bontempo, Villareal, distinct from those of the group. All of these
Asai, & Lucca, 1988) have done considerable values emphasize maintenance of the status quo,
theoretical and empirical work on individualism propriety, and avoidance of actions or inclinations
and collectivism as a personality attribute at the of individuals that might disturb the traditional
16-House.qxd 3/19/2004 5:00 PM Page 445
order. These are sociocentric values, appropriate used sound measurement techniques and was
in settings where the self lacks autonomous concerned with equivalence of measurement
significance but has meaning as part of the collecti- issues, a problem that has been associated with
vity (Miller, 1984). Cultures that emphasize Con- previous research (Hofstede, 1980) on individu-
servatism values are primarily concerned with alism and collectivism.
security, conformity, and tradition. (p. 101)
Smith, Dugan, and Trompenaars. Lastly, one
The other end of this dimension included val- other large-scale societal level study on individ-
ues related to what Schwartz labeled intellectual ualism and collectivism was conducted by
and affective autonomy, which were defined as Smith, Dugan, and Trompenaars (1996). This
study is based on Trompenaars’s (1985, 1993)
those values likely to be important in societies that value survey of more than 11,000 employees
view the person as an autonomous entity entitled from 58 countries. His original value question-
to pursue his or her individual interests and desires. naire was designed to capture five dimensions
Two related aspects of autonomy appear to be proposed by Parsons and Shils (1951), namely,
distinguishable: a more intellectual emphasis on universalism and particularism, achievement
self-direction and a more affective emphasis on and ascription, individualism and collectivism,
stimulation and hedonism.(p. 102) affectivity and neutrality, and specificity and
diffuseness. He also measured two other dimen-
Schwartz (1994) also predicted and con- sions: a personality variable, internal and exter-
firmed the relationship of these values to other nal locus of control (Rotter, 1966); and different
values in the circumplex. For example, as pre- orientations to time (Hall, 1959). Items used in
dicted, autonomy values were correlated with the questionnaire presented scenarios and asked
egalitarian commitment and with mastery respondents how they would act in a given
values, whereas conservation values were situation.
correlated with hierarchy values. As expected, Smith and colleagues (1996) analyzed this
autonomy and conservation were negatively database and found that the modal values of
correlated (r = −.89). Further analysis revealed employees varied substantially across cultures.
that nation-level scores on Schwartz’s values Using multidimensional scaling, they found
were also related to Hofstede’s 1980 nation- two interpretable dimensions, conservatism
level scores. Specifically, Hofstede’s individu- versus egalitarian commitment (label adopted
alism dimension was positively correlated with from Schwartz discussed above) and utilitar-
Schwartz’s autonomy value type (intellectual ian involvement versus loyal involvement.
and affective) and negatively correlated with the These dimensions were significantly related to
conservation value type. In addition, Hofstede’s Hofstede’s (1980) individualism and power
individualism dimension was also correlated distance dimensions. For example, collectivism
with the egalitarian commitment value type, was associated with a preference for closer
suggesting a close relationship between auton- work relations and higher involvement with
omy and egalitarianism. one’s company. Collectivism was also associ-
Schwartz and colleagues’ research is notable ated with particularism and a preference for
in several ways. First, it is based on a universal ascription over achievement. Further they found
theory of human values. It not only predicts a that individualism was associated with moder-
priori the nature of value dimensions, such as nity, which was measured by per capita
autonomy (individualism) and conservation gross national product, literacy rate, and life
(collectivism), but also describes the dynamics expectancy.
among these values and other values in the cir-
cumplex. Second, the project has been generally
replicated at both the culture and individual
Summary
levels of analysis, a point to which we will There has been a long tradition of research
return in the section on individual-level individ- on individualism and collectivism at the societal
ualism and collectivism. Finally, Schwartz level. Although different scholars have used
16-House.qxd 3/19/2004 5:00 PM Page 446
different methods, samples, and sometimes ways in which individualistic and collectivist
different terminology, there is a general similar- organizational cultures are different.
ity in the nature of the constructs at the societal In general, organizations that have individu-
level. Disagreement exists on the degree to alistic cultures would have members who con-
which the constructs are broad versus specific, sider themselves as largely independent of the
which often results in differences in expecta- organization. Employees would also assume
tions regarding whether the constructs are multi- that they are hired because of their unique skills
dimensional (Triandis, 1995) or unidimensional and abilities, rather than because of their rela-
(Hofstede, 1980; Schwartz, 1994) at the societal tionships or social background. In such cul-
level of analysis. In addition, disagreement tures, members would expect the organization
exists regarding how closely individualism and to offer them something they need, and would
collectivism are related to issues of power and be willing to leave the organization if their
hierarchy, with Schwartz and Hofstede arguing needs or goals were better served elsewhere.
they are highly related, and Triandis and col- More generally, personal attitudes would be
leagues arguing that measures of such con- important determinants of organizational
structs may be orthogonal. At the same time, the behaviors in individualistic organizational cul-
literature has several omissions, which are the tures. Furthermore, managers would assume
focus of the GLOBE study. First, most scholars that they need to compensate employees in a
have focused on individualism and collectivism manner that is rationally consistent with their
as value dimensions and, as will be discussed in capabilities and performance, and would
later sections, Project GLOBE adds to this tra- believe that individuals’ hard work and
dition by examining both values and practices successful performance is instrumental for
at the societal level. Second, much of the attaining rewards. As an example, in the orga-
research on individualism and collectivism at nizations within the United States the need for
the societal level has used items framed at the rational exchanges between an organization
individual level of analysis. However, from a and its members begins with the process of
levels-of-analysis perspective this is problem- recruitment and selection and continues until
atic, as discussed by Hanges and Dickson in employees are terminated. Throughout all of
Chapter 8 of this volume. GLOBE adds to the these activities, an underlying notion of fair-
literature by developing items that explicitly ness, which is based on the notion of equity,
refer to the particular levels of analysis. Finally, must prevail. Indeed, in the United States, orga-
GLOBE adds to this endeavor by taking a mul- nizations are accountable to external regulatory
tilevel approach to individualism and collec- agencies, such as federal and state govern-
tivism. In this spirit, we next turn to the nature ments, that define the rational procedures that
of the constructs at the organizational level of ensure fairness in organizations.
analysis. By contrast, organizations that have collec-
tivist cultures would have members who view
Individualism and Collectivism themselves as highly interdependent with the
organization. Generally speaking, the sharing of
at the Organizational Level
employees’ identity with the organization
Organizational culture has been a central would be so strong that the organization would
research issue in the organizational sciences become a part of members’ self-identity.
over the past few decades. At this level of Employees would assume that their relation-
analysis, culture is construed as organizational ships and duties and obligations, not merely
members’ shared beliefs and assumptions their unique attributes, are central in employ-
(Aycan et al., 2000; Schein, 1992). Although ment decisions. Further, managers would
there has not been much empirical research on assume that employees are willing to make per-
individualism and collectivism at the organiza- sonal sacrifices in their goals and desires in
tional level of analysis, theoretically speaking, order to fulfill obligations toward others in the
we may draw on Triandis’s (1995) four soci- organization. In return for their personal sacri-
etal defining attributes to help illuminate the fices, employees would expect the organization
16-House.qxd 3/19/2004 5:00 PM Page 447
would stand by their side even in hard economic organizational cultures can be differentiated on
times. Employees would also expect to be com- the degree to which they are collectivist versus
pensated in a manner that provides the greatest individualistic. They defined collectivist organi-
benefit to all members in the group, rather than zational cultures as those that emphasize shared
on a purely individual basis. In addition, objectives, interchangeable interests, and
employees would be less focused on their own commonalties among members. By contrast,
attitudes toward the job and more focused on individualistic organizational cultures are those
their duties and obligations to the organization. that emphasize individuals’ unique attributes
Thus, employees would view the nature of their and emphasize differences among employees.
relationship with the organization as one that is Chatman and colleagues argued that organiza-
less a matter of rational exchanges and more a tional membership is a very salient category for
matter of long-term relational exchanges. members of collectivist organizational cultures,
Some support for these tenets can be found as compared to individualistic organizational
in research at the organizational level of analy- cultures.
sis. In a large-scale study of organizations in Importantly, although organizational indi-
Denmark and the Netherlands, Hofstede, vidualism and collectivism is distinct from
Neuijen, Ohayv, and Sanders (1991) found two societal individualism and collectivism, the
dimensions that differentiated organizations two levels are expected to be interrelated.
that are relevant to the current discussion. The According to the model of culture-fit (Kanungo
first dimension was the degree to which organi- & Jaeger, 1990), societal-level culture affects
zations are parochial versus professional. In organizational work culture by shaping shared
parochial organizations, employees strongly managerial assumptions about the nature of
identify with their organization and believe that employees and how the organization needs to
organizational norms apply to their behavior at be structured for such employees. In addition,
home as well as on the job. They also believe Kanungo, Aycan, and Sinha (1999) also argue
that their social background is an important that managerial assumptions about the nature
determinant of hiring practices. By contrast, of the task are also influenced by institutional
employees in professional organizations iden- characteristics, such as ownership status (public
tify primarily with their job and consider their vs. private), industry (service vs. manufac-
home-life to be private. In these organizational turing), and resource availability (human and
cultures, employees believe that job compe- technological resources). Thus, according to
tence, not social background, is an important Kanungo and his colleagues, internal organiza-
determinant of hiring practices.3 The other rel- tional culture is shaped by the larger societal
evant dimension was labeled employee oriented context in which it is embedded.
versus job oriented. In employee-oriented orga- In some indirect support of some of these
nizations, organizations take responsibility for tenets, Aycan and colleagues (2000) found that
employee welfare, important decisions are across numerous countries, managers who per-
made by groups, and people feel their personal ceived the sociocultural context to be high on
problems are to be solved by the organization. collectivism (i.e., high on loyalty toward their
By contrast, in job-oriented organizations community) were likely to assume that employ-
people perceive that the organization is inter- ees within their organizations had high obliga-
ested only in the work they perform and not in tions toward others. This included the notion
their personal or family welfare. In such orga- that individuals should act in accordance with
nizations, employees report that decisions tend group needs and should compromise their own
to be made by individuals, rather than groups. wishes for the benefit of helping the group (see
Although neither of these dimensions were also Aycan, Kanungo, & Sinha, 1999).
labeled as collectivist versus individualistic, Along the same lines, Kim (1994) argued that
they clearly have overlap with components of in many Asian cultures organizations are erected
the constructs. in ways to sustain societal values and norms. As
In a more recent study, Chatman, Polzer, compared to individualistic cultures, which con-
Barsade, and Neale (1998) explicitly argued that struct institutions to reflect the importance of
16-House.qxd 3/19/2004 5:00 PM Page 448
individual rights, collectivist cultures construct and, as such, the nature of the dimensions can
institutions as extensions of the family to reflect vary at different levels of analysis.
the importance of the fulfillment of obligations Triandis and his colleagues were among the
(Kim, 1994). Similarly, Kashima and Callan first to investigate the constructs at the individ-
(1994) argued that in Japan, managers explicitly ual level of analysis. To clearly differentiate
design organizations through the use of meta- their discussions from societal-level individual-
phors that are familiar in the larger cultural con- ism and collectivism they adapted different
text (i.e., the Japanese ie, or “household”). In this terminology at this level, namely idiocentrism
way, organizations consciously create collec- and allocentrism, respectively (Triandis, Leung,
tivist organizational practices to be consonant Villareal, & Clack, 1985). This terminology
with prevailing sociocultural contexts. In Japan, allows for the fact that there can be idiocentrics
historically this has included the development of within collectivist cultures and allocentrics
long-term relationships with employees from within individualistic cultures. Since this study
recruitment to retirement. It has also included was conducted, research at this level has
family-like practices such as morning exercises, demonstrated that in the United States, idiocen-
singing of company songs on special occasions, trism and allocentrism are orthogonal at the
and assistance in helping employees find individual level (Gelfand et al., 1996), and indi-
spouses. Within such organizations, employees viduals have access to both idiocentric elements
make sacrifices for the company by putting in and allocentric elements (Trafimow, Triandis, &
service hours or unclaimed overtime hours Goto, 1991).4
(Kashima & Callan, 1994). Several scholars have found evidence for such
Although there is not as much research on individual differences within numerous nations.
individualism and collectivism at the organiza- Hui (1988) devised the individualism and collec-
tional level, there is increasing evidence that the tivism scale (INDCOL) and used it to reliably
constructs operate at this level of analysis and assess an individual’s level of individualism and
that they are influenced by the larger sociocul- collectivism in the Hong Kong Chinese context.
tural context. Research by GLOBE, discussed in Later studies at the individual level illustrated
later sections, adds to this endeavor in several that there are similar factor structures of individ-
ways. First, GLOBE explicitly developed mea- ualism and collectivism within the United States,
sures to assess individualism and collectivism at Japan, and Puerto Rico (Triandis et al., 1988).
the organizational level of analysis and used Likewise, Schwartz and Bilsky (1990) found that
organizational referents in its scales. Second, dimensions of values at the individual level were
GLOBE explicitly examined the relationship similar across many nations. Similar to the cul-
between societal and organizational individual- ture level, they found that values at the individual
ism and collectivism, furthering a multilevel level are organized according to openness to
perspective on the constructs. change versus conservation dimension. This
dimension was found in more than 90% of their
97 samples (Smith & Schwartz, 1997). Providing
Individualism and Collectivism
further converging evidence of the constructs at
at the Individual Level the individual level, Oishi, Schimmack, Diener,
In addition to examining individualism and and Suh (1998) recently found correlations
collectivism and related constructs at the societal between Triandis’s scales of idiocentrism and
and organizational levels, many scholars have allocentrism and Schwartz’s value measures. For
examined such distinctions at the individual example, vertical allocentrism was positively
level. In contrast to dimensions that differentiate correlated with tradition and conformity values,
nations or cultures, research at the individual vertical idiocentrism was positively corre-
level of analysis capitalizes on intracultural vari- lated with power and achievement, horizontal
ability and examines dimensions that are derived allocentrism was positively correlated with
from individual differences. As Triandis (1995) benevolence, and horizontal idiocentrism was
and Smith and Schwartz (1997) note, these positively correlated with achievement and
levels of analysis are statistically independent self-direction.
16-House.qxd 3/19/2004 5:00 PM Page 449
In more recent years, there has been an In addition, Smith and Schwartz (1997)
explosion of measures that have been developed review data that illustrate numerous social
to assess individualism and collectivism at background differences that are associated
the individual level (e.g., Bierbrauer, Meyer, & with individualism and collectivism, including
Wolfradt, 1994; Oyserman, 1993; Singelis, age, education, and socioeconomic status. Speci-
1994; Wagner & Moch, 1986; Yamaguchi, fically, younger and more educated individuals
1994). Because of this proliferation of mea- tend to be more individualistic than older and less
sures, there is some confusion at this level of educated individuals across numerous nations
analysis, as scholars develop measures that (Smith & Schwartz, 1997). With respect to social
assess different aspects of the construct, using class, research on individualism and collectivism
different methodologies, and often do not within cultures has generally been consistent
make reference to other measurements in the with Hofstede’s original findings at the societal
literature. On the other hand, such a varied level (discussed above): High socioeconomic
approach to measurement at this level of analy- classes tend to be more individualistic, whereas
sis is perhaps not surprising. As Triandis (1995) lower socioeconomic classes tend to be more
has warned, the constructs are highly broad, collectivist (Marshall, 1997; Triandis, 1994).
multidimensional, and can take on different cul- Interestingly, Marshall found that social class
ture-specific (or emic) attributes in different accounted for more variance in individualism
cultural contexts. Nevertheless, in the future it and collectivism than country of origin in a
would be useful to conduct construct validation comparison of Indonesia and New Zealand.
studies that include numerous measures cited in Finally, Huismans and Schwartz (1992) exam-
the literature. ined the relationship between values and overall
Despite this confusion, research does illus- religiousness, and found that some components
trate numerous demographic and social corre- of Schwartz’s dimension of conservation (i.e.,
lates of individualism and collectivism within tradition) are correlated with overall religiosity
cultures. Comparing across urban and rural and, further, that this correlation was even
locales within Estonia, Realo, Allik, and Vadi stronger among Israeli Jews.
(1997) found that collectivism is higher in rural
locations and individualism is higher in urban
locations, a finding that has also been replicated OVERALL SUMMARY
in Sri Lanka (Freeman, 1997). With respect to
race, in the United States, as compared to The constructs of individualism and collec-
European Americans, persons of color have tivism have received attention at numerous
scored higher on collectivism (defined as an ori- levels of analysis. Not surprisingly, as with
entation toward the welfare of one’s larger com- other areas within psychology and organi-
munity) and familism (defined as an orientation zational behavior, there is often confusion
toward the welfare of one’s immediate family; regarding levels of analysis (see also Klein,
Gaines et al., 1997; but see Osyerman et al., Dansereau, & Hall, 1994). First, there is the
2002, for a contrasting analysis). With respect question of whether the dimensions are similar
to gender, Kashima et al. (1995) found no dif- at different levels of analysis. As stated previ-
ference between men and women in Korea, ously, because such analyses are statistically
Japan, Australia, mainland United States, and independent, they are not necessarily similar. At
Hawaii on individualism (i.e., independence and the same time, it is also likely that there is con-
agency) and collectivism (i.e., concern for wel- ceptual similarity across levels of analysis for
fare of collectives). Gender differences were the constructs of individualism and collectivism
found, however, for a separate construct: (Schwartz, 1994). Second, there is the question
Emotional relatedness. More recent research of when it is appropriate to use a culture level of
also suggests that, at least within the United analysis versus an organizational or individual
States, men may be more likely than women level of analysis in cross-cultural research.
to define the self in terms of collectivities Schwartz argues cogently that this depends on
(Gabriel & Gardner, 1999). the type of question asked:
16-House.qxd 3/19/2004 5:00 PM Page 450
(in preparation) conclude that claiming that cultures focus on the well-being of the
collectivism is not suited for economic develop- collective, which requires less of an emphasis
ment is rather myopic and misleading. on time. In support of this, they found that soci-
Consistent with this analysis, if one exam- etal collectivism was negatively associated with
ines economic indicators besides gross national walking speed, the speed of postal requests, and
product (GNP), it appears that collectivism, with the accuracy of clocks across 31 countries.
not individualism, is more predictive of wealth. However, these relationships were nonsignifi-
For example, data from the Chinese Culture cant after the effects of economic well-being
Connection (1987) illustrate that across 22 (i.e., gross domestic product) were controlled,
Asian nations Confucian Dynamism scores are illustrating that economic factors play an impor-
highly correlated with average gross national tant role in pace of life.
growth (r = .70, p < .001). Likewise, Schwartz
(1994) reported that larger household sizes,
Health
which were also correlated with scores on his
conservation value dimension, also predicted Another indication of the well-being of nations
national wealth. relates to the health of its citizens. Research has
Lastly, adding further complexity to this illustrated that whether a culture is individualistic
picture, Schwartz (1994) has argued that the or collectivist has some health related correlates.
association between individualism and wealth Triandis et al. (1988) found that heart-attack rates
has changed over time since Hofstede’s (1980) were lower in collectivist cultures than individu-
original analysis. Specifically, Schwartz (1994) alistic cultures. Specifically, heart-attack rates
illustrated that the associations of GNP with his were the lowest in Japan (1.8 per 1,000) and
more recent nation-level scores on conservation among the Trappist monks (1 per 1,000), whereas
versus autonomy dimension is considerably they were the highest among U.S. European
weakened, accounting for half as much variance American populations (9.8 per 1,000). The
in his 1990 data, as compared with Hofstede’s Japanese in Hawaii had a heart-attack rate of
1970s data. Although methodological and con- 3.8 per 1,000, which further supported the idea
ceptual differences in the assessment of individ- that as people acculturate and become more indi-
ualism are an obvious culprit, it is also possible vidualistic they are likely to have more heart
that there has been a historical weakening of this attacks (Marmot & Syme, 1976).
association (Schwartz, 1994). Some research has focused on the relation-
Collectively, these findings put into question ship between individualism and collectivism
the assertion that individualism is a necessary and subjective well-being. At the societal level
requirement for the acquisition of national of analysis, Diener, Diener, and Diener (1995)
wealth, and further reinforces the assertion that found that individualism is positively correlated
there is no simple relationship between individ- with this indicator of health, even after other
ualism and collectivism and economic health. variables (i.e., income) are controlled. Yet at
We will return to this point when examining the individual level of analysis, a different
Project GLOBE results in later sections. picture has emerged. Sinha and Verma (1994)
found a positive association between allocen-
trism and psychological well-being, which was
Pace of Life
also moderated by social support. Under high
Pace of life refers to the speed, rate, and social-support conditions, allocentrism was
relative rapidity of activities within and across correlated with psychological well-being as
cultures (Levine & Norenzayan, 1999). In a more well as with cheerfulness, optimism, playful-
recent study, Levine and Norenzayan (1999) ness, and a lack of frustration, but under no-
argued that individualistic cultures have a faster social-support conditions the variables were not
pace of life than collectivist cultures. This is correlated.
based on the notion that individualistic cultures On related points, because social support is
focus on achievement, which requires a greater higher in collectivist cultures, levels of alien-
concern with time. By contrast, collectivist ation are expected to be higher in individualistic
16-House.qxd 3/19/2004 5:00 PM Page 452
• Individuals are integrated into strong • Individuals look after themselves or their
cohesive groups immediate families
• The self is viewed as interdependent • The self is viewed as autonomous and
with groups independent of groups
• Group goals take precedence over • Individual goals take precedence over group
individual goals goals
• Duties and obligations are important • Attitudes and personal needs are important
determinants of social behavior determinants of behavior
• People emphasize relatedness with groups • People emphasize rationality
• Ecologies are agricultural, and countries are • Ecologies are hunting and gathering, or industrial
often developing and wealthy
• There is a slower pace of life • There is a faster pace of life
• There are lower heart-attack rates • There are higher heart-attack rates
• There is lower subjective well-being • There is higher subjective well-being
• There are more extended family structures • There are more nuclear family structures
• Love is assigned less weight in marriage • Love is assigned greater weight in marriage
decisions decisions
• There are lower divorce rates • There are higher divorce rates
• Communication is indirect • Communication is direct
• Individuals are likely to engage in group • Individuals are likely to engage in activities
activities alone
• Individuals have fewer social interactions, but • Individuals have more social interactions, but
interactions tend to be longer and more intimate interactions tend to be shorter and less intimate
• Individuals make greater distinctions between • Individuals make fewer distinctions between
in-groups and out-groups in-groups and out-groups
NOTE: Although this table presents two extremes, it is important to recognize that these constructs represent a continuum and
that, furthermore, there is also within-culture variation.
rational determination and must be supple- behavior. Selection in such organizations can
mented with evidence that illustrates that the ter- be significantly influenced by the relations that
mination is based on specific job performance applicants have with other members within the
issues (Bhawuk, Bechtold, & Jones, 1998). organization or with institutions with whom the
By contrast, human resource management organization has contact. For example, Triandis
practices in organizations that have a collec- and Vassiliou (1972) found that employers in
tivist orientation are less likely to focus on Greece gave more weight to recommendations
rational exchanges between members and the of friends and relatives when making hiring
organization. Rather, long-term relational com- decisions, as compared with employers in the
mitments are generally established, and these United States. As such, hiring the “most quali-
relational commitments serve as guides for fied” person in collectivist cultures can involve
16-House.qxd 3/19/2004 5:00 PM Page 455
hiring a person with the best contacts and who were low on collectivism. In addition, job
relationships with the organization. Furthermore, seekers who scored high on collectivism were
because individual differences are not as impor- more likely to prefer long-term job security than
tant as a determinant for selection, it is likely individuals who scored low on collectivism. As
that training takes on much more importance such, the relationship between individualism
in collectivist cultures (Triandis, 2002; see also and collectivism and HRM has been found at
Kashima & Callan, 1994). Unlike in individual- multiple levels of analysis.
istic cultures, jobs are likely to be designed
around cohesive work groups so as to maximize
Employee Motivation
the social and the technical aspects of the job
(Erez, 1994). Such design systems would focus The nature of motivation varies in individu-
on team autonomy, team responsibility, and alistic and collectivist cultures. Most theories on
would try to maximize team-experienced mean- motivation developed within the United States,
ingfulness (Erez, 1994). an individualistic culture, focus on individual
In terms of compensation, considerations strivings based on personal needs (McClelland,
of what is equitable for the group, as well as 1967), rational calculations of possible out-
seniority and personal needs are also important comes (Vroom, 1964), and experienced psycho-
(Erez, 1994). Consistent with these notions, logical states (Hackman & Oldham, 1976).
Podsakoff, Dorfman, Howell, and Todor (1986) These perspectives highlight the importance of
found that contingent reward was not related to individuals’ interests and needs for understand-
employee performance in the Mexican cultural ing goal-directed behavior. By contrast, in col-
context. Similarly, Aycan and colleagues (1999) lectivist cultures the need to fulfill duties and
found that Indian managers reported having less obligations and to contribute to the group are
autonomy in their jobs and lower contingent central components of motivation (Kashima &
reward for their performance (a rational strat- Callan, 1994; Markus & Kitayama, 1991; Yu &
egy) than Canadian managers. Lastly, decisions Yang, 1994). These perspectives highlight the
to leave an organization would be heavily influ- importance of the social context for understand-
enced by the opinions of one’s in-group, in ing goal-directed behavior. For example, Yu
addition to market considerations (Wasti, 2000). and Yang (1994) argued that Chinese achieve-
Forced terminations are rare and are considered ment motivation is fundamentally socially
a last resort in these organizations. In this oriented, with the success of the group being a
respect, it is likely that poor performance is primary motivator. In an Islamic context, Abu-
more frequently tolerated and that it is the qual- Saad (1998) contrasted the notion of the
ity of the relationship between the individual Protestant Work Ethic (PWE) with the Islamic
and the organization that would have any effect Work Ethic (IWE). He found that whereas PWE
on downsizing decisions (Bhawuk, Bechtold, & includes an orientation toward personal success,
Jones, 1998). the IWE also measures the importance of orga-
Although this discussion has been focused nizational obligations and cooperation among
on cross-cultural differences, there is also sup- Islamic managers. Likewise, in the Japanese
port for these notions at the individual level. cultural context, Kashima and Callan (1994)
Ramamoorthy and Carroll (1998) found that argued that motivation in organizations is regu-
levels of individualist and collectivist orienta- lated through an amae-on-gimu exchange
tions of job seekers within the United States pre- between supervisors and subordinates. Within
dicted their reactions toward alternative human this system, subordinates seek to be accepted by
resource management practices. Specifically, and dependent upon their superiors, which is
they found that job seekers who were higher on referred to as amae (Doi, 1973). Amae generally
collectivism measures were less likely to prefer refers to the love and unconditional dependence
the use of selection tests for hiring, less likely to that exists between a parent and child. Within
prefer formal appraisal practices, and less likely organizations, when superiors fulfill amae, this
to prefer rewards and promotions based on produces obligations (gimu) among subordi-
individual merit compared with job seekers nates to repay such favors through hard work
16-House.qxd 3/19/2004 5:00 PM Page 456
Other studies have linked individualism and distinguished between transactional and
collectivism with conformity and cooperation in relational psychological contracts within cul-
groups. For example, in a meta-analysis Bond tures. From a cross-cultural perspective it is
and Smith (1996) found that conformity was likely that transactional contracts, which are
higher in nations that were high on collectivism, characterized by short time frames and specific
as assessed by both Hofstede (1980) and obligations, are more prevalent in individualis-
Schwartz (1994). Furthermore, numerous stud- tic cultures, whereas relational contracts, which
ies at different levels of analysis (team, individ- are characterized by long-term relationships
ual) have illustrated that collectivism is related with diffuse obligations, are more prevalent in
to cooperation. At the team level of analysis, collectivist cultures.
Eby and Dobbins (1997) illustrated that team Along similar lines, it is likely that the nature
collectivist orientation was related to team coop- of trust between employees and organizations
eration, which was in turn related to team per- varies in individualistic and collectivist cultures.
formance. Likewise, Cox, Lobel, and McLeod Trust has been described as taking on four
(1991) illustrated that groups composed of potential forms, and these forms include shal-
people with historically collectivist orientations low dependence, deep dependence, shallow
(Asians, Hispanics, and African Americans) interdependence and deep interdependence
were more cooperative than groups composed (Sheppard & Sherman, 1998). In individualistic
of Caucasians. At the individual level of analy- cultures, trust should theoretically take the form
sis, Wagner (1995) and Cox and colleagues of shallow or deep dependence. According to
(1991), found that collectivism was related to Sheppard and Sherman, the strength of trust in
cooperation in groups. Similarly, Moorman and such relationships is based on the employee’s
Blakely (1995) illustrated that individuals belief in the discretion and integrity of the orga-
higher on collectivism (Wagner, 1995) are more nization. If trust is betrayed and the employee’s
likely to engage in organizational citizenship trust is shallow, the employee may simply leave
behaviors (i.e., prosocial behaviors). These the organization. If trust is betrayed and the
studies are also consistent with Miller (1994), employee’s trust is deep, then the employee
who found that Indians are more likely to may look for a solution that allows him or her
believe that helping others is a matter of duty, to remain with the organization. In such organi-
whereas Americans are more likely to believe zations, this solution may involve the establish-
that helping others is a matter of choice. At the ment of a third entity to oversee and arbitrate
same time, these findings are likely moderated future conflict, such as an employee’s union. On
by whether the other group is an in-group or an the other hand, in collectivist cultures, trust
out-group (Triandis, 1995). should theoretically take the forms of “shallow
All in all, these findings illustrate that there interdependence” and “deep interdependence.”
is an emphasis on cooperative team processes The strength of trust would be a function of
in collectivistic cultures. Later, we will return to the level of interdependence between parties.
this issue when discussing the hypothesized Moreover, if trust is betrayed, individuals
relationship between collectivism and team- whose personal identities are defined, in large
oriented leadership behaviors as well. part, by their organizational membership are
likely to experience a tremendous amount of
Organizational Trust and dissonance.
Psychological Contracts
Although there has been little empirical Accountability
research in this area, it is likely that the nature of Organizations must rely on a division of
psychological contracts and organizational trust labor and make members accountable for task
varies in individualistic and collectivist cultures. accomplishment. However, the nature of
Psychological contracts refer to the nature of accountability is likely to vary in individualis-
the exchange relationship between employ- tic and collectivist cultures. In individualistic
ees and organizations. Rousseau (1995) has cultures, accountability is likely to rest with
16-House.qxd 3/19/2004 5:00 PM Page 458
Table 16.2 Higher Individualism and Collectivism for Organizations Versus Lower Individualism and
Collectivism for Organizations
Organizations That Score High on Collectivism Organizations That Score High on Individualism
• Members assume that they are highly • Members assume that they are independent
interdependent with the organization and of the organization and believe it is important
believe it is important to make personal to bring their unique skills and abilities to the
sacrifices to fulfill their organizational organization
obligations
• Employees tend to develop long-term • Employees develop short-term relationships, and
relationship with employers from recruitment change companies at their own discretion
to retirement
• Organizations take responsibility for • Organizations are primarily interested in the
employee welfare work that employees perform and not their
personal or family welfare
• Important decisions tend to be made by • Important decisions tend to be made by
groups individuals
• Selection can focus on relational attributes • Selection focuses primarily on employees’
of employees knowledge, skills, and abilities
• Jobs are designed in groups to maximize the • Jobs are designed individually to maximize
social and technical aspects of the job autonomy
• Training is emphasized more than selection • Selection is emphasized more than training
• Compensation and promotions are based on • Compensation and promotions are based on an
what is equitable for the group and on equity model, in which an individual is rewarded
considerations of seniority and personal needs in direct relationship to his or her contribution to
task success
• Motivation is socially oriented, and is based on • Motivation is individually oriented and is based
the need to fulfill duties and obligations and to on individual interests, needs, and capacities
contribute to the group
• Organizational commitment is based on • Organizational commitment is based on
expectations of loyalty and in-group individuals’ rational calculations of costs and
attitudes benefits
• Prosocial behaviors, or organizational • Prosocial behaviors, or organizational citizenship
citizenship behaviors, are more common behaviors, are less common
• Avoidant, obliging, compromising, and • Direct and solution-oriented conflict resolution
accommodating conflict resolution tactics tactics are preferred
are preferred
• Accountability for organizational successes • Accountability for organizational successes and
and failures rests with groups failures rests with individuals
individualistic culture, has now been replicated leadership (P), and another that was focused on
in a study in Japan (Misumi, 1985). Specifically, the maintenance of group relations, labeled main-
in his PM theory of leadership, Misumi found tenance-oriented leadership (M). Nevertheless,
that effective leader behaviors in Japan were there are some differences in PM leadership
composed of two dimensions, one that was theories in individualistic and collectivist cultures.
focused on the task, labeled performance-oriented First, the degree to which the two clusters of
16-House.qxd 3/19/2004 5:00 PM Page 460
behaviors are seen as distinct varies across on performance in Taiwan and in Mexico, both
cultures. Smith, Misumi, Tayeb, Peterson, and collectivist cultures (Hofstede, 1980). Partici-
Bond (1989) found that respondents in the United pative leadership, on the other hand, only had
States and Britain made clearer distinctions positive effects in the United States, presumably
between harmony maintenance and task accom- because high individualism reinforces beliefs in
plishment, whereas respondents in collectivist participation. However, there is some evidence
cultures perceived work group harmony to be inti- that participative leadership behaviors may be
mately related to task accomplishment. Second, preferred to delegation in some collectivist cul-
research has found that although there is a set of tures, such as Kuwait (Ali, Taqi, & Krishnan,
core P and M leadership behaviors that are 1997) and India (Aycan et al., 1999). For
enacted by leaders in both collectivist and indi- example, Aycan and colleagues (1999) found
vidualistic cultures, such general functions can be that Indian managers had stronger beliefs in the
fulfilled through different specific behaviors importance of employee participation, as com-
(Smith et al., 1989). pared to Canadian managers. At the individual
More specifically, behaviors linked to the M level of analysis, they also found that managers
function in Britain were directed more toward who reported that the sociocultural context was
the task than toward in-group maintenance, high on paternalism and loyalty toward commu-
whereas M behaviors in Japan and Hong Kong nity also reported that employees within the
had much more to do with group interaction. By organization had high obligations toward others
contrast, in Japan and Hong Kong, distinctive P and should use joint goal setting. Likewise, Ali
behaviors focused more on collective inter- and colleagues (1997) found that Kuwaitis, who
actions, whereas in Britain, P behaviors were were highly collectivist, had a strong preference
more task-centered. In addition, Smith and col- for consultative and participative decision-
leagues (1989) illustrated that in collectivist making styles. Thus, there is inconsistency in
cultures M functions are best characterized by a the literature regarding the relationship between
deep concern for subordinates and by the ability individualism and collectivism and participa-
to resolve personal difficulties in an indirect tion. In keeping with the above notion that
manner that ensures face saving (Smith et al., leader constructs may be defined differently
1989), whereas M functions in Britain were across cultures, it is possible that participation
related to work and not personal issues involv- is equally valued in both individualistic and
ing subordinates. Indeed, these findings are also collectivist cultures, but is manifested through
consistent with research in India by Sinha different behaviors.
(1984), who found that the most effective leader
exhibits both person-oriented and production-
Cognitive Tradition
oriented behaviors. Specifically, an effective
leader in India is one who is paternalistic and Cognitive perspectives on leadership focus
provides nurturance in exchange for subordi- on the process by which a person becomes
nates’ task accomplishment. Interestingly, the labeled as a leader and the characteristics that
nurturant-task style described by Sinha is simi- are perceived to be associated with effective
lar to leader-subordinate relationships in Japan, leadership. Within this approach, personal iden-
another collectivist culture, which are character- tities are defined, in large part, by their organi-
ized by amae, or indulgent dependence, discussed zational membership; for an individual to be
previously (Kashima & Callan, 1994; see also considered a leader, he or she must first be
Ling, 1989). categorized as a leader (Dorfman, 1998; Erez &
In other studies within the behavioral tradi- Earley, 1993; House, Wright, & Aditya, 1997;
tion, Dorfman and his colleagues (Dorfman & Shaw, 1990). During the cognitive categoriza-
Howell, 1988; Dorfman et al., 1997) found that tion process, a person’s attributes are compared
the effects of leaders’ directiveness and partici- with the attributes that potential followers believe
pativeness on performance varied across “prototypical” leaders to have. Such prototypes
individualistic and collectivist cultures. are thought to reflect a culture’s image of an ideal
Directive leadership had a positive impact leader (Hanges, Lord, & Dickson, 2000). As
16-House.qxd 3/19/2004 5:00 PM Page 461
such, the better the match between a person’s paternalism, group maintenance activities, face
characteristics and followers’ leader prototypes, saving, conformity, as well as charisma and lack
the more likely it is for the person to be consid- of social loafing. By contrast, leadership in
ered a leader (Lord & Maher, 1991). individualistic cultures emphasizes individual
Although these notions are theoretically discretion, autonomy, and task accomplishment
appealing, there have been very few direct (Erez & Earley, 1993; Hofstede & Bond, 1988;
empirical tests of them in the literature. Triandis, 1993). In individualist cultures there is
Dorfman (1998) theorized that in the individual- higher susceptibility to social loafing as well. A
istic culture of the United States, leader proto- summary of the implications of individualism
types reflect the cultural values of being and collectivism for leadership can be found in
independent, strong willed, and forceful, Table 16.3.
whereas in collectivist cultures such as Japan
prototypes reflect the cultural values of interde-
pendence, collaboration, and self-effacement. OVERVIEW OF GLOBE RESEARCH ON
One study that did directly assess leader proto- INDIVIDUALISM AND COLLECTIVISM
types found support for the notion that business
leader prototypes vary systematically as a func- As is evident from the previous discussion, the
tion of culture (Gerstner & Day, 1994). constructs of individualism and collectivism
have received much theoretical and empirical
attention in the literature over the past 20
Charismatic Tradition
years. It is clear that the constructs have
Various researchers have speculated that col- numerous economic, social, and organizational
lectivism at the societal and organizational correlates; operate at multiple levels of analy-
levels is associated with charismatic leadership. sis; and are often multidimensional in nature.
For example, Jung, Bass, and Sosik (1995) pro- The purposes of the GLOBE research project
vided a conceptual argument that transforma- of 62 nations have already been reviewed in
tional leadership would emerge more easily and detail in Chapter 2 by House and Javidan. In
would be more effective in collectivist as com- this section we review and interpret the results
pared to individualistic societal cultures because from the GLOBE study with respect to collec-
of the strong emphasis on group orientation, tivism and individualism. Before we turn to the
work centrality, and respect authority in collec- empirical results, we will briefly introduce the
tivist cultures. Meindl (1995) also argued that measures of the constructs within this research
for the work-group level the emergence of project.
charismatic leadership would be higher in First, in keeping with a levels perspective,
collectivist than individualistic organizational individualism and collectivism were assessed at
cultures. This is predicated on the notion that a both the societal and the organizational levels
readiness for charismatic leadership within of analysis in the GLOBE study. Previous
work groups is a function of the degree of cohe- research on the constructs has not examined
siveness or interconnectedness of relational these levels simultaneously, and as such, the
networks within a work group, which in turn, is current research builds on previous research
a function of collectivism (see Kashima et al., across different fields (e.g., cross-cultural psy-
1995). In support of this, Pillai and Meindl chology, organizational behavior). On the basis
(1998) found a significant positive correlation of the previous discussion we expected that
between work group collectivism and charis- societal-level individualism and collectivism
matic leadership (r = .37, p < .001). would be related to organizational-level indi-
vidualism and collectivism, providing evidence
for cross-level effects of the constructs.
Summary
Furthermore, this design enabled us to examine
Research suggests that leadership can vary in the joint contribution of both levels of analysis
collectivist and individualistic cultures. In col- to perceptions of effective leadership. We
lectivist cultures, leadership is associated with expected that collectivism scores at both levels
16-House.qxd 3/19/2004 5:00 PM Page 462
of organizational collectivism, and would whether children take pride in the individual
consequently have unique relationships with accomplishments of their parents and vice
effective leadership. versa, whether aging parents live at home with
Lastly, GLOBE measures of these constructs their children, and whether children live at
also made another important differentiation home with their parents until they get married
between collectivist values and collectivist (see Tables 16.4b and 16.4d for sample items).
practices. Previous research has included items For both Institutional Collectivism and In-
that focus only on values (Schwartz, 1994), Group Collectivism, there were As Is and
practices (Hofstede et al., 1991) or a mix of both Should Be versions of the scales. The As Is
in the same scale (Hofstede, 1980). The current scales were probed by the statement, “In this
research explicitly examined both values and society, people are generally . . .” and assessed
practices at both the societal and organizational existing practices in societies. The Should Be
level of analysis to determine whether any scales, on the other hand, have the same content
important variation between the two can be except they were probed by the statement, “In
captured with this distinction. this society, people should . . .” These scales
In what follows, we first describe GLOBE assessed societal values. Study participants
results for collectivism at the societal and orga- responded to all items on a 7-point scale in
nizational level of analysis. We then discuss which 1 = low collectivism (or high individu-
the relationship of societal and organizational alism) and 7 = high collectivism (or low indi-
collectivism scales and perceptions of effective vidualism). It is important to note that the items
leadership attributes. Please refer to Chapter 6 in these scales were constructed by GLOBE in
by House and Hanges for a further description ways that assessed Individualism and
of the 62 nations included in this study, sample Collectivism as opposites of the same contin-
characteristics, how scales were developed, and uum. As reviewed previously, whereas some
study procedures. scholars concur with the unidimensionality of
the constructs at the societal level (e.g.,
Hofstede, 1980; Schwartz, 1994), others argue
that they can be orthogonal (e.g., Triandis,
GLOBE MEASURES
1995). Factor analyses of the scales demon-
strated that they are all unidimensional.
Societal-Level Measures
It is also important to note that there were
Institutional Collectivism: The Institutional some differences across the practices and val-
Collectivism construct was measured through a ues scales for both Institutional and In-Group
set of four questions that were focused on the Collectivism scales. For example, the Institu-
degree to which institutional practices at the tional Collectivism practices scale included a
societal level encourage and reward collective unique item regarding the emphasis that is
action. Specifically, the questions assessed placed on being accepted in other groups that
whether group loyalty is emphasized at the is not included in the values scale. Likewise,
expense of individual goals, whether the the Institutional Collectivism values scale
economic system emphasizes individual or included an item regarding the preference
collective interests, whether being accepted by people have for individual versus team sports
other group members is important, and whether that was not included in the practices scale.
individualism or group cohesion is valued more In addition, the In-Group Collectivism prac-
in the society (see Tables 16.4a and 16.4c for tices scale focused exclusively on families,
sample items). children, and parents. By contrast, the In-Group
In-Group Collectivism: The Societal In- Collectivism values scale also included a focus
Group Collectivism construct was also opera- on the value of having pride in the society as a
tionalized by a set of four questions that whole. Therefore, although the scales mostly
assessed the degree to which individuals measure common elements, they also have
express pride, loyalty, and interdependence in some uniqueness and have some qualitative
their families. The items specifically measured differences.
16-House.qxd 3/19/2004 5:00 PM Page 464
In this society, leaders encourage group loyalty even if individual goals suffer: (reverse scored)
Strongly agree Strongly disagree
1 2 3 4 5 6 7
In this society, children take pride in the individual accomplishments of their parents: (reverse scored)
Strongly agree Strongly disagree
1 2 3 4 5 6 7
In this society, parents take pride in the individual accomplishments of their children: (reverse scored)
Strongly agree Strongly disagree
1 2 3 4 5 6 7
I believe that, in general, leaders should encourage group loyalty even if individual goals suffer: (reverse
scored)
Strongly agree Strongly disagree
1 2 3 4 5 6 7
I believe that the economic system in this society should be designed to maximize:
Individual interests Collective interests
1 2 3 4 5 6 7
In this society, children should take pride in the individual accomplishments of their parents: (reverse
scored)
Strongly agree Strongly disagree
1 2 3 4 5 6 7
In this society, parents should take pride in the individual accomplishments of their children:
Strongly agree Strongly disagree
1 2 3 4 5 6 7
16-House.qxd 3/19/2004 5:00 PM Page 465
In this organization, managers encourage group loyalty even if individual goals suffer: (reverse scored)
Strongly agree Strongly disagree
1 2 3 4 5 6 7
In this organization, group members take pride in the individual accomplishments of their group:
Strongly agree Strongly disagree
1 2 3 4 5 6 7
In this organization, group managers take pride in the individual accomplishments of group members:
Strongly agree Strongly disagree
1 2 3 4 5 6 7
I believe that in this organization, managers should generally encourage group loyalty even if individual
goals suffer: (reverse scored)
Strongly agree Strongly disagree
1 2 3 4 5 6 7
In this organization, the pay and bonus system should be designed to maximize:
Individual interests Collective interests
1 2 3 4 5 6 7
In this organization, group members should take pride in the individual accomplishments of their group:
Strongly agree Strongly disagree
1 2 3 4 5 6 7
In this organization, group managers should take pride in the individual accomplishments of group
members:
Strongly agree Strongly disagree
1 2 3 4 5 6 7
items related to the importance of group loyalty This illustrates that respondents across societies
versus individual goals, the pay and bonus sys- generally want more in-group and institutional
tem maximizing group or individual interests, collectivism than they have. Finally, it is also
and the emphasis on group cohesion versus interesting to note patterns for the variances of
individualism. In addition, the values version of the scales. For example, it is clear that there is
this scale also included additional items that lower variance across cultures (or more univer-
measure the value of achieving consensus and sal agreement) for In-Group Collectivism
the value of group versus individual work in the values, as well as Institutional Collectivism
organization. There was also some variation practices and values. By contrast, the GLOBE
across the practices and values versions of the results illustrate that societies are most variable
organizational In-Group Collectivism scale. on In-Group Collectivism practices.
Specifically, the values version included an Table 16.6 presents the correlations among
item concerning whether outsider statements all four Collectivism scales. As shown in this
against the organization should be bothersome table, the correlation between the societal
to members inside the organization. Thus, it is practices and values versions of the Institu-
important to note that there are some qualitative tional Collectivism scale is negative (r = −.61,
differences between the scales. p < 0.01). This indicates that the less a society
practices collectivism and emphasizes collec-
tive goals and interests, the more it values soci-
etal collectivism. Peng and Nisbett (1999) refer
GLOBE RESULTS: SOCIETAL LEVEL
to this phenomenon as “deprivation-based pref-
erences” at the individual level, wherein indi-
Descriptive Statistics
viduals express preferences for things that they
Table 16.5a presents the summary statistics believe they are lacking or things they have
for both the practices and values versions of taken for granted. By contrast, there is no sig-
the societal Institutional Collectivism scale, and nificant correlation between societal In-Group
Table 16.5b presents summary statistics for Collectivism for the practices and values scale,
societal In-Group Collectivism scale. These indicating that this same sentiment is not applic-
tables illustrate some interesting patterns. First, able to societal emphases on family pride and
it is evident that across all cultures, values and interdependence.
practices scores are higher for In-Group Interestingly, the table illustrates that the
Collectivism, as compared to Institutional Institutional Collectivism practices scale is
Collectivism. Second, it is interesting to note not at all correlated with either the In-Group
that for both the In-Group Collectivism and the Collectivism practices or values scale. As such, the
Institutional Collectivism scales, scores are Institutional Collectivism practices scale is highly
higher for values than they are for practices. distinct from both In-Group Collectivism scales.
16-House.qxd 3/19/2004 5:00 PM Page 467
In-Group .21
Collectivism practices
In-Group
Collectivism values
Band
A B C D
difference), where SED is a function of the bands the mean scores are not statistically
reliability of the scale of interest. As stated ear- different from each other.
lier, scores in band A are significantly higher Tables 16.7c and 16.7d present the country
than scores in bands B, C, or D, but within rankings and means for the Institutional and
16-House.qxd 3/19/2004 5:00 PM Page 469
Band
A B C
In-Group Collectivism values scales.6 In these are interpreted in more detail, after we
tables, the bands are based on the GLOBE describe the correlations of GLOBE
scale scores, and the countries are ranked from Collectivism scales with other GLOBE scales
the country with the highest mean to the and with other collectivism scales in the
country with the lowest mean. These rankings literature.
16-House.qxd 3/19/2004 5:00 PM Page 470
Band
A B C D
Band
A B C
Table 16.8a Correlations of Societal Institutional Collectivism and Other GLOBE Societal Dimensions
Institutional Institutional
GLOBE Scale Collectivism Practices Collectivism Values
Table 16.8b Correlations of Societal In-Group Collectivism With Other GLOBE Societal Dimensions
In-Group In-Group
GLOBE Scale Collectivism Practices Collectivism Values
included many of the Asian Dragons, which countries, and several Asian countries. Not
have been economically successful, yet using surprisingly based on previous culture studies,
Confucian values as a basis for such success. In cultures that score low on this syndrome include
addition, some Scandinavian cultures are also many Western cultures, including the United
characteristic of this syndrome, illustrating a States, Australia, England, Canada, Germany,
combined focus on collective interests and and some Scandinavian cultures as well. Thus,
nonassertiveness. This cultural syndrome may it is clear from these results that societies such
be reflective of a dual emphasis on collectivism as those found in Scandinavia can score high
and self-transcendence values found in on certain forms of collectivism (Institutional
Schwartz’s circumplex of values. It may also Collectivism) yet score very low on other forms
be reflective of Triandis’s notion of “horizontal of collectivism (In-Group Collectivism). It is
collectivism” (Triandis, 1995; Triandis & also worth noting that this form of collectivism
Gelfand, 1998), which emphasizes collective seems to overlap with Triandis’s (1995;
interests in which power differences practices Triandis & Gelfand, 1998) notion of “vertical
are not as important. collectivism.”
By contrast, In-Group Collectivism practices GLOBE scores on Collectivism values (or
seem to be part of a cultural syndrome in which what should be) also reflect distinct cultural
there are close ties among family members, and syndromes. Cultures characterized by high
in which people are concerned with others, are scores on Institutional Collectivism values,
respectful of authority, and have fewer rules. As which feature a strong focus on collective
illustrated in the previous rankings, countries interests, are more likely to be characterized by
that are characterized by this cultural syndrome practices wherein there are close family ties,
include many Latin American and African authority is respected, short-term focus and
16-House.qxd 3/19/2004 5:00 PM Page 474
men tend to be dominant, and there are few value scores. As can be seen, of the four
rules and little structure. As can be seen in this GLOBE scales, In-Group Collectivism prac-
description, this syndrome is similar to that tices have the strongest correlations with these
found for societies characterized by In-Group other collectivism scales.
Collectivism practices, which is consistent In particular, as expected, there is a
with correlations reported previously between strong negative correlation between In-Group
these two scales (r = .43, p < .01). And, as can Collectivism practices and Hofstede’s (1980)
be seen in the rankings for this scale, societies measure of individualism (r = −.82; p < .01).
that score high on this syndrome include This fits with Hofstede’s conceptual definition
numerous Latin and Central American cultures of collectivism. In addition, recall that
and Spain, and several Middle Eastern cultures GLOBE’s societal In-Group Collectivism scale
(i.e., Kuwait and Turkey). Many of these was in part derived from Triandis and col-
nations have not had economic prosperity in leagues’ (1986) culture-level family integrity
the recent past, yet perhaps are in transition scale, and that the latter was also previously
and in search of new ways to organize in the correlated with Hofstede’s (1980) scores on
21st century. By contrast, cultures that score collectivism (r = .78). Thus, the current results
low on this syndrome include many Western also replicate this relationship 15 years after
and Scandinavian cultures, and several Eastern Triandis conducted his research.
European cultures (e.g., Hungary, Poland, Indeed, a comparison of Hofstede’s country
Slovenia). Notably, it also includes several rankings on individualism with the country
Confucian Asian cultures, such as Japan, rankings of GLOBE’s In-Group Collectivism
Korea, and Hong Kong. This is perhaps reflec- practices reveals that, for the most part, the
tive of a decrease in emphasis on some forms same countries that were considered individual-
of collectivism in Confucian Asian cultures istic in Hofstede’s (1980) dataset would still be
(see also Oyserman et al., 2002, for similar classified as individualistic in the GLOBE
findings on Japan and Korea). dataset. Indeed, there is remarkable consistency
Finally, societies with high In-Group among the countries that are were in the top
Collectivism value scores also tend to be char- third of Hofstede’s most individualistic cultures
acterized by practices wherein authority is and the countries that are in the most individu-
respected and wherein there are few rules alistic band in the GLOBE data (band C). These
and little structure, more short-term orientation, countries include the United States, Australia,
and less performance orientation. As can be Great Britain, Canada, the Netherlands, New
seen, this syndrome is somewhat similar to that Zealand, Denmark, and Sweden. Likewise,
found for societies characterized by Institutional the countries that fell into the most collectivist
Collectivism values, which is consistent with third in Hofstede’s dataset also tended to fall
correlations between these two scales reported into the most collectivist band (band A) within
previously (r = .29, p < .05). Interestingly, the the GLOBE dataset. This includes Venezuela,
United States, although falling low on Institu- Colombia, Thailand, Portugal, Mexico, and the
tional Collectivism values scores, is in the Philippines. Overall, these results suggest that
middle ranking on this scale, suggesting that rankings on societal individualism and collec-
respondents in the United States believe tivism have been remarkably stable over the last
they should be more focused on this form of 30-year period.
collectivism. Within this framework of similarities, there
are also some interesting differences. For
Correlations of GLOBE example, Hong Kong was in the most collec-
Collectivism and Hofstede and Schwartz tivist third in Hofstede’s dataset, but was clus-
tered in band B within the GLOBE dataset. This
Table 16.9 displays the correlations among may be due to the rapid economic growth within
the four GLOBE societal Collectivism scores, this country in the past 20 years, which may
Hofstede’s (1980) individualism scores, and have led to increased social independence, as
Schwartz’s (1994) embeddedness and autonomy discussed previously in this chapter. Of course,
16-House.qxd 3/19/2004 5:00 PM Page 475
Table 16.9 Correlations Between GLOBE’s Societal Collectivism Scales and Hofstede’s (1980) and
Schwartz’s (1994) Scales
Schwartz’s Schwartz’s
Hofstede’s Schwartz’s Intellectual Affective
Individualism Embeddedness Autonomy Autonomy
GLOBE Scales Scale Value Scale Value Scale Value Scale
differences in scale characteristics and samples the conceptual definition of conservatism that
between the two studies could also be operative, was offered by Schwartz (1994), which
necessitating caution in interpreting such included the importance of close-knit harmo-
specific comparisons. nious relations in which the interests of the
Table 16.9 also illustrates that other GLOBE person are not viewed as distinct from those of
scales also correlated with Hofstede’s (1980) the group.
measure. Specifically, there was a significant In sum, these correlations reveal that of the
negative correlation between Institutional two societal level Collectivism scales used in
Collectivism values and Hofstede’s individ- the GLOBE study, the In-Group Collectivism
ualism scores (r = −.55, p < .01) although this practices scale is the most highly correlated
relationship is weaker than the In-Group with other well-established societal-level col-
Collectivism practices scale. Thus, according to lectivism scales in the literature. Thus, this scale
these results, Hofstede’s (1980) ranking is most received substantial convergent validity. It is
similar to the ranking on GLOBE’s In-Group interesting that the In-Group Collectivism prac-
Collectivism practices, but also shares some tices scale correlates with several of Schwartz’s
variance with the GLOBE Institutional Collec- values scales (positively with embeddedness,
tivism values scale. and negatively with intellectual and affective
With respect to Schwartz’s (1994) value autonomy). Although the In-Group Collec-
scales, Table 16.9 illustrates that GLOBE tivism values scale was also correlated with
In-Group Collectivism practices is positively Schwartz’s measure of embeddedness, it was
correlated with embeddedness (r = .66, p < .01) nonsignificant. As reviewed previously, how-
and negatively correlated with intellectual ever, it is possible that there was a restriction of
autonomy (r = −.59, p < .01) and affective range of other GLOBE scales. Indeed, the In-
autonomy (r = −.67, p < .01). This too, fits with Group Collectivism practices scale had the
16-House.qxd 3/19/2004 5:00 PM Page 476
highest variability, which facilitated its respondents place such a strong emphasis on
correlation with other scales in the literature. In-Group Collectivism (both practices and
Based on these results, it is also evident that values, see below), this creates a vantage point
GLOBE’s Institutional Collectivism practices from which societal practices are perceived
scale seems to measure a different aspect of as actually being too individualistic or are not
collectivism from those previously examined at to be trusted.
the societal level of analysis. For example, as By contrast, the clusters with the highest
discussed in the previous section, Institutional Institutional Collectivism values scores were
Collectivism practices seem to be part of a cul- Latin America, the Middle East, and Southern
tural syndrome wherein such cultures are future Asia. On the other hand, Nordic Europe, Anglo
focused and performance oriented yet seek to countries, and Eastern Europe scored among the
accomplish such orientations through practices lowest on Institutional Collectivism values.
that emphasize being concerned about others, It is also interesting to note that respon-
and not being assertive or power dominating. dents in most clusters desire more of
This syndrome is not well captured by either Institutional Collectivism than they actually
Hofstede’s (1980) measure or by the embedded- have (i.e., values scores are higher than prac-
ness versus autonomy distinction in Schwartz tices scores). This is the case for the Middle
(1994). It may however, reflect, the combina- East, Eastern Europe, Germanic Europe, Latin
tion of embeddedness and transcendence in America, Latin Europe, Southern Asia, and
Schwartz’s theory. Sub-Saharan Africa cultures. There may be
different reasons, however, for these findings,
Regional Differences depending on the region. For example, it may
and GLOBE Societal Collectivism be that Germanic European respondents have
a longing for more collectivism because of
In Table 16.10a, the GLOBE countries are their focus on individualistic practices. On the
sorted into 10 clusters and the respective means other hand, respondents in the Middle East,
for Institutional Collectivism practices and val- Sub-Saharan Africa, and Latin America clus-
ues are presented (see Chapter 10 by Gupta and ters may have higher scores on values than
Hanges for a discussion of the construction of practices because of the fact that they are
clusters). Table 16.10b presents the mean undergoing rapid cultural changes in political
scores for In-Group Collectivism practices and and economic structures (see the Annual
values for each cluster. Analyses of variance Report, 2001, from the World Bank, for argu-
revealed that the clusters’ means are statis- ments along these lines). The only clusters
tically different for each of the four scales: wherein societal practices are rated higher
Institutional Collectivism practices (F[9, 51] than values are Confucian Asia, Anglo, and
= 6.22, p < .01), Institutional Collectivism Nordic Europe. This illustrates that respon-
values (F[9, 51] = 9.36, p < .01), In-Group dents in such regions express a desire for less
Collectivism practices (F[9, 51] = 14.98, institutional collectivism than they report hav-
p < .01), and In-Group Collectivism values ing. Again, this may be due to different rea-
(F[9, 51] = 6.47, p < .01). sons, depending on the region. Respondents in
Table 16.10a illustrates several interesting Anglo regions, which are ruggedly individual-
patterns. First, the clusters with the highest istic, may highly devalue Institutional Collec-
Institutional Collectivism practices scores tivism but not In-Group Collectivism (see
were Nordic Europe and Confucian Asia. The below) because of the historical distrust of
clusters with the lowest scores on Institutional collective structures and goals imposed
Collectivism practices were Latin America, through colonization. By contrast, respondents
Latin Europe and Germanic Europe. This is in Confucian Asian cultures may be less
consistent with the previous discussion of the inclined to endorse institutional collectivism
country rankings. The scores for Latin because of rapid changes toward individual-
America may, at first glance, seem surprising. ism within the structure of societies (e.g., less
However, it may be that since Latin American focus on lifetime employment). Clearly, these
16-House.qxd 3/19/2004 5:00 PM Page 477
Table 16.10a Culture Clusters and Institutional Collectivism—Practices and Values Mean Scores**
(Continued)
16-House.qxd 3/19/2004 5:00 PM Page 478
Costa Rica
Ecuador
El Salvador
Guatemala
Mexico
Venezuela
NOTE: Means with the letter superscript a are significantly higher than the others at p < .05, b are not significantly different
from the others, and c are significantly lower than the others.
* Higher scores indicate more collectivism.
** Number of cases: 61.
arguments are merely speculative and need to Turning to the In-Group Collectivism scores,
be tested in future research, especially because Table 16.10b also reveals some interesting
it is argued that cultural practices and values patterns. First, the clusters with the highest
change very slowly across generations (Hofstede, In-Group Collectivism practices scores were
2001). Southern Asia, Middle East, Eastern Europe,
16-House.qxd 3/19/2004 5:00 PM Page 479
Table 16.10b Culture Clusters and In-Group Collectivism—Practices and Values Mean Scores**
In-Group In-Group
Collectivism Collectivism
Culture Cluster Practices Mean Scores Values Mean Scores
(Continued)
16-House.qxd 3/19/2004 5:00 PM Page 480
In-Group In-Group
Collectivism Collectivism
Culture Cluster Practices Mean Scores Values Mean Scores
Costa Rica
Ecuador
El Salvador
Guatemala
Mexico
Venezuela
NOTE: Means with the letter superscript a are significantly higher than the others at p < .05, b are not significantly different
from the others, and c are significantly lower than the others.
* Higher Scores indicate more collectivism.
** Number of cases: 61.
Latin America, and Confucian Asia. By con- this scale is most highly correlated with
trast, Anglo, Germanic Europe, and Nordic Hofstede’s (1980) rankings of nations. This is
Europe clusters score the lowest on In-Group also reflected in these cluster rankings, which
Collectivism practices. As discussed previously, are similar to patterns identified by Hofstede.
16-House.qxd 3/19/2004 5:00 PM Page 481
Thus, even with different methods and with the Collectivism and these social and economic
passage of 30 years, the rankings on this form of indicators were computed. The results are orga-
collectivism remain similar. nized into the following categories: Economic
Table 16.10b also illustrates that the health, success in science and technology,
clusters with the highest In-Group Collec- human condition, family and friends, achieve-
tivism values scores were Latin America and ment, political ideology, religion, and gender
Anglo. Particularly low scores were found for equality. For all of these analyses, we also
Confucian Asia and Germanic Europe. This is report whether the relationships between
important, as much research on the constructs Institutional and In-Group Collectivism change
has previously focused on Asian and American when controlling for national wealth (GNP per
comparisons in which Asian respondents capita, based on 1998 data). As noted in Chapter
are seen as desiring In-Group Collectivism. 7 by Javidan and Hauser, In-Group Collectivism
However, these data illustrate that respondents practices are highly negatively correlated with
in Anglo cultures actually express a desire GNP per capita (r = −.76, p < .01), making it
for more of this form of collectivism than difficult to separate the two analytically.
Confucian Asian cultures. Perhaps because it Nevertheless, we report both analyses below for
has historically been low in this region. the interested reader.
Finally, it is notable that the Anglo and the
Nordic Europe clusters have large discrepancies
GLOBE Individualism–Collectivism
between respondents’ In-Group Collectivism
values and practices. Specifically, scores for
and Economic Health
these clusters have much higher values than As discussed in previous sections, the
practices on In-Group Collectivism, which is question of whether individualism or collec-
perhaps not surprising given that both regions tivism is associated with healthier economies
have been characterized as having low In-Group has been the focus of some research. Whereas
Collectivism in the past. Other clusters, such as some scholars have asserted that individualism
Germanic Europe, also have higher scores on leads to greater economic growth, others have
values than practices, although they are not as suggested that collectivism is also associated
pronounced. Finally, two clusters—namely with economic growth. The GLOBE findings
Southern Asia and Confucian Asia—have on these issues are presented in Tables 16.11a
higher scores on practices than values, which and 16.11b.
may be indicative of the changes in interdepen- Specifically, as discussed in Chapter 7 by
dence among family members as such regions Javidan and Hauser, economic health was
acquire more affluence (see previous discussion assessed with four measures: Economic Pros-
on the relationship between collectivism and perity, Public Sector Support for Prosperity,
wealth, and the Annual Report, 2001, from the Societal Support for Competitiveness, and the
World Bank). Global Competitiveness Index from the World
Economic Forum (1979).
Table 16.11a illustrates that societal
GLOBE Measures and
Institutional Collectivism has some relation-
Other Economic and Social Indicators
ships with economic health. Three of the
In this section, we describe how four economic indicators are significantly corre-
Individualism–Collectivism is related to eco- lated with Institutional Collectivism scores.
nomic and social indicators. The indicators of Institutional Collectivism practices are posi-
interest have been extracted from three major tively correlated with Economic Prosperity,
databases: The World Economic Forum’s Public Sector Support for Economic Prosperity,
Competitiveness Rankings, the United Nations and Competitiveness Index. Societies reported
Human Development Report, and the World to be practicing Institutional Collectivism tend
Values Survey. Correlations between the to be more competitive and more economically
practices and values scales of both GLOBE prosperous. They also seem to have public sec-
Institutional Collectivism and In-Group tors that are supportive of economic success.
16-House.qxd 3/19/2004 5:00 PM Page 482
Institutional Institutional
Economic Health Collectivism Practices Collectivism Values
Economic prosperity .33* −.48**
N = 57 N = 57
Public sector support for prosperity .36* −.46**
N = 40 N = 40
Societal support for competitiveness .28 −.29
N = 40 N = 40
WEF competitiveness index .40** −.47**
N = 41 N = 41
In-Group In-Group
Economic Health Collectivism Practices Collectivism Values
Economic prosperity −.78** −.29*
N = 57 N = 57
Public sector support for prosperity −.67** −.27
N = 40 N = 40
Societal support for competitiveness −.14 −.11
N = 40 N = 40
WEF competitiveness index −.45** −.27
N = 41 N = 41
Institutional Institutional
Science and Technology Collectivism Practices Collectivism Values
Success in basic science .39* −.53**
N = 40 N = 40
In-Group In-Group
Science and Technology Collectivism Practices Collectivism Values
Success in basic science −.45** −.44**
N = 40 N = 40
become nonsignificant after controlling for The results in Table 16.11c show that
GNP per capita is controlled for. The signifi- Institutional Collectivism practices are posi-
cant relationships between In-Group Collec- tively related to success in basic science,
tivism values scores and economic prosperity whereas Institutional Collectivism values are
also become nonsignificant after controlling for negatively related to success in basic science.
GNP. Because GNP per capita and economic These findings suggest that in societies with
prosperity are positively related (r = .91, p < high Institutional Collectivism practices scores,
.01), the above findings suggest that prosperity there is greater interest in basic science and
is driven by higher GNP and not necessarily there is also greater cooperation between aca-
lower In-Group Collectivism. demia and the business world. The greater
incidence of technology transfer between uni-
GLOBE Individualism–Collectivism and Success versities and corporations may be due to a gen-
in Science and Technology. We now turn to a eral societal focus on collective interests and
discussion of GLOBE Individualism–Collecti- cooperation within collectivistic societies in the
vism and success in science and technology, as interest of the betterment of the entire society.
well as the relationship between individual- As illustrated in Table 16.11d, both In-Group
ism–collectivism and societal attitudes toward Collectivism values and practices are negatively
science and research more generally. Success in and significantly correlated with success in
basic science reflects the amount of technology basic science. However, both of these corre-
transfer that occurs between corporations and lations become nonsignificant after controll-
universities, the ability of research to enhance ing for GNP per capita. Thus, the positive
economic and technological development, as relationship between individualism and success
well as the ability of science and technology to in basic science is likely driven by higher GNP
arouse the interest of youth. per capita in individualistic cultures.
16-House.qxd 3/19/2004 5:00 PM Page 484
Institutional Institutional
Human Condition Collectivism Practices Collectivism Values
In-Group In-Group
Human Condition Collectivism Practices Collectivism Values
many cultures with high scores on In-Group society was indexed by items regarding the
Collectivism are among the least wealthy and, extent to which family is important in life,
therefore, have less access to medical care. people are motivated to make their parents
Alternatively, these findings may be due to the proud, and parents are supposed to make sacri-
fact that the construct of the human condition is fices for their children. Respect for family and
itself partly cultural and in some cases Western. friends refers to beliefs about loving and respect-
For example, an indicator of societal health in ing one’s parents unconditionally, exerting effort
this analysis refers to “personal security and to live up to the expectations of friends, and liv-
private property,” which reflects a Western ing with one’s parents. As can be seen in Table
notion of health that may be inappropriate for 16.11g, Institutional Collectivism values are
evaluating societal health in collectivistic cul- positively correlated with the strength of family
tures. If strength of family ties were included as ties scale. Likewise, as shown in Table 16.11h,
an indication of the human condition (as will be In-Group Collectivism practices and values are
discussed below), then we would see that col- positively related to strength of family ties, and
lectivist cultures fare better on other indicators. In-Group Collectivism practices are also corre-
Likewise, alcohol and drug abuse in the work- lated with respect for family and friends. Not
place may not be a problem in many develop- surprisingly, this is consistent with the con-
ing countries, but malaria, ailments due to the structs and with previous research on the impor-
poor quality of drinking water, and the avail- tance of family ties in many collectivist cultures
ability of hygiene facilities may be more crucial (Triandis, 1995).
issues in these countries. As such, it may be GLOBE Individualism–Collectivism and
more meaningful to look at specific health Achievement. This section involves correlations
issues and collectivism rather than general among collectivism, achieving results, and ini-
indices of health. tiative. Achievement is defined in terms of the
amount of freedom and control that people feel
GLOBE Individualism–Collectivism and Family they have in the course of life, the extent to
and Friends. The strength of family ties in a which people persevere until they are satisfied
16-House.qxd 3/19/2004 5:00 PM Page 486
Institutional Institutional
Family and Friends Collectivism Practices Collectivism Values
In-Group In-Group
Family and Friends Collectivism Practices Collectivism Values
with the results of a task, and the importance after controlling for GNP per capita. This
that people place on doing a job that provides suggests that it is GNP per capita, and not
a sense of accomplishment. Initiative is a collectivism, that explains achieving results.
measure of the importance that people place Because GNP per capita is an index of
on being able to use initiative in their jobs and societal industrialization (and, thus, indirectly
on having a responsible job. As Table 16.11i a measure of achieving results), this makes
shows, Institutional Collectivism values are intuitive sense.
positively correlated with Initiative. This rela- GLOBE Individualism–Collectivism and
tionship may be explained by the tendency for Political Ideology. In this section we explore the
cultures that endorse Institutional Collectivism relationship between collectivism and six facets
values to also believe that they should be more of political ideology. These are described in
Performance Oriented, as discussed previously Chapter 7 by Javidan and Hauser and consist of
in the section on correlations with other disdain for democracy, passiveness, lack of
GLOBE dimensions. voice, dislike for democracy, role of govern-
As shown in Table 16.11j, In-Group ment, and stability. Similar to what has been
Collectivism values are significantly positively found in other sections, Institutional Collecti-
correlated with achieving results, although we vism has fewer significant correlations with
note this correlation becomes nonsignificant the political ideology variables than does
16-House.qxd 3/19/2004 5:00 PM Page 487
Institutional Institutional
Collectivism Practices Collectivism Values
In-Group In-Group
Collectivism Practices Collectivism Values
In-Group Collectivism. The only significant injustices because of their obligations to famil-
correlation in Table 16.11k—the positive corre- ial needs such as the need for a job to support
lation between passiveness and Institutional the family. Likewise, having voice in govern-
Collectivism values—becomes nonsignificant mental decisions and the community is likely to
after controlling for GNP. be a less salient issue in these societies, given
Table 16.11l shows that In-Group Collec- that the focus is on more immediate in-groups
tivism practices are positively correlated with and their survival.
passiveness, lack of voice, role of government, GLOBE Individualism–Collectivism and
and stability, although the correlation involving Religion. As discussed previously, aspects of
stability becomes nonsignificant after control- collectivism such as strong group identity have
ling for GNP. The positive correlation of In- been evident in religions across the globe for
Group Collectivism practices and passiveness centuries. In this section, we explore the rela-
illustrates that the phenomena of signing tionship between Individualism–Collectivism
petitions, boycotting, attending demonstrations, and religiousness in modern-day societies. The
and joining unofficial strikes are less frequent components of religiousness are described in
in these societies. It is possible that clusters Chapter 7 by Javidan and Hauser. These compo-
with respondents that scored high on In-Group nents are religious devotion and religious
Collectivism practices are likely to tolerate dogma.
16-House.qxd 3/19/2004 5:00 PM Page 488
Institutional Institutional
Political Ideology Collectivism Practices Collectivism Values
In-Group In-Group
Political Ideology Collectivism Practices Collectivism Values
Institutional Institutional
Religious Ideology Collectivism Practices Collectivism Values
Religious devotion −.24 .38*
N = 38 N = 38
Religious dogma .11 .22
N = 37 N = 37
In-Group In-Group
Religious Ideology Collectivism Practices Collectivism Values
Religious devotion .49** .38*
N = 38 N = 38
Religious dogma .49** .35*
N = 37 N = 37
As seen in Table 16.11m, the only significant Collectivism practices are negatively correlated
correlation is found between Institutional with gender equality. In-group-oriented societies,
Collectivism values and religious devotion, such as those found in the Middle East and
although this positive correlation becomes non- Southern Asia, tend to be male dominated.
significant after controlling for GNP. Table
16.11n shows that both In-Group Collectivism
Summary: Societal Collectivism
practices and values are significantly and posi-
tively correlated with religious devotion and As the cumulative findings in this section
religious dogma. After controlling for GNP, reveal, the Institutional Collectivism and In-
however, all of these correlations become Group Collectivism scales measure different
nonsignificant. aspects of collectivism. In particular, the two
scales show different patterns of correlations
GLOBE Individualism/Collectivism and Gender with other GLOBE scales and existing collec-
Equality. Tables 16.11o and 16.11p present tivism scales in the literature, different country
results related to Collectivism and gender equal- and cluster rankings, and different relationships
ity. Gender equality is discussed in Chapter 7 with external data. Likewise, values and prac-
by Javidan and Hauser. As can be seen from tices measure different components and, indeed,
these tables, gender equality as reported by the are negatively related for Institutional Collecti-
World Values Survey is not correlated with vism. Although institutional collectivism con-
Institutional Collectivism. In contrast, In-Group structs have traditionally been seen as broad
16-House.qxd 3/19/2004 5:00 PM Page 490
Institutional Institutional
Collectivism Practices Collectivism Values
In-Group In-Group
Collectivism Practices Collectivism Values
and complex (Triandis, 1995), they are often telecommunication industries, although the
discussed as unitary in the literature. As such, differences in means were marginally signifi-
GLOBE findings provide additional needed cant only for the In-Group Collectivism prac-
complexity that can help researchers understand tices (F = 1.38, p > .10 and F = 2.87, p > .05,
and differentiate among different “syndromes” respectively). For Institutional Collectivism val-
of collectivism. Next we turn to GLOBE results ues, the financial industry had a lower mean
at the organizational level, before turning to than the telecommunications industry (5.04)
GLOBE leadership results. and the food industry (5.00), and these differ-
ences were significant (F = 4.15, p < 0.05). For
In-Group Collectivism values, the financial
GLOBE RESULTS: services had a higher mean than the other two
ORGANIZATIONAL LEVEL industries, but these differences were not statis-
tically significant (F = 1.51, p > .10). As a gen-
Descriptive Statistics eral pattern, it is interesting to note that the
food industry tends to have the most collectivist
Table 16.12 presents the summary statistics practices; in contrast, the financial industry
for the Organizational Collectivism I (Institu- tends to have the most Collectivism In-Group
tional Collectivism) and Organizational Collec- values.
tivism II (In-Group Pride and Loyalty) practices
and values scales, across all 62 cultures that par-
Summary: Organizational Collectivism
ticipated in the GLOBE study. As with other
chapters in this book, the results are separated The findings reveal that at the organizational
by industry. For both Institutional and In-Group level, both measures of Collectivism tend to be
Collectivism practices, the food industry had a valued more than they are practiced. For both
higher mean than the financial services and scales, the means on the Collectivism values
16-House.qxd 3/19/2004 5:00 PM Page 491
Table 16.12 Summary Statistics for Organizational Institutional and In-Group Collectivism for Three
Industries
Standard
Variable n Minimum Maximum Mean Deviation
Financial Industry
Institutional Collectivism practices 130 1.38 5.57 4.18 0.60
Institutional Collectivism values 130 3.83 5.83 4.85 0.41
In-Group Collectivism practices 130 2.85 5.69 4.64 0.51
In-Group Collectivism values 130 3.56 6.58 5.66 0.57
Food Industry
Institutional Collectivism practices 91 2.17 5.20 4.31 0.50
Institutional Collectivism values 91 3.89 6.14 5.00 0.50
In-Group Collectivism practices 91 3.00 6.00 4.79 0.50
In-Group Collectivism values 91 3.57 6.41 5.52 0.58
Telecommunications Industry
Institutional Collectivism practices 53 2.90 5.82 4.25 0.61
Institutional Collectivism values 53 3.68 6.00 5.04 0.57
In-Group Collectivism practices 53 3.48 5.56 4.62 0.54
In-Group Collectivism values 53 3.76 6.50 5.58 0.64
are consistently higher than the means for Specifically, practices for societal Institutional
the Collectivism practices scales, and this find- Collectivism have a significant and strong
ing is constant across all three industries (see positive relationship with organizational
Table 16.12). Institutional Collectivism practices ( p < .01).
Unexpectedly, In-Group Collectivism practices
at the societal level were not related to In-Group
Collectivism practices at the organizational
GLOBE RESULTS: RELATIONSHIP level. However, we found a significant and even
BETWEEN SOCIETAL AND stronger relationship between societal Collecti-
ORGANIZATIONAL PRACTICES AND vism values and organizational Collectivism
SOCIETAL AND ORGANIZATIONAL VALUES values (p < .01) for both Institutional and In-
Group Collectivism.
As indicated in Chapter 2, the GLOBE theoreti- Therefore, in three of four tests, our analyses
cal model postulates that societal practices and support a principal proposition in the GLOBE
values affect organizational practices and theoretical model (i.e., Proposition 3, Figure 2.1,
values. For this chapter, we conducted four Chapter 2, by House & Javidan); societal
hierarchical linear models (HLMs) to test these cultural values and practices affect organiza-
hypotheses for organizational Individual– tional cultural values and practices.7 This is
Collectivism practices and values. We tested the particularly true for values. In both tests involv-
GLOBE hypothesis regarding the effect of ing values, societal Institutional Collectivism
societal culture on organizational culture by values are associated with organizational
conducting HLM analyses in which organiza- Collectivism values. Thus, organizations in
tional Collectivism was predicted by societal societies with high scores on institutional col-
Collectivism (for both Institutional and In- lectivist values tend to emphasize collective
Group Collectivism). These analyses supported rewards over individual rewards. The results
our hypotheses in three of the four tests. also indicate that the more a society values
16-House.qxd 3/19/2004 5:00 PM Page 492
Charismatic/Value-Based
(second order)
Charismatic I: Visionary Foresight, intellectually stimulating, future-oriented, prepared,
anticipatory, plans ahead, inspirational, visionary, able to
anticipate
Charismatic II: Inspirational Enthusiastic, positive, encouraging, morale booster, motive
arouser, confidence builder, dynamic, motivational
Charismatic III: Self-sacrifice Risk taker, self-sacrificial, convincing
Integrity Honest, sincere, just trustworthy
Decisive Willful, decisive, logical, intuitive
Performance oriented Improvement-oriented, excellence-oriented, performance-oriented
In-Group Collectivism, the more organizations societal and organizational levels of analysis.
within that society are reported to practice On the basis of our previous review (see
organizational In-Group Collectivism. Table 16.3), we generally expected that societal
Finally, although not illustrated in these collectivism would be negatively related to
tables, it is important to note that the societal Autonomous leadership and positively related
scales for Collectivism are highly correlated to Team-Oriented and Charismatic/Value-
across industries, demonstrating that societal Based leadership. At the same time, we also
scores on these scales are generalizable across expected that different forms of collectivism
industries (see Hanges & Dickson, Chapter 8, might have different leadership correlates. We
this volume, for statistical details). also examined whether the effects would be
found at the organizational level of analysis.
The measures for the leadership attributes
that were of primary interest (Autonomous
GLOBE FINDINGS: INDIVIDUALISM leadership, Team-Oriented leadership, and
AND COLLECTIVISM AND IMPLICIT Charismatic/Value-Based leadership) are pre-
THEORIES OF EFFECTIVE LEADERSHIP sented in Table 16.13. We will also provide a
discussion of the relationship between
In this last section, we present GLOBE find- Individualism–Collectivism and other leadership
ings that relate Individualism–Collectivism to attributes (i.e., Humane-Oriented, Self-Protective,
reported effective leadership dimensions at the and Participative leadership) for exploratory
16-House.qxd 3/19/2004 5:00 PM Page 493
purposes (see Chapter 8 by Hanges & Dickson Team-Oriented leadership was positively
for specific measures of these dimensions). correlated with Institutional Collectivism values
(r = .26, p < .05) and In-Group Collectivism
practices (r = .30, p < .05) and values (r = .37,
Culture Cluster Rankings for
p < .01). In other words, those cultures in which
Predicted Culturally Endorsed respondents value collective goals and interests,
Leadership Theories and practice and value family pride and interde-
As discussed in Chapter 8, there are signifi- pendence are more likely to endorse collabora-
cant differences in values and practices for every tive and team-integrating leader behaviors as
cultural dimension among the 61 societal cul- effective behaviors.
tures (i.e., recall that GLOBE uses the term soci- The leader behaviors listed in Table 16.13
etal cultures rather than nations). Table 16.14 comprising Team-Oriented leadership are
presents the average perceived effectiveness perceived to be in the service of collectivistic
scores for each cluster for the three leadership practices and norms. Finally, Charismatic/
dimensions. Analyses of variance were con- Value-Based leadership was positively corre-
ducted to evaluate whether the differences across lated with In-Group Collectivism values
clusters are statistically significant. Results (r = .32, p < .05). This relationship suggests that
indicate that cluster differences are significant charismatic leadership is more highly endorsed
for Charismatic/Value-Based, Team-Oriented, by managers in those cultures that value family
and Autonomous leadership (p < .05). Post hoc pride and interdependence. Those societies in
analyses demonstrate that the Latin America which managers express a desire for increased
cluster has a higher score for Team-Oriented family pride and interdependence are more
leadership than all others, which, in turn, have receptive to charismatic leadership, and this
higher scores than the Middle East. For finding supports hypotheses that have been
Charismatic/Value-Based leadership, the Anglo, advanced in the literature (e.g., Jung et al.,
Latin America, and Southern Asia clusters have 1995; Meindl, 1995; Pillai & Meindl, 1998) and
the highest scores and the Middle East the low- that were discussed in previous sections.
est scores. For the Autonomous dimension, the Indeed, the emergence of charismatic leadership
Eastern Europe cluster has the highest scores and appears to be more likely in societies with high
Latin America has the lowest. In-Group Collectivism values, presumably
because the readiness for charismatic leadership
is a function of the degree of interconnectedness
Correlation Between Societal within workgroups, which, in turn, is assumed
Collectivism and Leadership Dimensions to be a function of societal collectivism. In the
Table 16.15 displays the correlations following section, we report the results of the
between the leadership dimensions and the four more sophisticated HLM analyses to examine
societal Collectivism scales. As the table illus- the relationships between culture and culturally
trates, the correlations are generally consistent endorsed leadership theories (CLTs).
with theoretical expectations. The only sig-
nificant correlation involving Autonomous
Hierarchical Linear Modeling Analyses
leadership, however, was with Institutional
Collectivism values, and this correlation was We tested for the relationship between
negative (r = −.48, p < .01) as hypothesized. culture and the CLT leadership dimensions
This indicates that managers in those cultures by using hierarchical linear modeling. An
that desire to emphasize collective goals and overview of HLM analyses and a detailed dis-
interests more than they currently do are less cussion of how we conducted these analyses are
likely to emphasize Autonomous leadership. provided in Chapter 11 by Hanges, Dickson,
This relationship makes conceptual sense, as and Sipe. This present chapter presents relation-
independent, individualistic, unique, and ships between one culture dimension (Individu-
autonomous leader behaviors are incongruent alism and Collectivism) and the six CLTS.
with the promotion of collective interests. Competitive tests of all culture dimensions and
16-House.qxd 3/19/2004 5:00 PM Page 494
Table 16.14 Perceived Effectiveness Scores for Leadership Scales Based on Culture Clusters
Colombia
Costa Rica
Ecuador
El Salvador
Guatemala
Mexico
Venezuela
CLTs are presented in Chapter 21 by Dorfman read through the results discussed below, it
and colleagues.8 In general, as Dorfman argues may be helpful to view Figures 16.1 and 16.2
in Chapter 21, GLOBE expected that societal for a visual summary. These figures, however,
and organizational values will be more strongly show only results regarding cultural values,
related to CLT leadership dimensions than soci- not practices. (All HLM coefficients are
etal and organizational practices. (See Dorfman presented in Table 21.10 of Chapter 21 by
argument in Chapter 21.) As indicated previ- Dorfman et al.)
ously, our notions of values and CLT leadership Specifically, we explored whether Collecti-
dimensions represent idealized concepts of vism values and practices were related to
how the world Should Be in contrast to prac- Team-Oriented, Autonomous, Charismatic/
tices, which represent the world As Is. As you Value-Based, Self-Protective, Participative,
16-House.qxd 3/19/2004 5:00 PM Page 496
Figure 16.1 Institutional Collectivism Cultural Values as Drivers for CLT Leadership Dimensions
* Only statistically significant relationships for values (not practices) are shown (p < .05; see Table 21.10, Chapter 21 by
Dorfman et al.)
CLT = Culturally endorsed leadership theory
O = Organizational level
S = Societal level
Charismatic (S,O)
• Organizational loyalty
• Pride in In-Group
+ Team Oriented (O)
Figure 16.2 In-Group Collectivism Cultural Values as Drivers for CLT Leadership Dimensions
* Only statistically significant relationships for values (not Practices) are shown (p < .05; see Table 21.10, Chapter 21 by
Dorfman et al.). The most important leadership CLT relationships are bolded (i.e., relationship is significant at both society
and organization levels of analyses or highest HLM coefficient within each level of analysis). Results for Practices are
described in the text.
CLT = Culturally endorsed leadership theory
O = Organizational level
S = Societal level
5.9% of the organizational and societal variance organizational level. The organizational
for this dimension. All of this explained variance Institutional Collectivism practices scores were
was associated with forces operating at the orga- significantly positively related (p < .01) to the
nizational level. The organizational Institutional Team-Oriented leadership dimension, as were
Collectivism practices scores were significantly the Organizational Institutional Collectivism
positively related to the Charismatic/Value- values scores (p < .01). Team-Oriented leader-
Based leadership dimension (p < .01), as were ship is more likely to be perceived as contribut-
the organizational Institutional Collectivism val- ing to effective leadership in organizations with
ues scores (p < .01). Charismatic/Value-Based high Institutional Collectivism practices and
leadership is more likely to be perceived as con- values scores.
tributing to effective leadership in organizations
• Participative leadership. Institutional
with high Institutional Collectivism practices
Collectivism practices and values scores were
and values scores.
significantly related to the Participative leader-
• Team-Oriented leadership. Institutional ship dimension and explained a total of 6.8% of
Collectivism practices and values scores were the organizational and societal variance. Of the
significantly related to the Team-Oriented lead- total explained variance, 13.6% was associated
ership dimension and explained a total of 6.6% with forces operating at the organizational level
of the organizational and societal variance for of analysis. The remaining explained variance,
this dimension. All of the explained variance 86.4%, was associated with forces operating at
was associated with forces operating at the the societal level of analysis.
16-House.qxd 3/19/2004 5:00 PM Page 498
We suspect that the emphasis on face-saving attributes. Consistent with this, not surprisingly,
in both types of Collectivist cultures may be due the results across levels are very consistent for
to different reasons. For example, it may be due cultural values, but less so for cultural practices.
to the importance of respect for authority in Consistent with a multilevel perspective, the
societies that emphasize In-Group Collectivism, data showed that Collectivism and effective
yet it may be due to the importance of non- leadership were not always consistent across
assertiveness and humane orientation in levels and, in fact, had some notable reversals
societies that practice Institutional Collectivism in effects, particularly for cultural practices. For
(i.e., in Scandinavian cultures). One surprising example, although Participative leadership
result, however, is that the less a society values practices were negatively associated with effec-
In-Group Collectivism, the more it views Self- tive leadership at the societal level, they were
Protective leadership as effective. As discussed positively associated with effective leadership
previously, the In-Group Collectivism scale is at the organizational level for In-Group
associated with the syndrome in which societal Collectivism values, as well as Institutional
practices focus on authority differentiation, and Collectivism values.
in which people believe that there should be Likewise, whereas Autonomous leadership
more of a focus on family ties and concern for was negatively associated with effective leader-
others. As such, the desire for such cultural ele- ship at the societal level for Institutional
ments could possibly be served by attention to Collectivism values, the results were mixed at
face-saving. As noted above, these effects are the organizational level. For example, organiza-
based entirely at the societal level of analysis; tions that practice Institutional Collectivism—
Collectivism at the organizational level is not those organizations that focus on group over
related to Self-Protective leadership. individual interests—were less likely to
Second, Institutional and In-Group Collecti- perceive autonomy as being associated with
vism practices at the societal level were consis- effective leadership. However, the more organi-
tently negatively associated with Participative zations practice group pride and loyalty
leadership, and the societal level accounted for (Collectivism II), the more likely Autonomous
the vast majority (greater than 80%) of variance leadership was seen as effective. Because the
in this CLT. These effects illustrate that those latter form of collectivism focuses on having
societies that practice Institutional or In-Group pride in individuals’ accomplishment within
Collectivism are less likely to perceive that groups, it is perhaps understandable that unique-
participativeness is an important component ness and independence in leadership is also
of effective leadership. Again, this may be for embraced in these organizations. However, it is
different reasons, depending on the nature of worth noting that the amount of variance
collectivism. For example, this relationship may explained for this CLT is very low, thus render-
be due to the reliance on individual authority ing these findings not very substantial.
figures in societies that emphasize In-Group Overall, these analyses indicate that the extent
Collectivism, yet it may be due to the reliance to which the collectivistic cultural practices and
on collectivities (e.g., government, top manage- values are emphasized in organizations and
ment groups) to take care of others in societies societies is reflected in the kind of leadership that
that practice institutional collectivism. This is viewed as effective. Our findings support
explanation, of course, needs substantiation in the argument that culture plays a role in influ-
future research. encing the content of leader attributes that are
considered effective. Furthermore, the results
were consistent with predictions regarding
DIFFERENCES IN PERCEPTIONS OF Collectivism values and leadership (see Figures
LEADERSHIP ACROSS LEVELS 16.1 and 16.2). The leadership attributes that
were commonly associated across both types of
As stated previously, Dorfman argues that cul- Collectivism values included Charismatic/
tural values, not practices, are more compatible Value-Based, Team-Oriented, Participative, and
with CLTs as the latter reflect desired leadership Humane-Oriented leadership.
16-House.qxd 3/19/2004 5:00 PM Page 502
The third defining attribute focuses on how the self was referred to as communal sharing), vertical and
is viewed vis-à-vis the larger society, or how the self horizontal relationships (authority ranking and
interacts with the society. Those with independent equality matching, respectively), and individualism
concept of self do what they like to do, or what they (market pricing). He also illustrated that the typology
think is good for them (i.e., they pursue their individual is consistent with Rokeach’s (1973) analysis of politi-
desires, attitudes, values, and beliefs). Because this cal systems. He discussed political systems that highly
meets the need of most of the people in a culture in value both equality and freedom, which correspond to
which most people have an independent concept of self, horizontal individualism (social democracy; e.g.,
the individualistic society values people doing their Australia, Sweden). Systems that he discussed as valu-
own thing. However, people with interdependent con- ing equality but not freedom correspond to our con-
cept of self inherit many relationships and learn to live ceptualization of horizontal collectivism (e.g., the
with these interdependencies. Part of managing the Israeli kibbutz). Those systems that value freedom but
interdependencies is to develop goals that meet the not equality correspond to our notion of vertical indi-
needs of more than one’s own self. In the process of vidualism (e.g., competitive capitalism and market
taking care of the needs of one’s in-group members, a economies such as in the United States). Lastly, those
social mechanism evolves in collectivist cultures that is societies that value neither equality nor freedom
driven by norms. Thus, for those with interdependent correspond to vertical collectivism (e.g., fascism or the
concept of self it is much easier cognitively to resort to communalism of traditional societies with strong
methods that have been tried in the past for interacting leaders) in our conceptualization.
with people at large. A sense of duty guides them 3. We would argue that the terms parochial and
toward social norms in both the workplace and inter- professional have unnecessary pejorative connota-
personal relationships. Individualists, on the other hand, tions. We think many people in collectivist countries
are more concerned about their personal attitudes and will find the labels offensive because it insinuates
values. They care much less than collectivists about that it is parochial not to separate work and family,
what their family members have to say, let alone the and that to do so is professional.
extended family, friends, or neighbors. Hence, the dif- 4. Unfortunately, the terminology allocentrism and
ference between following one’s own attitude versus idiocentrism has not been used widely in the literature.
norms of the society becomes a salient difference Given this, in the rest of this discussion we use the more
between individualist and collectivist cultures. commonly used terms, collectivism and individualism.
The fourth defining attribute focuses on the 5. Our response bias correction procedure identi-
interpersonal relationships, or the nature of social fied response bias in some countries for this scale.
exchange between self and others. When the self is We recomputed the predicted response bias corrected
viewed as independent, interpersonal relationships scale score for each country:
are developed to meet the need of the self to maxi- Institutional Collectivism practices: France, 4.20
mize the benefits to the self. Thus, social exchange is (no change in band); Indonesia, 4.27 (no change in
based on the principle of equal exchange, and people band); Morocco, 4.18 (moves from band C to band
form new relationships to meet their changing needs B); Philippines, 4.37 (moves from band A to band B);
based on cost-benefit analysis. Thus, individualists Qatar, 4.78 (moves from band B to band A); and
are rational in their social exchange. On the other Taiwan, 4.30 (moves from band A to band B).
hand, those with an interdependent concept of self In-Group Collectivism practices: France, 4.66 (no
and relationships that are inherited are likely to view change in band); Morocco, 6.37 (no change in band);
their relationships as long term in nature and, there- Qatar, 5.07 (no change in band).
fore, are unlikely to break a relationship even if it is For a complete discussion of this procedure and
not cost-effective. Thus, collectivists value relation- all response bias corrected scores, see Appendix B.
ships for their own sake, and nurture them with 6. The response bias correction procedure identi-
unequal social exchanges over a long period of time. fied a substantial amount of response bias in some
2. Triandis (1995) illustrated that that this four- countries for this scale. We recomputed the predicted
way typology articulates exceptionally well with some response bias corrected scale score for each country:
of the literature that has examined varieties of cultural Institutional Collectivism values: France, 5.27
patterns. For instance, Fiske (1992) has discussed cul- (moves from band B to band A); and Morocco, 5.34
tural patterns that correspond to collectivism (which (moves from band B to band A).
16-House.qxd 3/19/2004 5:00 PM Page 505
In-Group Collectivism values: Ecuador, 5.81 Ali, A. J., Taqi, A. A., & Krishnan, K. (1997).
(moves from band A to band B); France, 5.88 (no Individualism, collectivism, and decision styles
change in band); Morocco, 6.03 (moves from band B to of managers in Kuwait. Journal of Social
band A); and New Zealand, 6.54 (no change in band). Psychology, 137(5), 629–637.
For a complete discussion of this procedure and Aycan, Z., Kanungo, R. N., Mendonca, M., Yu, K.,
all response bias corrected scores, see Appendix B. Deller, J., Stahl, G., & Kurshid, A. (2000).
7. As reported in Chapter 20 by Brodbeck, Impact of culture on human resource manage-
Hanges, Dickson, Gupta, and Dorfman, we found that ment practices: A 10 country comparison.
all the cultural dimensions of organizational cultural Applied Psychology: An International Review,
values and practices significantly differed across 49(1), 192–221.
societies. Although important, this prior analysis did Aycan, Z., Kanungo, R. N., & Sinha, J. B. P. (1999).
not identify the particular aspect of societal differ- Organizational culture and human resource
ences that was related to organizational culture. In the management practices: The model of culture
present chapter, we found that societal and organiza- fit. Journal of Cross-Cultural Psychology, 30(4),
tional Individual–Collectivism practices were signif- 501–526.
icantly related for Institutional Collectivism only Bakan, D. (1966). The duality of human existence.
(R2 Total = 2.8%, R2 Societal = 56.9%, p < .01). We Chicago: Rand McNally.
found even stronger results for societal and organiza- Bellah, R. N., Madsen, R., Sullivan, W., Swidler, A.,
tional Individual–Collectivism values (R2 Total for & Tipton, S. M. (1985). Habits of the heart:
Institutional Collectivism = 14.2%, R2 Societal = Individualism and commitment in American life.
70.9%, p < .01; R2 Total for In-Group Collectivism = Berkeley: University of California Press.
11.1%, R2 Societal = 52.8%, p < .01). As discussed in Barry, H., III, Child, I. L., & Bacon, M. K. (1959). A
Chapter 11 by Hanges, Dickson, and Sipe, the R2 cross-cultural survey of sex differences in
Total considers all levels of analysis (i.e., individual, socialization. Journal of Abnormal and Social
organizational, and societal) whereas the R2 Societal Psychology, 55, 327–332.
isolates the societal level portion of the dependent Berry, J. W. (1976). Sex differences in behavior and
variable and indicates the percentage of variance cultural complexity. Indian Journal of Psychology,
accounted for by the predictor at only this level. 51, 89–97.
Whereas GLOBE has primarily taken the conserva- Berry, J. W. (1994). Ecology of individualism
tive approach and reported the R2 Total in GLOBE, a and collectivism. In U. Kim, H. C. Triandis,
number of scholars suggest that R2 Societal provides C. Kagitçibasi, S.-C. Choi, & G. Yoon (Eds.),
a more accurate description of aggregated relation- Individualism and collectivism: Theory, method
ships. For further discussion, see the paper by Lance and applications (pp. 77–84). Thousand Oaks,
and James (1999). CA: Sage.
8. Results between the single HLM and multiple Berry, J. W., Poortinga, Y. H., Segall, M. H., &
HLM tests will likely differ somewhat. The differ- Dasen, P. R. (1992). Cross-cultural psychology:
ences between the results of the multiple HLMs and Research and applications. New York: Cambridge
single HLMs are conceptually similar to the differ- University Press.
ences between a multiple regression analysis and a Berry, J. W., Trimble, J. E., & Olmedo, E. L. (1986).
correlation coefficient. Table 21.10 in Chapter 21 by Assessment of acculturation. In W. J. Lonner &
Dorfman et al. presents both single and multiple J. W. Berry (Eds.), Field methods in cross-
HLM coefficients. In addition, the relationships cultural research. Cross-cultural research and
for all culture dimension values are summarized in methodology series (Vol. 8, pp. 85–110).
Chapter 3 by Javidan, House, and Dorfman. Thousand Oaks, CA: Sage.
Bhawuk, D. P. S. (2001). Evolution of culture
assimilators: Toward theory-based assimilators.
REFERENCES International Journal of Intercultural Relations,
25(2), 141–163.
Abu-Saad, I. (1998). Individualism and Islamic work Bhawuk, D. P. S., & Udas, A. (1996). Entrepreneurship
beliefs. Journal of Cross-Cultural Psychology, and collectivism: A study of Nepalese entrepre-
29(2), 377–383. neurs. In J. Pandey, D. Sinha, & D. P. S. Bhawuk
16-House.qxd 3/19/2004 5:00 PM Page 506
Dumont, L. (1986). Essays on individualism. Gaines, S. O., Jr., Marelich, W. D., Bledsoe, K. L.,
Chicago: University of Chicago Press. Steers, W. N., Henderson, M. C., Granrose, C. S.,
Durant, W. (1935). Our oriental heritage. New York: et al. (1997). Links between race/ethnicity and
Simon & Schuster. cultural values as mediated by racial/ethnic identity
Durant, W. (1939). The life of Greece. New York: and moderated by gender. Journal of Personality
Simon & Schuster. and Social Psychology, 72, 1460–1476.
Durkheim, E. (1933). The division of labor. Chicago: Gelfand, M. J., Triandis, H. C., & Chan, K. S. (1996).
Free Press. Individualism versus collectivism or versus
Earley, P. C. (1989). Social loafing and collectivism: authoritarianism? European Journal of Social
A comparison of the U.S. and the People’s Psychology, 26, 397–410.
Republic of China. Administrative Science Georgas, J., Mylonas, K., Bafiti, T., Poortinga, Y. H.,
Quarterly, 34, 565–581. Christakopoulou, S., Kagitçibasi, C., Kwak, K., &
Earley, P. C. (1993). East meets West meets Middle Ataca, B. (2001). Functional relationships in the
East: Further explorations of collectivist and nuclear and extended family: A 16-culture study.
individualist work groups. Academy of Manage- International Journal of Psychology, 36, 289–300.
ment Journal, 36, 319–348. Gerstner, C. R., & Day, D. V. (1994). Cross-cultural
Earley, P. C. (1994). The individual and collective comparison of leadership prototypes. Leadership
self: An assessment of self-efficacy and training Quarterly, 5(2), 121–134.
across cultures. Administrative Science Quarterly, Griffith, S. B. (1963). Sun Tzu: The art of war.
39, 89–117. Oxford: Oxford University Press.
Earley, P. C., & Gibson, C. B. (1998). Taking stock Gudykunst, W. B., Gao, G., & Franklyn-Stokes, A.
in our progress on individualism and collec- (1996). Self-monitoring and concern for social
tivism: 100 years of solidarity and community. appropriateness in China and England. In
Journal of Management, 24, 265–304. J. Pandey, D. Sinha, & D. P. S. Bhawuk (Eds.),
Eby, L. T., & Dobbins, G. H. (1997). Collectivistic Asian contributions to cross-cultural psychol-
orientation in teams: An individual and group- ogy (pp. 255–267). New Delhi: Sage.
level analysis. Journal of Organizational Gudykunst, W. B., Gao, G., Schmidt, K. L. Nishida,
Behavior, 18(3), 275–295. T., Bond, M. H., Leung, K., et al. (1992). The
Encyclopedia Britannica. (1953). Vol. 12 (p. 256a). influence of individualism-collectivism, self-
Erez, M. (1994). Toward a model of cross-cultural monitoring, and predicted outcome value on
industrial and organizational psychology. In communication in in-group and out-group rela-
H. Triandis, M. Dunnette, & L. Hough (Eds.), tionships. Journal of Cross-Cultural Psychology,
Handbook of industrial and organizational psy- 23, 196–213.
chology (2nd ed., Vol. 4, pp. 559–607). Palo Hackman, J. R., & Oldham, G. R. (1976). Motivation
Alto, CA: Consulting Psychologists Press. through the design of work: Test of a theory.
Erez, M., & Earley, P. C. (1993). Culture, self-identity, Organizational Behavior and Human Decision
and work. New York: Oxford University Press. Processes, 16(2), 250–279.
Fiske, A. P. (1992). The four elementary forms of Hall, E. T. (1959). The silent language. New York:
sociality: Framework for a unified theory of Doubleday.
social relations. Psychology Review, 99, Hanges, P. J., Lord, R. G., & Dickson, M. W. (2000).
689–723. An information-processing perspective on leader-
Freeman, M. A. (1997). Demographic correlates of ship and culture: A case for connectionist archi-
individualism and collectivism: A study of social tecture. Applied Psychology: An International
values in Sri Lanka. Journal of Cross-Cultural Review, 49(1), 133–161.
Psychology, 28(3), 321–341. Herzberg, F., Mausner, B., & Snyderman, B. B. (1959).
Gabriel, S., & Gardner, W. L. (1999). Are there “his” The motivation to work. New York: John Wiley.
and “hers” types of interdependence? The impli- Hofstede, G. (1980). Culture’s consequences. Beverly
cations of gender differences in collective versus Hills, Sage.
relational interdependence for affect, behavior, Hofstede, G. (1994). Foreword. In U. Kim, H.
and cognition. Journal of Personality and Social Triandis, C. Kagitçibasi, S.-G. Choi, & G. Yoon
Psychology, 77(3), 642–655. (Eds.), Individualism and collectivism: Theory,
16-House.qxd 3/19/2004 5:00 PM Page 508
methods, and applications (pp. ix–xiii). Thousand of transformational leadership and collectivistic
Oaks, CA: Sage. cultures. Journal of Leadership Studies, 2, 3–18.
Hofstede, G. (1997). Cultural constraints in manage- Kagan, D. (1966). The ancient near east and Greece.
ment theories. In R. P. Vecchio (Ed.), Leadership: London: The Macmillan Group, Ltd.
Understanding the dynamics of power and influ- Kagitçibasi, C. (1997). Individualism and collectivism.
ence in organizations (pp. 465–483). Notre Dame, In J. W. Berry, M. H. Segall, & C. Kagitçibasi
IN: University of Notre Dame Press. (Eds.), Handbook of cross-cultural psychology
Hofstede, G. (2001). Culture’s consequences (2nd (2nd ed., pp. 1–50) Boston: Allyn & Bacon.
ed.). Thousand Oaks, CA: Sage. Kanungo, R. N., Aycan, Z., & Sinha, J. B. P. (1999).
Hofstede, G., & Bond, M. H. (1988). The Confucius Organizational culture and human resource
connection: From cultural roots to economic management practices: The model of culture fit.
growth. Organizational Dynamics, 16, 4–21. Journal of Cross-cultural Psychology, 30(4),
Hofstede, G., Neuijen, B., Ohayv, D. D., & Sanders, 501–526.
G. (1991). Measuring organizational cultures. Kanungo, R. N., & Jaeger, A. M. (1990). Introduc-
Administrative Science Quarterly, 35, 286–316. tion: The need for indigenous management in
Holtgraves, T. (1997). Styles of language use: developing countries. In A. M. Jaeger & R. N.
Individual and cultural variability in conversa- Kanungo (Eds.), Management in developing
tional indirectness. Journal of Personality and countries (pp. 1–23). London: Routledge.
Social Psychology, 73, 624–637. Kashima, Y., & Callan, V. J. (1994). The Japanese
House, R. J., Wright, N. S., & Aditya, R. N. (1997). workgroup. In H. C. Triandis, M. D. Dunnette,
Cross-cultural research on organizational lead- & L. M. Hough (Eds.), Handbook of industrial
ership: A critical analysis and a proposed theory. and organizational psychology (2nd ed., Vol. 4,
In P. C. Earley & M. Erez (Eds.), New perspec- pp. 606–649). Palo Alto, CA: Consulting
tives on international industrial and organiza- Psychologists Press.
tional psychology (pp. 535–625). San Francisco, Kashima, Y. Y., & Kashima, E. M. (1998). Culture
CA: Lexington Press. and language: The case of cultural dimensions
Hui, C. H. (1988). Measurement of individualism- and personal pronoun use. Journal of Cross-
collectivism. Journal of Research on Personality, Cultural Psychology, 29(3), 461–486.
22, 17–36. Kashima, E. S., & Kashima, Y. (1999). Negotiation
Hui, C. H., & Triandis, H. C. (1986). Individualism- of the self in interpersonal interactions: Japan-
collectivism: A study of cross-cultural researchers. Australia comparison. In J. Adamopoulos & Y.
Journal of Cross-Cultural Psychology, 20, Kashima (Eds.), Social psychology and the cul-
296–309. tural context (pp. 189–202). Thousand Oaks,
Hui, C. H., Yee, C., & Eastman, K. L. (1995). The CA: Sage.
relationship between individualism-collectivism Kashima, Y., Yamaguchi, S., Kim, U., Choi, S.-C.,
and job satisfaction. Applied Psychology: An Gelfand, M., & Yuki, M. (1995). Culture, gen-
International Review, 44, 276–282. der and self: A perspective from individualism-
Huismans, S., & Schwartz, S. H. (1992). Religiosity collectivism research. Journal of Personality
and value priorities: A study of Protestants, and Social Psychology, 59, 925–937.
Catholics, and Jews. In I. Saburo & Y. Kashima Kateb, G. (1992). The inner ocean: Individualism
(Eds.), Innovations in cross-cultural psychology and democratic culture. Ithaca, NY: Cornell
(pp. 237–249). Bristol, UK: Swets & Zeitlinger. University Press.
Huntington, S. P. (1996). The clash of civilizations Kennedy, P. (1987). The rise and fall of the great
and the remaking of world order. New York: powers: Economic change and military conflict
Simon & Schuster. from 1500 to 2000. New York: Random House.
Iyengar, S. S., & Lepper, M. R. (1999). Rethinking Kim, U. (1994). Individualism and collectivism:
the value of choice: A cultural perspective on Conceptual clarification and elaboration. In
intrinsic motivation. Journal of Personality and U. Kim, H. Triandis, C. Kagitçibasi, S.-G.
Social Psychology, 76, 349–366. Choi, & G. Yoon (Eds.), Individualism and col-
Jung, D., Bass, B., & Sosik, J. (1995). Bridging lead- lectivism: Theory, methods, and applications
ership and culture: A theoretical consideration (pp. 19-41). Thousand Oaks, CA: Sage.
16-House.qxd 3/19/2004 5:00 PM Page 509
Kim, U., Triandis, H. C., Kagitçibasi, C., Choi, S.-C., Mead, M. (1961). Cooperation and competition
& Yoon, G. (Eds.). (1994). Individualism and among primitive peoples. Boston, MA: Beacon
collectivism: Theory, method, and applications. Press.
Thousand Oaks, CA: Sage. Meindl, J. R. (1995). The romance of leadership as a
Klein, K. J., Dansereau, F., & Hall, R. J. (1994). follower-centric theory: A social constructionist
Levels issues in theory development, data col- approach. Leadership Quarterly, 6(3), 329–341.
lection, and analysis. Academy of Management Miller, J. G. (1984). Culture and the development of
Review, 19, 195–229. everyday social explanation. Journal of
Kluckhohn, F., & Strodtbeck, F. (1961). Variations in Personality and Social Psychology, 46,
value orientation. Evanston, IL: Row, Peterson. 961–978.
Lance, C. E., & James, L. R. (1999). A proportional Miller, J. G. (1994). Cultural diversity in the morality
variance accounted for index for some cross- of caring: Individually-oriented versus duty-
level and person-situation research designs. oriented interpersonal codes. Cross-Cultural
Organizational Research Methods, 2, 395–418. Research, 28, 3–39.
Lester, D. (1996). Individualism and divorce. Misumi, J. (1985). The behavioral science of leader-
Psychological Reports, 76(1), 258. ship: An interdisciplinary Japanese research pro-
Levine, R., Sato, S., Hashimoto, T., & Verma, J. (1995). gram. Ann Arbor: University of Michigan Press.
Love and marriage in eleven cultures. Journal of Mobley, W. H. (1977). Intermediate linkages in the
Cross-Cultural Psychology, 26(5), 554–571. relationship between job satisfaction and employee
Levine, R. V., & Norenzayan, A. (1999). The pace of turnover. Journal of Applied Psychology, 62(2),
life in 31 countries. Journal of Cross-Cultural 237–240.
Psychology, 30, 178–192. Moorman, R. H., & Blakely, G. L. (1995).
Ling, W. (1989). Pattern of leadership behavior Individualism–collectivism as an individual dif-
assessment in China. Journal of Psychology in ference predictor of organizational citizenship
the Orient, 32(2), 129–134. behavior. Journal of Organizational Behavior,
Lord, R. G., & Maher, K. J. (1991). Leadership and 16(2), 127–142.
information processing: Linking perceptions Morris, M., Avila, R. A., & Allen, J. (1993).
and performance. Cambridge, MA: Unwin Individualism and the modern corporation:
Hyman. Implications for innovation and entrepreneur-
Markus, H., & Kitayama, S. (1991). Culture and self: ship. Journal of Management, 19(3), 595–612.
Implications for cognition, emotion and motiva- Muchinsky, P. M. (2000). Psychology applied to
tion. Psychological Review, 98, 224–253. work: An introduction to industrial and organi-
Markus, H. R., Kitayama, S., & Heiman, R. J. (1997). zational psychology (6th ed.). Belmont, CA:
Culture and “basic” psychological principles. In Wadsworth.
E. T. Higgins & A. W. Kruglanski (Eds.), Social Oishi, S., Schimack, V., Diener, E., & Suh, E. M.,
psychology: Handbook of basic principles (1998). The measurement of values and
(pp. 857–913). New York: Guilford. individualism-collectivism. Personality and
Marmot, M. G., & Syme, S. L. (1976). Acculturation Social Psychology Bulletin, 24, 1177–1189.
and coronary heart disease in Japanese- Oyserman, D. (1993). The lens of personhood:
Americans. American Journal of Epidemiology, Viewing the self, and others, in a multicultural
104(3), 225–247. society. Journal of Personality and Social
Marshall, R. (1997). Variances in levels of individu- Psychology, 65, 993–1009.
alism across two cultures and three social Oyserman, D., Coon, H. M., & Kemmelmeier, M.
classes. Journal of Cross-Cultural Psychology, (2002). Rethinking individualism and collec-
28, 490–495. tivism: Evaluation of theoretical assumptions and
Matsui, T., Kakuyama, T., & Onglatco, M. U. (1987). meta-analyses. Psychological Bulletin, 128, 3–72.
Effects of goals and feedback on performance in Parsons, T. (1949). Essays in sociological theory;
groups. Journal of Applied Psychology, 72(3), Pure and applied. New York: Free Press.
407–415. Parsons, T., & Shils, E. A. (1951). Toward a general
McClelland, D. (1967). The achieving society. theory of action. Cambridge, MA: Harvard
New York: Free Press. University Press.
16-House.qxd 3/19/2004 5:00 PM Page 510
Peng, K., & Nisbett, R. E. (1999). Culture, dialectics, Rotter, J. (1966). Generalized expectancies for
and reasoning about contradiction. American internal versus external control of reinforce-
Psychologist, 54, 741–754. ment. Psychological Monograph, 80(1), 1–28.
Pillai, R., & Meindl, J. R. (1998). Context and Rousseau, D. (1995). Psychological contracts in
charisma: A “meso” level examination of the organizations: Understanding written and unwrit-
relationship of organic structure, collectivism, ten agreements. Thousand Oaks, CA: Sage.
and crisis to charismatic leadership. Journal of Schein, E. H. (1992). Organizational culture and
Management, 24(5), 643–671. leadership. San Francisco: Jossey-Bass.
Podsakoff, P. M., Dorfman, P. W., Howell, J. P., & Schwartz, S. H. (1992). Universals in the content and
Todor, W. D. (1986). Leader reward and structure of values: Theoretical advances and
punishment behaviors: A preliminary test of a empirical tests in 20 countries. In M. Zanna (Ed.),
culture-free style of leadership effectiveness. Advances in experimental social psychology,
Advances in International Comparative Manage- (Vol. 25, pp. 1–66). New York: Academic Press.
ment, 2, 95–138. Schwartz, S. H. (1994). Beyond individualism
Putnam, R. D. (2000). Bowling alone: The collapse and collectivism: New cultural dimensions
and revival of American community. New York: of values. In U. Kim, H. C. Triandis,
Simon & Schuster. C. Kagitçibasi, S.-C. Choi, & G. Yoon (Eds.),
Ramamoorthy, N., & Carroll, S. J. (1998). Individualism and collectivism: Theory,
Individualism/collectivism orientations and method, and applications (pp. 85–122).
reactions toward alternative human resource Newbury Park, CA: Sage.
management practices. Human Relations, 51(5), Schwartz, S. H., & Bilsky, W. (1990). Toward a
571–588. theory of universal structure and content of val-
Realo, A., Allik, J., & Vadi, M. (1997). The hierar- ues: Extensions and cross-cultural replications.
chical structure of collectivism. Journal of Journal of Personality and Social Psychology,
Research in Personality, 31(1), 93–116. 58, 878–891.
Redfield, R. (1956). Peasant society and culture: An Segall, M. H., & Kagitçibasi, C. (1997). Introduction.
anthropological approach to civilization. In J. W. Berry, M. H. Segall, & K. Kagitçibasi
Chicago: University of Chicago Press. (Eds.), Handbook of cross-cultural psychology
Redding, S. G., Norman, A., & Schlander, A. (1994). (Vol. 3, pp. xxv–xxxv). Needham, MA: Allyn &
The nature of individual attachment to the Bacon.
organization: A review of East Asian variations. Shaw, J. B. (1990). A cognitive categorizations
In H. C. Triandis & M. D. Dunnette (Eds.), model for the study of intercultural manage-
Handbook of industrial and organizational ment. Academy of Management Review, 12,
psychology (2nd ed., Vol. 4, pp. 647–688). Palo 626–645.
Alto, CA: Consulting Psychologists Press. Sheppard, B. H., & Sherman, D. M. (1998). The
Reykowski, J. (1994). Collectivism and individualism grammars of trust: A model and general impli-
as dimensions of social change. In U. Kim, H. C. cations. Academy of Management Review, 23(3),
Triandis, C. Kagitçibasi, S.-C. Choi, & G. Yoon 422–437.
(Eds.), Individualism and collectivism: Theory, Singelis, T. M. (1994). The measurement of inde-
method, and applications (pp. 276–292). pendent and interdependent self-construals.
Newbury Park, CA: Sage. Personality and Social Psychology Bulletin, 20,
Rhee, E., Uleman, J. S., & Lee, H. K. (1996). 580–591.
Variations in collectivism and individualism Singelis, T. M., Triandis, H. C., Bhawuk, D. P. S., &
by in-group and culture: Confirmatory factor Gelfand, M. (1995). Horizontal and vertical
analyses. Journal of Personality and Social dimensions of individualism and collectivism: A
Psychology, 71, 1037–1054. theoretical measurement refinement. Cross-
Riesman, D., Denny, R., & Glazer, N. (1950). The Cultural Research, 29, 240–275.
lonely crowd: A study of the changing American Sinha, D. (1988). The family scenario in a developing
character. New Haven, CT: Yale University Press. country and its implications for mental health:
Rokeach, M. (1973). The nature of human values. The case of India. In P. R. Dasen, J. W. Berry, &
New York: Free Press. N. Sartorius (Eds.), Health and cross-cultural
16-House.qxd 3/19/2004 5:00 PM Page 511
Triandis, H. C., McCusker, C., & Hui, C. H. (1990). Context-specific measurement of individualism-
Multimethod probes of individualism and collectivism on the individual level: The
collectivism. Journal of Personality and Social individualism-collectivism interpersonal assess-
Psychology, 74, 118–128. ment inventory. Journal of Cross-Cultural
Triandis, H. C., & Vassiliou, V. A. (1972). Psychology, 28(6), 743–767.
Interpersonal influence and employee selection Wheeler, L., Reis, H. T., & Bond, M. H. (1989).
in two cultures. Journal of Applied Psychology, Collectivism-individualism in everyday social
56, 140–145. life: The Middle Kingdom and the melting pot.
Trompenaars, F. (1985). The organization of meaning Journal of Personality and Social Psychology,
and the meaning of organization: A comparative 57, 79–86.
study of the conceptions of organizational struc- World Bank Annual Report (2001). Retrieved
ture in different cultures. Unpublished doctorial January 9, 2004, from http:www.worldbank.
dissertation, Wharton School of Management, org/annualreport/2001/wbar2001.htm.
University of Pennsylvania, Philadelphia. World Economic Forum. (1979). Report on interna-
Trompenaars, F. (1993). Riding the waves of culture. tional competitiveness. Geneva, Switzerland:
London: Nicholas Brealey. Author.
Vroom, V. H. (1964). Work and motivation. Yamaguchi, S. (1994). Empirical evidence on collec-
New York: John Wiley. tivism among the Japanese. In U. Kim, H. C.
Wagner, J. A., III. (1995). Studies of individualism- Triandis, C. Kagitçibasi, S.-C. Choi, & G. Yoon
collectivism: Effects on cooperation in groups. (Eds.), Individualism and collectivism: Theory,
Academy of Management Journal, 38, 152–170. method, and applications (pp. 175–188).
Wagner, J. A., III, & Moch, M. K. (1986). Newbury Park, CA: Sage.
Individualism-collectivism: Concept and mea- Yu, A.-B., & Yang, K.-S. (1994). The nature of
surement. Group and Organizational Studies, achievement motivation in collectivist societies.
11, 280–304. In U. Kim, H. C. Triandis, C. Kagitçibasi, S.-C.
Wasti, S. A. (2000). Organizational commitment Choi, & G. Yoon (Eds.), Individualism and col-
in a collectivist culture: The case of Turkey lectivism: Theory, method, and applications
[Abstract]. Dissertation Abstracts International (pp. 239–250). Newbury Park: CA: Sage
Section A: Humanities and Social Sciences, Yutang, L. (1938). The wisdom of Confucius.
60(9-A), p. 3547. New York: Modern Library.
Weissman, M. D., Matsumoto, D., Preston, K.,
Brown, B. R., & Kupperbusch, C. (1997).
17-House.qxd 3/19/2004 5:01 PM Page 513
17
POWER DISTANCE
DALE CARL
VIPIN GUPTA
MANSOUR JAVIDAN
513
17-House.qxd 3/19/2004 5:01 PM Page 514
and enactment of power, and cross-cultural individual is enhanced through this expertise. It
researchers, who have explored the existence of is based on the personal power of the individual,
power distance differences across societies. We rather than on power emanating from the posi-
shall discuss these two research traditions sepa- tion itself. In an increasingly technological
rately, although their insights are interrelated in world, this source of power is becoming more
shedding light on this complex and sometimes important within many organizations. Finally,
controversial fact of human existence. referent power refers to a subordinate’s feeling
of oneness with the leader and a desire to iden-
tify with, emulate, and internalize the values of
The Psychological Stream and Power the superior. This is another form of personal
Sources of Power power (Yukl, 2002). Referent power is opera-
tionalized by showing consideration for the
One of the most fundamental questions con- needs and feelings of subordinates, by express-
cerning power is what gives an individual or a ing feelings of trust, acceptance, and concern for
group influence over others? French and Raven their welfare, and by challenging them through
(1959) advanced a five-category classification a personal appeal to move beyond the formal
scheme that identified the various sources or requirements of their roles (Bass, 1985; French &
bases of power that they believed were “espe- Raven, 1959; Hinkin & Schriesheim, 1989;
cially common and important.” They consist of House, 1977; Yukl, 2002).
coercive, reward, legitimate, expert, and refer- Despite the implicit logic of the French
ent power. The coercive power base was defined and Raven typology and its wide use in power
by French and Raven as being dependent on research, there have been several concerns
fear. It rests on the application, or threat of about the lack of conceptual consistency regar-
application, of physical harm or other unpleas- ding the source or origin of the influence
ant outcomes as a form of punishment for the (Hinkin & Schriesheim, 1989; Yukl, 2002).
lack of compliance. Reward power, at the other Over the years, research has indicated
extreme, is based on the opposite motivation: several additional sources of power. First, Burt
People enact positive behaviors to obtain valu- (1992) identified structural holes, defined as the
able rewards controlled by the powerful person. holes in the social structure, as an important
Rewards can be intangible, such as praise, source of power. The structural holes arise if
friendliness, and acceptance, or tangible, such different subgroups of a society are not fully
as increased pay, promotion, or access to infor- interconnected. As a result, some members gain
mation, to name a few. a position of power by virtue of their ability to
Legitimate power is the power that is vested provide a connection among the unconnected
in a person as the result of his or her position subgroups (Burt, 1992). The unique connection
within the formal hierarchy. The subordinate provided by the members in power allows them
responds to a request or demand because the to gain a larger share of the value of goods and
superior has the right to request it and the services exchanged or traded among the various
subordinate has the obligation to comply. subgroups. Second, information generates
Legitimate power has been referred to alterna- power not on the basis of one’s expertise or
tively as “authority” and “position power” knowledge, but because of one’s ability to
(Stogdill, 1974). This power base includes both rapidly and efficiently discover or obtain rele-
coercive and reward power; however, as Yukl vant information and to be a good listener. Third,
(2002) noted, the person in the position must be credibility indicates the respect one commands
perceived as having a legitimate right to the because of repeated interaction and past behav-
position if he or she will be allowed to exert ior (Milgrom & Roberts, 1992). Through their
any influence. Expert power is the ability to behavior, people can accrue a reputation for
influence on the basis of technical expertise, themselves in society, which helps them influ-
special skills, or special knowledge that are ence the behavior of others. Fourth, visibility
both relevant and necessary to the organization. enhances power by making one’s presence
The credibility, and hence the respect, of the felt—people who have a pleasing, helpful, and
17-House.qxd 3/19/2004 5:01 PM Page 515
supportive personality tend to have more McClelland and Burnham determined that
resourceful networks and command more effective managers were characterized primarily
trust. Fifth, charisma produces power through by their need for power. This need takes two
infectious qualities of leadership and influ- forms in organizations: first, the personalized
ence, involving a leader’s aura, dynamism, power type, who strives for dominance and
and persuasiveness (Weber, 1976). In addition, who seeks to pursue personal goals and second,
research on influence behavior has indicated the socialized power type, who seeks to further
several influence tactics that also add to one’s the goals of subordinate individuals, groups,
power (Yukl, 2002). These include rational per- and the organization. McClelland and Burnham
suasion, exchange tactics, legitimate requests, found the latter to be the most effective form
and personal appeals (including ingratiation). of power.
How is the need for power enacted? On
the basis of a study of Harvard University
Need for Power
undergraduates, Winter (1973) saw the need
Perhaps the most foundational treatment of for power, or the need for “having impact,”
the concept of power emerged from Maslow’s enacted in three basic ways. First, impact can
hierarchy of needs. His pyramid presented five be achieved through strong action, such as
categories of human needs, starting at the base assaults or aggression; by giving help, advice,
with physiological needs, followed by safety, or assistance; by controlling someone; by
social, esteem, and self-actualization. The top persuading someone; or by trying to impress
two needs shed some light on a person’s need someone. Second, it can be achieved through
for power. The need for esteem included inter- arousing a strong emotion in someone, ranging
nal need factors such as self-respect and auton- from fear to desire. Third, it can be expressed
omy, whereas external esteem needs included by striving to enhance one’s own personal repu-
factors such as growth, achieving one’s poten- tation, thereby gaining status.
tial, and self-fulfillment. Both of the categories But in many of today’s societies, in which
can be viewed from the perspective of gaining overt expressions of power are unacceptable,
individual power, free from stifling dictates how do people satisfy their need for power?
or obligations from more powerful others. McClelland (1975) tested Winter’s perspectives
According to Maslow, this yearning was not on a group of 50 white-collar and blue-collar
simply a desire, but rather a fundamental need workers in the United States. On the basis of his
of mankind. Although this theory has since been research, he identified four ways in which men
discredited in the psychological community with a high need for power express or satisfy
(Whaba & Bridwell, 1976), Maslow’s work has this need. First, they read sex and sports maga-
been very influential over the years. zines and watch violent television programs.
McClelland (1961, 1975) and McClelland Second, they accumulate prestigious posses-
and Burnham (1976) focused on the needs of sions, like a special car or valuable art objects.
managers within organizations. McClelland had Third, they engage in competitive sports in
originally determined that managers are moti- which there is “man-to-man” competition such
vated by three basic and nonconscious needs— as football, baseball, tennis, and basketball.
the need for achievement, the need for affiliation, Finally, they tend to join organizations in which
and the need for power. The need for achieve- they can hold office and, thus, have an impact
ment is the drive to excel and to strive to suc- on others. Other research on power indicates
ceed, the need for affiliation is a desire for that high power-motivated individuals tend to
friendly and close interpersonal relations, and inhibit group discussion, with the result that
the need for power is the nonconscious desire fewer alternatives may be considered and a sub-
to have an impact and to gain status. McClelland optimal decision reached (Fodor & Smith,
referred to the need for power (nPow) to dis- 1982). Further, high power-motivated individu-
tinguish it from other conceptualizations con- als, if their influence attempts are thwarted,
cerning desire for power. In studying 49 line exhibit more stress than low power-motivated
managers in a large American corporation, individuals (Fodor, 1985).
17-House.qxd 3/19/2004 5:01 PM Page 516
In addition, the social beliefs, values, and McClelland (1970) provided the following
practices of societies tend to carry over not emotive description of the negative reaction to
just to the organizations, but also to a whole unbridled power in many societies:
range of cultural landscapes, through influences
on leadership patterns. One of the insights of It is a fine thing to be concerned about doing
the Charismatic/Value-Based theory of leader- things well (n Achievement) or making friends
ship (House, 1977) is that because societal cul- (n Affiliation), but it is reprehensible to be con-
ture influences people’s values, the enactment, cerned with having influenced others (n Power).
acceptance, and effectiveness of specific leader The vocabulary behavioral scientists use to
behaviors and global leader behavior patterns describe power relations is strongly negative in
are also accordingly modified (Triandis, 1995). tone . . . one finds these people depicted as harsh,
The nonconscious power motivation of sadistic, fascist, Machiavellian, prejudiced, and
American presidents, which was measured neurotic. Ultimately, many claim, the concern
using content analysis of their inaugural for power leads to Nazi-type dictatorships, and
addresses, was found by Winter and Stewart the exploitation of helpless masses who have lost
(1977) to be predictive of the frequency of the their freedom. Even less political terms for power
United States entry into war (r = .62, p < .01) than these have a distinctly negative flavor—
and the frequency of assassination attempts dominance-submission, competition, zero-sum
on the presidents (r = −.81, p < .01). House, game (if I win, you lose). (p. 32)
Spangler, and Woycke (1991) postulated that an
exceptionally high need for power explains why In France, Etienne de la Boetie wrote in 1954
charismatic leaders develop the persuasive in his Discours de la Servitude Voluntaire that
skills to influence others and to gain satisfaction a tyrant has no other power than that which is
from leading. Consistent with this reasoning, given to him, and that the problem is less the
need for power was positively related to behav- tyrant than the “voluntary servitude” of his sub-
ioral charisma of the United States presidents. jects (Hofstede, 1980). This historically accurate
Why would a person, who has his or her own statement refers to a society’s acceptance of
innate need for power, willingly accept another the power that can be exerted by an influential
person’s power over him or her? The most obvi- person, which is the essence of power distance.
ous answer is that there are some things that we As will be discussed in the following section on
have to do to achieve need satisfaction. Staying cross-cultural power distance, the acceptance of
in the organization and being subjected to the unbridled power is limited in many societies by
whims and dictates of a superior are often pre- several cultural factors. Moreover, as McClelland
ferable to leaving the organization. There is also and Burnham (1976) proposed, leaders are more
a conditioned human acceptance of some degree effective when they are motivated by fulfilling
of legitimate authority. In addition, people in the needs of the organization or the concerns of
positions of power are often accepted because subordinates. Within organizations, as opposed
of their personal power, whether it is based on to political positions in some countries, there are
expertise or referent power (French & Raven, usually checks and balances that will not allow a
1959). Many of the charismatic attributes and despotic leader to remain in place for very long.
behaviors (Bass, 1985; House, 1977) play a role In summary, the need for power perspective
in convincing subordinates to comply because suggests that individuals vary in their motives and
of their wish to please and to emulate the leader. attitudes toward the use of power, and that power
Western psychologists have also been con- differences may be accepted if need for power is
cerned with the “dark side” of power. This con- socialized or is exerted by a person of influence.
cern exposes the potential harm that a powerful
individual can cause among hapless subordi-
The Role of Gender
nates or followers due to their wish to comply
Concerning the Use of Power
with the wishes of a despot. Examples in the
political sphere include Adolph Hitler, Idi The discussion of the sources of power,
Amin, Augusto Pinochet, and James Jones. the need for power, and the enactment of power
17-House.qxd 3/19/2004 5:01 PM Page 517
has focused primarily on men’s attitudes and rated their instructors higher than did the male
behaviors. But at least half of the world’s popu- students if the instructor used legitimate and
lation is female. Is there a gender divide on the referent power.
acceptance and enactment of power? (Please In summary, it can be concluded that women
refer to Emrich, Denmark, & Den Hartog’s have different conceptions of power and how
Chapter 14, Gender Egalitarianism, for a more power should be enacted. Women tend to be
detailed discussion.) more sensitive to the use of legitimate and ref-
We might be tempted to think of political erent power, and are expected to enact their spe-
leaders such as Margaret Thatcher and Benazir cific role demands to be accepted positively in
Bhutto to underscore the necessity for “successful the society. Further, there are evidently gender
women” to play the power game to succeed, both stereotypes concerning the acceptance of vari-
in politics and in business. However, this male- ous dimensions of power that, although declin-
oriented perspective might be shallow, inaccurate, ing (Bartol & Butterfield, 1976), still exist in
and self-serving from a female perspective. most societies.
The literature that deals with the female
relationship with power is not straightforward,
The Cross-Cultural
because it tends to blend the concepts of power
with expected female managerial styles and the
Stream and Power Distance
effectiveness of autocratic leadership. For The term power distance was coined by
example, Bartol and Butterfield (1976) claimed Mulder to mean “the degree of inequality in
that women are evaluated more favorably than power between a less power Individual (I) and
men if they use a “consideration style,” whereas a more powerful Other (O), in which I and O
men were evaluated more positively than belong to the same (loosely or tightly knit)
women if they used a structured approach. social system” (Mulder, 1977, p. 90). Building
Haccoun, Sallay, and Haccoun (1978) deter- on the work of Mulder and his colleagues
mined that male, nonmanagement subordinates (Mulder, 1971, 1976, 1977; Mulder, Ritsema
judged a woman using a directive management van Eck, & De Long, 1971), Hofstede con-
style to be less effective than a man using a cluded that, “the basic issue involved . . . is
similar style. Similarly, Campbell, Bommer, human equality. Inequality can occur in areas
and Yeo (1993) concluded that gender has no such as prestige, wealth, and power; and differ-
impact on leadership style. Jago and Vroom ent societies put different weights on status con-
(1982) noted that if stereotype gender roles sistency among these areas” (Hofstede, 2001,
are not met, women are perceived to be auto- p. 75). Following Hofstede (2001), the GLOBE
cratic and are judged negatively, whereas project definition of Power Distance is “the
men are assessed positively. Smith and Smits degree to which members of an organization or
(1994) concluded that female leaders generally society expect and agree that power should be
influence the workplace differently, despite the shared unequally.”
fact that the female and male leaders’ charac- Inequality among various members of a
teristics might be similar. society has been a concern since time immemo-
Another stream of research has examined the rial. In ancient Greece in approximately 350
differences in the perceptions of male versus BCE, Plato, although appreciating a fundamen-
female subordinates. Using canonical correla- tal need for equality among people, endorsed a
tions, Ronk (1993) concluded that gender of the society in which an elite class would exercise
subordinate appears to play a role in the accep- leadership. Those who were educated and
tance of some of the dimensions of power. belonged to the upper class were moral equals
There was no significant difference between the given the same status as others in that class, but
male and female students’ acceptance of the use different from those who were laborers or who
of coercive, reward, and expert power by their worked in nonscholarly jobs. The English
professors. However, there were some statisti- philosopher Thomas Hobbes (1588–1679)
cally different reactions by gender to other maintained that individual behavior is moti-
power dimensions. Specifically, female students vated by the desire for pleasure and avoidance
17-House.qxd 3/19/2004 5:01 PM Page 518
of displeasure. His view was that human nature lower power. In low power distance countries
is largely self-seeking, guided by “egotistical such as Scandinavia and the Netherlands, each
hedonism.” Therefore, power was essential to individual is respected and appreciated for
protect one’s person and goods. This need for what that person has to offer, and people expect
power often led to insecurity and even war, access to upward mobility in both their class and
which Hobbes considered an integral part of the their jobs. Within low power distance cultures,
human condition. the distaste for large power differentials is often
More recently, Beteillie (1977) maintained based on the beliefs that power corrupts, and
that inequality is “one of the central problems that excessive power results in the abuse of
of every human society” (p. ix). Building on power, from which people in less powerful posi-
Rousseau’s (1972) distinction between natural tions have no recourse. The history of many
or physical inequalities and moral or political nations will bear out these concerns, but they
inequalities, he suggested that “society is are not necessarily relevant in some high power
inconceivable without conventions and rules, distance cultures. For example, in the Buddhist
and that these constitute the seedbed of what religion and the Confucian philosophy, power
may be called social as opposed to natural distance implies a reciprocal arrangement that
inequality” (p. 4). The functional necessity of has traditionally protected the less powerful in a
unequal power through force and domination to relationship.
hold societies together was a common theme Whatever our perspective on the uses and
among many of his French contemporaries, abuses of power, many of the roots of power
including Mountesque, Durkeim, Pareto, distance, whether high or low, are historically
Mosca, and Michels. Dahrendorf (1969) sum- derived and satisfy cultural expectations within
marized this view succinctly: “human society societies (Cullen, 2001). It is our interpretation
without inequality is not realistically possible that the acceptance of a certain level of power
and is therefore ruled out” (p. 42). Runciman distance within societies can be traced to four
(1969) proposed three dimensions of social fundamental phenomena. These are the predom-
inequality that are evident, to various degrees, inant religion or philosophy, the tradition of
in most cultures: class structure, power, and democratic principles of government, the exis-
status. tence of a strong middle class, and the propor-
Differences in power distance are by no tion of immigrants in a society’s population. We
means confined to Western thought. In China discuss each of these phenomena below.
around 500 BCE, Confucius spoke of five hier-
archical relationships, each with its norms and
The Role of Predominant
duties. The relationships were ruler–subject,
Religions and Philosophies
father–son, older brother–younger brother,
husband–wife, and senior friend–junior friend. Many of our values and beliefs are pro-
In these relationships, the junior partner owed grammed at an early age by families and basic
the senior respect and obedience; the senior institutions, which in turn are guided by our
partner, in turn, owed the junior protection, group’s religion or philosophy. The distinction
consideration, help, support, and assistance in is made here between a religion, which includes
personal and spiritual matters. Within India, some relationship with one or more deities or
occupation, geography, kinship, and other fac- spirits and generally with an afterlife, and a
tors combined to create a complex class struc- philosophy such as Confucianism, which is
ture, later termed the caste system. This caste concerned with human interrelations and our
system provided each class with a degree of present life. In both cases, they provide funda-
power legitimated and derived from one’s mental influences on a society’s value systems,
hereditary social class. and they provide the basis for many of the laws,
In high power distance cultures such as rituals, and rites to which we are exposed. As
France, some individuals are perceived to have will be discussed, they also provide a founda-
a higher overall rank whose power is unques- tion for our acceptance or rejection of high
tionable and virtually unattainable by those with power distance.
17-House.qxd 3/19/2004 5:01 PM Page 519
The religions that will be discussed are Antioch, Rome, and Constantinople. Reinforcing
Christianity (Catholicism and Protestantism), this hierarchy were the bishops, who in turn
Islam, Hinduism, and Buddhism. Confucianism, were responsible for the ordination and admin-
a philosophy, will also be examined for its influ- istration of lower level priests and deacons. The
ences on power distance within many Eastern Catholic Church (“catholic” meaning “univer-
societies. Several historically important religions sal”) established a synthesis of religion, culture,
will not be considered in this chapter because of and governmental and social structure that was
the relatively small number of adherents or their referred to as Christendom, or the domain of
geographical dispersion. Judaism, for example, Christianity (Oxtoby, 1996). The power and
comprises only 14 million followers worldwide, influence of the church hierarchy became more
with more than 7 million in the Americas, 4 pronounced as the result of mutual support
million in Europe, and 3 million in Israel. Shinto between the Church and heads of state, with
is a purely native Japanese religion that has Church representatives holding increasingly
fewer than 130 million followers, most of whom powerful positions within royal courts through-
also follow Confucian and Buddhist traditions. out Europe. The Catholic Church eventually
The following synopsis of the fundamental asserted its power even over heads of state:
teachings of the world’s most predominant reli- Pope Gregory VII, in 1075, issued a decree that
gions and philosophies is intended to provide the Pope could depose emperors through
some insight into their influences with respect excommunication from the Church.
to the acceptance of a higher or lower degree of The Roman Catholic Church, whose head-
power distance. quarters in the Vatican in Rome is a separate
state, generally reinforced the status quo within
Christianity. Christianity is a religion based on the countries where it settled. While it main-
the belief that Jesus of Nazareth (Christ) was the tained the power and influence of its own repre-
son of the one God and that he is Lord and sentatives, it conditioned the common man and
Savior of the world. Upon his death, he became woman to accept his or her own fate to ensure
part of God within the Holy Trinity. In his life- social stability.
time 2,000 years ago in Palestine, Jesus gained By the late 14th century CE, the Roman
notoriety by challenging the establishment, both Catholic Church was facing criticism for lavish
secular and religious. He argued for personal and corrupt practices, and for the Church’s pre-
freedom, individual determination of one’s fate sumed jurisdiction over people’s personal salva-
through personal prayer, the need for meekness, tion. In 1517, a German monk named Martin
and concern for all people, regardless of their Luther nailed a list of 95 propositions on a
status in life. church door in Wittenberg, Germany, outlining
After his death, Christianity grew to an inde- injustices that he felt should be corrected in the
pendent imperial religion by the 4th century Roman Catholic Church. His primary concerns
CE. Christians now live in every part of the were related to how sins should be forgiven.
world, constituting a quarter of the world’s pop- Rather than accepting the practice that specific
ulation. Christians are the majority in Europe, sins are confessed through a priest of the
throughout the Western Hemisphere, and in Church, he argued that the divine grace of Jesus
Australia and New Zealand. In Africa, the reaches out to every human being who seeks
modern spread of Christianity has produced repentance individually, regardless of their
Christian majorities in many countries south of merit or performance. In essence, Luther was
the Sahara. In Asia, only the Philippines has a taking priesthood off the pedestal of status and
Christian majority, whereas South Korea has a authority that it had enjoyed since the time of
significant minority, approximately one third of Constantine (Oxtoby, 1996).
the population. The subsequent religious movement against
Despite its unstructured beginnings, Chris- the Roman Catholic Church was called Prote-
tianity became more formally organized as it stantism, based on the concept of “protesting”
spread. By the 4th century CE, it was headed against many practices of the established
by five patriarchs in Alexandria, Jerusalem, Christian Church. Fundamentally, these sects
17-House.qxd 3/19/2004 5:01 PM Page 520
were opposed to centralized control and When Hagar was either sent away (Judaic and
hierarchy, preferring instead to allow a person to Christian Bibles) or escorted by Abraham
interpret and apply the lessons in the scriptures (Qur’an) to settle in the valley of Makkah
individually. The Lutheran sect, based on (present-day Mecca according to the Muslim tra-
Luther’s writings, became the predominant reli- dition), she and Ishmael eventually ended up in
gion in Northern Germany and the Scandinavian present-day Saudi Arabia, where they became the
countries. The other two main sects that emerged progenitors of the Muslim Arabs.
from the Protestant Reformation in the 16th cen- The history of Islam as a formal religion
tury were Calvinists and Anglicans. Since that started with Mohammed (also Muhammed and
time, more than 100 Protestant sects have numerous other spellings in the West). He was
emerged, each based on some derivation of the born in approximately 570 CE somewhere near
basic Protestant tenet that the individual needs Makkah. While sitting in the solitude of a cave,
no one to intercede on his or her behalf with he was confronted by the angel Jibril (Gabriel to
Christ and God. The Protestant religions, in gen- Jews and Christians), who taught him the mes-
eral, support the concept of equality of status sages that Mohammed should preach to his
before God, egalitarianism of access to God, people. The message was that the people should
individualist assertion, and hence lower power worship only one God (Allah), they should care
distance before other human beings. for the orphaned, they should feed the hungry,
Roman Catholicism was carried by both they should assist the oppressed and destitute,
missionaries and state representatives to parts and they should offer hospitality to the traveler.
of Africa, countries in Southern Asia, and to Although it is perhaps needless to say, his ideas
most countries in Latin America, where it con- were completely against the status quo. The
tinues to exert influence. Protestantism spread Qur’an (Koran), the holy book of Islam, reports
throughout much of northern Europe and to the the reaction of the established leaders: “We
English-speaking British colonies of Canada, found our fathers following a custom, and we
the United States, Australia, and New Zealand. shall follow in their footsteps” (Qur’an, 40:22).
Although present-day Catholicism is more They did not want their hierarchical society
benign than in previous centuries, it still sup- destabilized by this latest prophet.
ports the status quo in many societies, and it According to Islamic scripture, Mohammed
continues to recognize women as unsuitable to was the last and greatest prophet from Allah. On
hold the highest positions within the Church the surface, his message was simple: Every
establishment. Consequently, societies that have person has equal value before Allah. Live a pure
been primarily Roman Catholic tend to be high life by following the five pillars of Islam:
in power distance, whereas Protestant societies Declaring that there is no God except Allah and
prefer lower power distance. that Muhammad is the Messenger of God;
establishing regular worship; paying zakat alms
Islam. Islam is an innately personal and empow- in service of the poor and needy; observing the
ering religion that is the primary religion from fast of Ramadan; and performing the hajj pil-
Northern Africa in the west to Pakistan and grimage. But the Qur’an makes an important
Bangladesh in the East, and also in Malaysia and distinction between Islam and faith. Islam is
Indonesia. Islam is the third of the historic outwardly an all-encompassing blend of reli-
monotheistic (one God) religions that sprung gious, social, and legal institutions, as opposed
from the Middle East, following Judaism and to Iman (faith), which is an inner belief in God
Christianity. In many ways, it is comparable to expressed through worship and moral relations
Judaism, with which it shares many of the same with others. Thus, Islam provides a template for
prophets, with Adam as the first prophet. Its the- both religious and temporal institutions.
ological split from Judaism is traced to Abraham, Islamic scholars have argued over the cen-
a Jewish patriarch who was the first prophet to turies about whether Islam actually endorses a
embrace one God. Abraham had two sons, hierarchical society. One of the basic Qur’anic
Ishmael, his first born by a servant named Hagar, principles is consultation between the leaders and
and Isaac, the second son born by his aging wife. the followers. However, immediately following
17-House.qxd 3/19/2004 5:01 PM Page 521
Muhammad’s death, the institution of the other countries. Hinduism is also known as
caliphate was devised by the elders of the Vaidika Dharma, meaning “religion of the
Muslims to vest both religious and temporal Vedas,” the ancient Hindu scriptures. The
authority in one leader. On the basis of Vedas teach that the soul is divine, only held
Muhammad’s practice of asserting legal, politi- in the bondage of matter; perfection will be
cal, and religious authority, the caliphs became reached when this bond bursts, providing free-
his successors, similar to their former institution dom from death and misery. This bondage can
of Sub-Sahara chief. They were both the imam, fall off only through the mercy of God, and this
or chief religious leader, and the commander of mercy is given only to the pure; thus, purity is
the Muslims in times of peace and war. However, the condition of His mercy. The fundamental
after the fourth caliph, there was so much infight- message of the Hindu religion is that it is
ing among various tribes and factions that the through constant struggle to purify one’s
caliphs were split into different camps, and there physical, mental, and spiritual body that one can
ceased to be just one caliph. As Islam spread realize communion with the God.
across the Middle East, North Africa, Southern The doctrine of karma is the answer provided
Europe, and into Asia, the imam held a primarily by Hindus to the questions of why suffering and
religious position at the side of what were often inequalities exist in the world: Why should one
considered irreligious monarchical rulers. Sunnis, person be different from another in his looks,
the largest Islamic group, accept this split abilities, and character? Why is one born a king
between governance and religion, but the Shias, and another a beggar? A just and merciful God
who represent one sixth of Muslims and are cannot create such inequalities. The word karma
based primarily in Iran, believe that religion and literally means “deed” or “action,” but implies
state should be inseparable. the entire cycle of causative forces and their
Islam does not endorse high power distance; effects. According to the Law of Karma (liter-
on the contrary, unlike the Roman Catholic ally “action”), every human action—in thought,
Church, Islam has no hierarchy in its network word, or deed—inevitably leads to results, good
of mosques, nor are there any positions of or bad, depending upon the ethical quality of the
authority within the mosques. However, what action. The Law of Karma conserves the ethical
ultimately matters in the discussion of power effects of all actions, and conditions people’s
distance is that most countries that embrace future lives according to their actions in previ-
Islam as their principle religion tended histori- ous lives. Karma is neither predestination nor
cally to be highly hierarchical in nature. fatalism. The past karma of an individual con-
Because Islam is a newer religion, the traditions sists of two components: constant and variable.
of supreme rulers (e.g., in Saudi Arabia, the The past karma tends to carry over to the present
United Arab Emirates, Kuwait, Oman, Syria, birth, and some of it remains constant and
and Iraq) or elitist governments (e.g., in North beyond a person’s control because it determines
Africa) have not been changed by the accep- the person’s family and the country of birth,
tance of Islam by the country’s majority. A physical characteristics, and the dominant social
good example is Iran, which is a highly funda- and religious environment. The effects of past
mentalist and religious Islamic state, but one in karma remain latent in the subconscious mind
which the power distance practices continue. of the child in the form of samska-ras (natural
Within government ministries, the senior offi- habits and tendencies). By manipulating this
cials are normal people who dress and act like variable part of the past karma with the use of
average people, but they also exert substantial initiative and free will, an individual can break
power over those in lesser positions. Evidently, the primal, innate barriers of past karma and
Islam and a high power distance culture can realize a better present and a worthy future.
coexist comfortably. Hindu religious thought embodies a great
variety of ideas, principles and practices, giving
Hinduism. Hinduism is the religion of more rise to various religious schools (samprada-yas).
than 1 billion Hindus, most of whom live in Each school venerates the Supreme Deity, but
India, but there are large populations in many through the veneration of a different personal
17-House.qxd 3/19/2004 5:01 PM Page 522
god who represents a particular aspect of the power. The major conclusion to be drawn from
Ultimate Reality (Brahman). Each school has this discussion is that within Indian and other
temples, guru lineages, religious leaders, pilgrim- Southern Asian societies, power distance was
age centers, monastic communities, and sacred historically very high, and it remains very high
literature. However, all of these schools believe in practice today.
in the central doctrines of Hindu religion.
One of the main aspects of Indian social life Buddhism. In approximately 530 BCE, along the
associated with Hinduism is the caste system. Ganges River in northern India, a boy named
There are four different levels within this system Sakyamuni was born to Queen Mahamaya and
of class hierarchy: At the top, brahman (priests King Suddhodana. At the age of about 29 years,
and guru, the spiritual teacher), then kshatriya while on a chariot ride through the royal park,
(rulers and warriors who were ready to give Sakyamuni saw “four great sights” that altered
their lives up for the safety of the masses, and his perceptions of life: A sick man, an old and
so accorded a premium position), vaishya (mer- suffering man, a dead man, and a monk. This
chants, artisans, and agriculturalists, who served experience focused his thoughts on the vanity
the society through their services and earned of youth, on health, and on life in general, and
compensation to make their living), and shudra toward the role of religion. Sakyamuni fled the
(those at the unskilled and lower-income strata). palace that night, abandoning his privileged
Each of these categories (varnas) largely reflect birthright to enter into a bodhisattva stage of
the occupational roots of the families from learning. After many years of study and medita-
which its members have descended. tion, he achieved the supreme human status of
The Vedas, principal Hindu scriptures, do becoming the Buddha, or Enlightened One.
not talk of castes in terms of ancestral lineages. Without dwelling on the rich history of
However, during the post-Vedic times, amid Buddhism, there are three major “vehicles” of
considerable social and political chaos arising Buddhism that reflect the diversity of the cul-
from external invasions, the caste system came tures into which this religion was introduced.
to be used to maintain the social positions of Buddhists today accept what they call “the
the wealthy, to ensure the sharing of economic Three Jewels” of Buddhist teaching, which are
production, and to ensure an orderly society. the Buddha himself, his teachings, and the
Attempts were made to convince people that “Order of Disciplines” in achieving personal
caste is a preordained and hereditary institution enlightenment. Fundamentally, followers of
with divine sanction. Moreover, although the this religion are encouraged to progress from
caste system is not formally a part of Hinduism, becoming more compassionate, to becoming
the theory of cycles of rebirth, with the possi- more generous, to detaching from worldly
bility of birth in a higher caste, was linked to desires, to becoming more focused mentally on
carrying out one’s duties in accordance with spiritual wisdom and purity.
caste rules. Thus, Hinduism was used as the There are Buddhists in almost every country
vehicle to integrate religion with a highly rigid in Europe, Asia, and North America, and they
social structure. form a high percentage of the population in Tibet
Although the caste system was officially and in the Southern Asian countries of Thailand,
outlawed after India became an independent Cambodia, and Myanmar (Burma). Buddhism
country in 1947, the practice still exists in the also has a substantial number of adherents in
minds of many Hindus today. They continue to Vietnam, Laos, China, Korea, and Japan.
marry within, and associate with, their former From a power distance perspective, Buddhism
caste. The poorer castes are limited in educa- is somewhat fatalistic but, on the whole, neutral.
tional opportunities by their poverty and current It is dedicated to the concern of an individual’s
lack of education, although attempts are being personal path to enlightenment, in which societal
made by several private sector organizations to hierarchy is irrelevant. Because of the cultural
improve their plight. At the top of the social overlays on this religion, it is difficult to disen-
spectrum, the governing class is still seen as tangle the role of power distance within different
entrenched and protective of its own vestiges of Buddhist teachings. Included in this philosophy
17-House.qxd 3/19/2004 5:01 PM Page 523
is the acceptance of the worldly social position loyalty, whereas junior people are, in turn, entitled
into which we are born. Sakyamuni taught that to support.
a person should be judged by his or her charac- Societies in which Confucian traditions are
ter rather than by social status. Although rulers strong include the People’s Republic of China,
are accepted as a functional necessity, a good Taiwan, Singapore, the Koreas, and Japan.
Buddhist government should provide the basic Within these cultures Confucianism formalizes
human needs for all its citizens, and the punish- the principle of high power distance, which is
ment of crimes should be fair and equal for all the essence of its message. Confucian institu-
people. In the balance, it appears that the level of tions embody a very hierarchical, bureaucratic
power distance accepted within a society would structure, with a patriarchal expectation of total
not be influenced by the Buddhist religion. loyalty and obedience. Age is accorded wisdom,
an important determinant of one’s status, which
Confucianism. Unlike the religions discussed in turn influences the system of rewards and
previously, Confucianism is a philosophical tra- promotions. But Confucianism, with its stress
dition. It was developed originally in China to on reciprocal obligations between senior and
deal with social order, veneration of ancestors, junior, mitigates the Western concern about the
and transmitting the wisdom of the past to later potential abuse of power by those in positions at
ages. The ideology was first espoused by a the top of the hierarchy. The major conclusion
man called Confucius in the Western world to be drawn from this discussion is that societies
(552–479 BCE). “Confucius” is the Latin trans- that embrace Confucianism, from Chinese to
lation of K’ung fu-tzu, or “Master K’ung,” who Japanese, are predisposed by this philosophy to
was a man named K’ung Ch’iu, from Shandong accept high power distance as a fundamental
province. Little is known of his early life, except characteristic of an orderly society.
that he travelled through many Chinese feudal
states seeking a ruler who would follow his Summary. The literature suggests that the
advice, but he did not find such a ruler. In his respect for experience and tradition in the
later years, he devoted his time to teaching Confucian and Hindu societies, and the empha-
his disciples, and his teachings were not fully sis on hereditary class roles and spiritual leaders
recorded and systemized until approximately in the Hindu, Islamic, and Roman Catholic
200 years after his death. societies, predispose members of these societies
Although there is a religious side to to accept strong power distance. In contrast,
Confucianism in some texts on the subject, its emphasis on individual initiative for enacting
adherents do not, for example, have temples one’s dream and attaining high status in the
dedicated to Confucian gods. Rather, Confucius Protestant societies, make these societies less
and subsequent scholars following his tradition accepting of power distance. Similarly, the
provided ethical answers to the questions of Buddhist societies are expected to endorse low
life’s meaning and order in society, including levels of power distance due to their thrust on
teachings on the virtue of humaneness (jen). bridging the social castes and their emphasis on
Confucianism addresses the “Five Relation- a community spirit.
ships” that deal with the norms and duties of five
pivotal relationships in society: ruler–minister,
Tradition of Democratic
father–son (filial piety), husband–wife, elder and
Principles of Government
younger brother, and senior friend–junior friend.
The responsibilities implied by these relationships Most developed countries have emerged, or
emphasize a vertical hierarchy based on age and are in the process of emerging, from feudalistic
seniority. The senior person is expected to pro- societies. Feudalism is a system of governing in
vide support and encouragement for the lower- which an upper class (nobility) has certain well-
status person, whereas the lower-status person is defined responsibilities to the ruler, in return for
expected to give loyalty and respect to the senior the use of land exploited with the labor of a
person. This relationship extends to society at peasantry (serfs), and whose produce was traded
large; senior people are entitled to respect and using the services of the merchants. Whereas
17-House.qxd 3/19/2004 5:01 PM Page 524
early feudalism was developed mainly to allow Democracy was actually a Greek term that
monarchs to maintain large armies of mounted evolved several centuries BCE. Depending
troops by giving the nobles—who “owned” the upon the theoretical school, it involved govern-
armies—land in exchange for service, classic ing “by the people” or “for the people.” In its
feudalism consisted of a more political relation- early incarnations in ancient Greece, democracy
ship between a monarch and the nobility. was perceived as governing for the people,
The breakdown of feudalism was initiated which was undertaken by the social elite. In
by the landed nobility, who opposed the author- today’s democratic societies, the term “democ-
itarian rule of the monarchs. Through intrigue racy” has several standard, but not universal,
and war, they managed to wrestle absolute characteristics:
power from the king. But what made the ulti-
mate difference in changing the power estab- As most commonly understood by social scientists
lishment was the growth of strong, town-based today, the term democracy denotes a regime in
merchants. They became increasingly necessary which the authority to govern derives from the
to process and distribute agricultural com- consent of the majority of the people. In practice
modities from the countryside to the cities and such consent is expressed through arrangements
to far-flung trading empires, which again whereby certain people acquire and exercise gov-
shifted the balance of actual power to the ernment power on the basis of regular, free, com-
guardians of commerce in the towns (Blum, petitive elections by all adults, whose votes have
1978). Commensurate with their increased equal value. The electoral principle of democracy
importance in the economic prosperity of their is intertwined with basic liberal principles, that is,
societies, the merchants increased their demand civil liberties—including freedom of speech, of
for representation at the government level. In association, and of participation in the contest for
this new capitalist environment, the town mer- power—without which free elections would not be
chants became the bourgeoisie, who could no possible. (Etzioni-Halevy, 1997, p. xxiii)
longer be ignored.
In the West, the road to democracy was the Western Europe’s last remaining dictator-
result of protracted contests over several cen- ships in Greece, Portugal, and Spain fell in the
turies among various interest groups that made 1970s. In other parts of the world, the emer-
alliances with and against each other. For gence of democracy is still going on today. In
example, in England the aristocracy extracted Latin America, most military dictatorships
numerous concessions from the royalty during collapsed in the 1980s. A movement toward
the 15th and 16th centuries, particularly con- democracy is also evident in East Asia: The
cerning how taxes were raised. In Russia during Philippines, South Korea, Japan, and Taiwan all
the 18th century, the nobility had its obligations have forms of democracy. Since the end of the
to the tsarist autocracy removed. In Germany, 1980s, the former communist regimes in
the aristocracy fought furiously to gain a voice Eastern Europe have been forced to grapple
in government rather than suffer under the dic- with democracy, whereas South Africa accepted
tates of the Great Elector. In France, another democracy in 1994 with the dismantling of
hotbed of political activity in the late 16th cen- apartheid and the country’s first free elections.
tury, the establishment of the First Republic Although there are inequalities in all democ-
marked the end of an absolute monarchy and ratic societies, particularly economic inequities
subsequently virtually eliminated much of the in capitalist states, the mechanisms of democ-
upper class. The result was an abrupt change to racy, such as free elections, civil liberties, and the
a more democratic society, but one that would separation of powers, serve to limit the oppres-
take many years to actually fulfill its promise. sive might of the state and the power of the
Each of the movements led initially to limited state’s elite. Even if one believes that the masses
democracy, and subsequently to full democratic have no real power and virtually no influence
rights for all men, and later for women. over government policies, which is the contention
These early efforts in European countries laid of most Marxist theorists, it is still evident in
the foundation for present-day democracy. democracies that business, academia, the press,
17-House.qxd 3/19/2004 5:01 PM Page 525
and trade unions have usually gained a degree of a large power difference between themselves
autonomy from the state. Thus, they continue to and their superiors. The societies with a large
disperse centralized power to the interests that middle class tend to be pluralist and embrace
they represent. diversity and equal opportunity. As Hofstede
We cannot escape from at least a brief dis- (1980) explained, “in pluralist societies, new
cussion of the role of communism in this debate. members will be more easily admitted into
Communism is a system of socialism enshrined elites than in elite societies, because the middle
in a draconian system of government. In some groups in the pluralist society are stepping
ways, it might be viewed as a form of democ- stones to the top dog ranks” (p. 96). Examples
racy “for the people,” but this is not an argu- of such societies are most historically capitalist
ment that most people are willing to accept. countries, including Australia and countries in
Like other more traditional forms of autocracy, Western Europe and North America.
it is a political system that perpetuates the Countries with virtually no middle class,
inequalities in society, although the new heads including many in Latin America, the Arabian
of the governing elite have changed from Gulf region, Asia, and Africa, have a class
the formerly disenfranchised to the newly structure comprised of a small upper class and a
entrenched as a result of their first-mover posi- large majority of lower-class citizens. As a
tion within the dominant political party. This is result of their lack of access to decision making
not democracy, whether in China, North Korea, and their inability to make substantial changes
Vietnam, or Cuba. This is a form of government in their future, members of the lower class are
that strongly emphasizes the levels of status forced to accept high power distance. At the
inequality. same time, members of the elite expect a large
What is most important about the role of power distance, and try to reinforce the rigid
democracy in predicting a society’s level of class structure.
power distance is the tradition, or comfort level It should be noted that this stereotypical split
gained over time, with the fundamental elements between countries with a large, powerful middle
of a democratic government. We expect that, in class and those without a middle class has
societies that have a long experience with the changed considerably in the past 20 years. As
democratic rights of free speech, free associa- countries around the world become less agrar-
tion, and access to the contest for governmental ian, with a commensurate higher education level
power sharing, there will be less power stratifi- and access to global information, middle classes
cation and centralization. have emerged in most countries. This new class,
with its expectations of mobility and power
sharing, is putting considerable pressure on the
The Role of the Middle Class
ruling regimes to change the distribution of
The middle class is a heterogeneous group power within their societies. However, it gener-
whose members are relatively wealthy in terms ally takes many years to effect the transition
of disposable and discretionary income, tend from a high power distance to a low power dis-
to have a higher education, and hold managerial tance culture, even when widespread expecta-
or administrative jobs rather than positions as tions already exist for low power distance.
physical laborers. They not only believe that Consequently, societies with a large, established
they are responsible for their own lives, but also middle class will have a lower level of power
that they are in the position to make substantive distance than societies with a newly emerging
choices concerning their futures, including their middle class.
education, their careers, and where they live.
Because of their financial security and the
The Role of a High
crucial roles that they play within their
Proportion of Immigrants
country’s economy, members of the middle
class demand access to decision-making bodies Immigrants leave their native countries for a
at all levels of government and within their variety of reasons, notably religious or political
organizations. Consequently, they do not accept persecution, relative poverty, access to higher
17-House.qxd 3/19/2004 5:01 PM Page 526
education, or access to higher paying jobs. The Roman Catholic society that is exposed to
common denominator among these factors democracy and a middle class would reduce its
is the desire to change their position within level of power distance over time, but it would
society. When arriving in their new country, still have a higher level of power distance than a
they often have no extended family as a support Protestant country with a democratic tradition
group, nor do they have communal obligations and a large middle class. Finally, the role of a
and strictures to uphold. Rather, they are forced large proportion of immigrants in a given
to use their own initiative to succeed. society reinforces the low power distance trend
It should be noted that there is a considerable that is in all likelihood already present.
difference between countries that accept immi- We propose one corollary: Regardless of the
grants as citizens and those that allow foreign religion, any society that has neither a democra-
workers who cannot attain citizenship. Foreign tic tradition nor an established middle class will
workers cannot exert their independence, nor have a relatively high level of power distance.
are they allowed to share power in any way. In our discussion of Islam, we suggested that
This is particularly evident in some countries in several of the Islamic principles were funda-
Western Europe, throughout the Arabian Gulf mentally opposed to high power distance, but
states, and in some countries in Southern Asia. that both authoritarian regimes and Islam could
Countries that have traditionally accepted a coexist comfortably. Although this corollary
large proportion of immigrants, notably Canada, does not provide an explanation, it does serve as
the United States, and Australia, are highly indi- a basis for predicting a society’s relative level of
vidualist and democratic, and they have a large, power distance.
mobile middle class. There are no barriers to
entry into the middle class that cannot be over-
come through a mixture of talent, personal How Family Power Values Are Taught
initiative, contacts, and sometimes a bit of luck. Several studies have compared the motives
Even if one enters the country as a menial for having children in a variety of countries
laborer, the expectation is that advancement is (Darroch, Meyer, & Singarimbun, 1981; Fawcett,
possible and that nobody controls one’s destiny. 1983; Kagitçibasi, 1982). They have identified
Consequently, within countries with a large pro- two sets of values attributed to having children:
portion of immigrants, power distance, both As Economic and psychological support. In many
is (practices) and Should be (values) is expected less-developed countries, children are seen as
to be very low. the source of economic support for parents in
their old age. They are expected to be obedient
and to take care of their parents in their old age.
Which Factor Has the
Iranians refer to their children as “the cane for
Greatest Influence on Power Distance?
the hand in old age.” In contrast, those who are
We have treated the four primary factors motivated by psychological support, such as
affecting a society’s level of power distance Americans and Germans, dismiss such a notion,
as virtually distinct influences. In reality, they because children are expected to be autonomous
are often interdependent, and it is difficult to after they leave the family home.
consider one aspect in isolation. However, we Another stream of literature reinforces
speculate that a society’s predominant belief the significance of indoctrinating children at
system and its religion or philosophy will have an early age in cultural values, this one from
the most profound and enduring influence on the perspective of personal empowerment.
power distance. This will then be moderated McClelland (1961) set out to propose a psycho-
somewhat by a democratic tradition and the logical perspective on Weber’s (1976) chronicle
existence of a strong middle class, which go of the Protestant Reformation. Weber’s hypo-
virtually hand in hand in the development of thesis was that the Protestant values of self-
a society. Moreover, both factors can be reliance and hard work led to the evolution of
expected to exert similar influence toward low- the spirit of modern and rational capitalism.
ering the level of power distance. Hence, a McClelland proposed a more detailed process in
17-House.qxd 3/19/2004 5:01 PM Page 527
which Protestant values led to greater emphasis if the differences in resources such as wealth,
among parents on teaching their children to set education, and physical characteristics allow
high standards and to value independence. people to exert social influence that is dispropor-
Winterbottom (1958) had shown that mothers tionate to the social benefits generated from such
of 8- to 11-year-old boys in families with a high influence. An unequal distribution of power per
need for achievement tended to set high perfor- se is not equivalent to power distance, unless
mance expectations for their sons at an early power is used to secure special rank or advan-
age. Other data demonstrate that Protestant tages. The people in more powerful positions,
parents tend to put more emphasis on the impor- whether in organizations or societies, might
tance of independence, knowledge, planning, demand that others accept their positions unquest-
and achieving (McClelland, Rindlisbacher, & ingly. Or, they could create barriers so that the
de Charms, 1955; Rosen & D’Andrade, 1959). underprivileged might not develop new skills
Such characteristics are perceived as necessary and, therefore, not move on to positions at par or
to succeed in low power distance societies. Also, even better than their positions.
according to Hofstede (1980), “the boss–subor-
dinate relationship is a basic human relationship
which bears fundamental resemblance to even Early Studies on Power Distance
more fundamental relationships earlier in life: Haire, Ghiselli, and Porter (1966) were the
that of parent and child . . . and of teacher and first prominent researchers to explore differ-
pupil” (p. 97). ences in preferences for power among different
In summary, the role of the family in learn- cultures, although they did not refer directly to
ing social values is very influential. Families the concept of power distance. They used a
are pivotal in developing the values and atti- questionnaire based on a modified version of
tudes of their children, which then serves as our Maslow’s hierarchy of needs to assess how the
primary source of reference concerning collec- needs of managers from 14 countries were ful-
tive values and culture. Acceptance of a low or a filled in their current positions. The dimensions
high power distance is conditioned by families in their questionnaire that potentially measure
and support groups from an early age. some aspects of power distance across cultures
were composed of autonomy and self-actualiza-
Summary tion. The items presented in their questionnaire
explored the following power-related concerns:
Above we discussed the psychological and
cross-cultural perspectives on power, and their Autonomy
implications for the concept of power distance.
Next, we review the concept of power distance – The authority connected with management
as a dimension of culture. position.
the 14 countries into five groups, which they The next major study of power distance was
labeled Nordic-European (Denmark, Germany, undertaken by Mulder (1977) in the Netherlands.
Norway, and Sweden), Latin-European (Belgium, As the societies become weaker in power
France, Italy, and Spain), Anglo-American distance, the underprivileged tend to reject their
(England and the United States), Developing power dependency. Laboratory experiments in
(Argentina, Chile, and India), and Japan (by social and organizational settings in the Nether-
itself). For this analysis, what is important are lands, a low power distance culture, have shown
the different mean standardized scores that attempts by people to seek “power distance reduc-
the various groups demonstrated with respect tion” (Mulder, 1977). The findings indicated the
to autonomy and self-actualization. Positive following:
values indicate greater fulfillment of need
than for the average manager across all 14 1. More powerful individuals attempt to maintain
countries; negative values indicate lesser or to enhance the power distance from the less
fulfillment. Their results are presented in powerful people.
Table 17.1a.
2. The greater the power distance from the less
The implications of their findings with
powerful person, the greater the attempts to
respect to power distance are complicated.
enhance it by the power holder.
Summarily, it would appear that the Nordic-
Europeans who were sampled were very content 3. Less powerful individuals attempt to reduce
with the fulfillment of their need for power, the power distance between themselves and
Anglo-Americans were somewhat frustrated, more powerful people.
and the other clusters had a higher preference
4. The smaller the power distance from the more
for power than that experienced within their
powerful person, the stronger the tendency to
current jobs. The importance of this study is
reduce it by the individuals with low power.
that countries can be clustered on the basis
of their preference for power. Further, there are
The above set of findings suggests a condition
substantive differences that might be explained
of quasi-equilibrium in which power holders
by the influence of the four criteria explained
have achieved a certain distance from those who
previously: the predominant religion or philoso-
lack power and this distance is difficult for the
phy, an established tradition of democracy, the
latter to bridge.
long-term existence of a middle class, and the
proportion of immigrants in each country.
However, we have to note that the Anglo-
Hofstede’s Seminal Study
American societies sampled in this study appear
to have a greater preference for power than our Chronologically, the next major research on
influencing factors would suggest. this construct was conducted by Geert Hofstede
17-House.qxd 3/19/2004 5:01 PM Page 529
(1980). Hofstede used IBM data in this ground used as a feature of the organizational “regime”
breaking study that assessed commonalities and (the superior’s decision-making style), in con-
differences in various fundamental value con- trast to a characteristic of the organizational
structs across cultures. Building on Mulder’s “climate” indicated in item (a).
work and Aston’s research that identified cen-
tralization as a key attribute of organizations c. What decision-making style would you prefer
(Pugh, 1976), Hofstede used the IBM survey your boss to have?
data to compute a power distance index for eval-
uating the degree of power distance in each In contrast to the second item (b) concern-
country. Hofstede’s power distance index was a ing the perceived manager, item (c) deals with
composite measure computed using the preferences. Hofstede noted that in societies in
responses to three questions: which superiors endorse large power distances,
subordinates prefer a dependent relationship
a. How frequently, in your experience, are with the superiors, or go to the other extreme of
employees afraid to express disagreement with counter-dependent relationship by preferring a
their managers? (mean score across respon- superior who does not decide at all and instead
dents in a society) goes by democratic majority.
1 2 3 4 5 Building on Mulder’s (1977) and Pugh’s
Very Frequently Very Seldom (1976) frameworks, Hofstede maintained that
his power distance index is a measure of the
The measure was based on Patchen’s (1965) extent to which a society accepts the unequal
index of “willingness to disagree with supervi- distribution of power in institutional and organi-
sors.” Patchen’s measure was correlated nega- zational environments. According to Hofstede,
tively with employee control over work goals members of high power distance cultures accept
and interest in work innovation. The theo- large differences in power among the various
retical base for its use at the societal level levels of a hierarchy, whereas members of low
derived from Whyte (1969), who wrote, with power cultures expect a lower level of power
respect to Venezuela, that in a highly stratified differential. Subordinates in high power dis-
society where all powers are concentrated in the tance cultures accept artifacts such as titles,
hands of the superior, subordinates learn that it ranks, and commensurate privileges and status,
can be dangerous to question a decision of the whereas those in low power distance cultures
superior. expect virtually equal treatment on the basis of
their self-worth and their contributions to the
b. How would you describe the actual decision- organization, regardless of status, seniority, or
making style of your boss? age. In low power distance cultures, superiors
still have authority, but employees are not duty-
To measure power distance in boss– bound to be respectful or hold a sense of awe
subordinate relationships in various countries, toward their superiors.
Hofstede’s questionnaires presented four alter- The respondents were the first-level supervi-
native decision-making styles: (a) autocratic sors in the sales and service offices of IBM in
(“tells”), (b) persuasive/paternalistic (“sells”), 41 countries and three regions. In countries with
(c) consultative (“consults”) and (d) democratic lower scores on Hofstede’s power distance
(majority votes, “joins”). In 1970, the fourth cate- index there was limited dependence of subordi-
gory was renamed as participative (“consensus”). nates on bosses and a preference for consulta-
Hofstede noted that to understand the boss’s tion. Subordinates approached and critiqued
influence on subordinate behavior, a subjective their bosses quite freely. In contrast, high power
description of the boss’s decision-making distance index was associated with significant
behavior is relevant. He found that item (b) is dependence of the subordinates on bosses. The
significantly correlated with item (a), and there- subordinates let the boss decide, whether on the
fore included it as part of the power distance basis of personal preferences or on majority
measure. The perceived manager question was vote, rather than actively discussing their point
17-House.qxd 3/19/2004 5:01 PM Page 530
Total 38 −22 97 47
of view. An additional analysis of power population size, and a further 7% from wealth
distance index in 38 occupations, using data (1970 gross national product [GNP] per capita).
from the United Kingdom, France, and West Power distance index was lower in countries at
Germany subsamples, was also conducted. As higher latitudes where, due to colder, adverse cli-
shown in Table 17.1b, the power distance index mate, a greater social mobility had traditionally
tended to be higher in lower-skilled occupations been supported, and power differences could
than among professional workers such as engi- never be stabilized as continuous technological
neers and scientists. innovations rendered older wealth bases obsolete.
However, another analysis using data from The robustness of Hofstede’s power distance
four occupations in 11 countries showed that in index has been established to a fairly high degree
high power distance countries, the power dis- by independent replication in several studies.
tance index was high across all occupations. In Hoppe (1993) included Hofstede’s original
contrast, in low power distance countries, power distance questions in a survey of high-
whereas the employees in higher skilled occu- positioned alumni from an Austrian conference
pations had a low power distance index, the center. The power distance index scores of the 18
employees in lower skilled occupations had overlapping countries in Hoppe’s and Hofstede’s
power distance indices similar to their counter- studies were significantly correlated (r = .67,
parts in high power distance countries. These p < .01). Similarly, Helmreich and Merritt
findings suggested that power distance tends to (1998) surveyed commercial airline pilots in
emerge in all countries, but with higher educa- 23 countries and included Hofstede’s original
tion the people in some societies are able to questions in the survey. The power distance
reduce power distance within organizations. index scores for the 21 overlapping countries
Hofstede (1980) also proposed that the history were strongly correlated with the IBM scores
of power distance philosophy in countries is a (r = .76, p < .01). Also, Shane (1994) used
function of the climate. He found that 43% of Hofstede’s questions in a survey of more than
the variance in power distance index scores could 6,000 employees of six organizations in 32
be predicted from the geographical latitude (of nations in 1991–1992. The power distance index
the country’s capital), an additional 8% from scores were again strongly correlated with the
17-House.qxd 3/19/2004 5:01 PM Page 531
IBM scores (r = 0.54, p < 0.01) for 29 overlapping n = 26, p < .05). These findings suggest that
nations. managers or leaders in societies in which high
In addition, Hofstede reported a high correla- power distance is reported make greater
tion between his power distance index and demands on subordinates, but do so with clarity.
scores on the following items, using Gordon’s For further validation, Hofstede (2001) corre-
(1976) survey of interpersonal values: lated his power distance index with several items
from Inglehart’s World Values Survey. In
1. Greater conformity: Doing what is socially cor- societies that scored high on Hofstede’s power
rect, following regulations closely, doing what distance index, children were encouraged to learn
is accepted and proper, being a conformist. the qualities of “hard work” and “obedience” at
home and were not encouraged to learn the qual-
2. Less independence: Not having the right to do ity of “independence.” In addition, Hofstede’s
whatever one wants to do, being free to make power distance index was correlated positively
one’s own decisions, being able to do things in with the percentages of people choosing equality
one’s own way. to be more important than freedom. That is, the
power distance index, which reflects more
In higher power distance countries, student inequality, was negatively related with the value
samples in 17 countries endorsed the values of of inequality. Thus, Hofstede’s power distance
greater conformity (r = .80) and less indepen- index likely does not reflect power distance val-
dence (r = .79). The differentiation between ues, but instead reflects power distance practices.
the values of conformity and independence was This interpretation is supported empirically in
less stark using managerial respondents. Thus, other studies and will be reported and discussed
in countries in which high power distance is below in the section discussing GLOBE Power
reported, the less powerful tended to endorse the Distance and Hofstede’s power distance index.
values of conformity and dependence more
strongly than the more powerful people.
Other Studies on the
Significant evidence also links power dis-
Construct of Power Distance
tance with role overload as well as lack of role
ambiguity in organizations. Peterson and col- Another important study was the Chinese
leagues (1995) found that Hofstede’s measure of Culture Connection described by House and
power distance is positively correlated with the Javidan in Chapter 2. The members of the
national level of role overload (r = .42). Later, Chinese Culture Connection developed mea-
van de Vliert and van Yperen (1996) found that, sures of Chinese cultural values (Chinese
although Hofstede’s power distance measure Culture Connection, 1987, p. 145). For the
was positively correlated with the national levels purpose of this discussion, power distance
of role overload (r = .44, n = 20, p < .05), the correlated significantly with integration (r =
correlation disappeared when average tem- − .58) and with moral discipline (r = .55). In
perature of the capital city was partialed out of societies that practice power distance, there is
the correlation, most likely reflecting higher less emphasis on integrating different societal
power distance countries closer to the equator. groups. Also, power differences can be legiti-
Temperature was positively correlated with mated and sustained only if those in power
the role overload (r = .56, n = 21, p < .01), and maintain and enforce a sense of moral disci-
remained significant even after controlling for pline. It is important to note that both integra-
the power distance. In a larger sample, Peterson tion and moral discipline correlated significantly
and Smith (1997) again found a positive correla- with Hofstede’s individualism index.
tion between Hofstede’s power distance measure Are power distance and individualism really
and national levels of role overload (r = .64, distinct dimensions? Hofstede reported a corre-
n = 26, p < .01), even after controlling for the lation of .67 between the power distance and
average temperature of the capital city. Further, individualism-collectivism indices. However, he
power distance was negatively correlated with concluded that they are conceptually different,
the national levels of role ambiguity (r = −.42, because power distance refers to emotional
17-House.qxd 3/19/2004 5:01 PM Page 532
dependence on more powerful persons, whereas (1998), who used a 5-point scale to measure
individualism is concerned with personal emo- achievement–ascription with more than 30,000
tional independence. respondents. In response to the statement, “The
Also recall from Chapter 2 that the exis- respect a person gets is highly dependent on their
tence of a cultural dimension concerned with family background,” the highest ascribed status
power inequalities in various societies was also countries were in the Middle East (Kuwait, Saudi
explored by Schwartz (1999). Schwartz devel- Arabia, Oman, and Bahrain), Southern Asia
oped a 2-dimensional graph consisting of (Thailand, Hong Kong, and the Philippines), and
seven value domains, one of which was hier- Africa (Kenya and Burkina Faso). Austria and
archy. In this framework, power differences Serbia were also among this group. The highest
are expressed as a hierarchy value, which relies achieved status countries were Nordic (Norway,
on “hierarchical systems of ascribed roles to Denmark, Finland, and Sweden) and Anglo
insure socially responsible behavior” in which (Ireland, New Zealand, the United Kingdom, the
people seek “to comply with the obligations United States, Canada, and Australia).
and rules attached to their roles” (Schwartz, The next study of importance, also revie-
1999). The values typifying a hierarchical cul- wed in Chapter 2, was conducted by Smith,
ture, that is, one that scores high on Schwartz’s Dugan, and Trompenaars (1996). Achievement–
hierarchy measure, are social power, authority, ascription was measured with six items desig-
humility, and wealth. Hierarchical cultures are ned by Trompenaars using a 5-point Likert scale
deemed to be in opposition with the voluntary format. Completed questionnaires were retur-
cooperation, maintained through egalitari- ned by 8,841 respondents from 43 countries.
anism value-oriented norms, that encourage Multidimensional scaling (MDS) solutions
people to feel concern for everyone’s welfare of dimensionality were computed in a 3-
by internalizing a commitment to voluntary coop- dimensional solution, two of which were labeled
eration with others. Schwartz argued that invol- conservatism–egalitarian commitment and utili-
untary dependence on roles, enforced through tarian involvement—loyal involvement. Power
some kind of authority, is at the essence of distance scores from Hofstede (1980) are related
hierarchical power distance, whereas coop- to both dimensions, with higher power distance
erativeand voluntary negotiation characterizes associated with conservatism and loyal involve-
low power distance and egalitarianism. How- ment. The first dimension of the MDS solution
ever, Schwartz did not provide any indepen- correlated 0.68 ( p < 0.01) with Hofstede’s
dent validation linking the construct of power distance index and −0.81 ( p < 0.01) with
hierarchy–egalitarianism with the forced and Hofstede’s individualism. Smith and colleagues
involuntary dependence on roles. concluded that their first dimension has to do
The work by Trompenaars (1993) also offers with the nature of one’s obligations to groups
insight on this dimension. One of the dimensions and organizations, only one aspect of which is
derived by Trompenaars, using factor analysis captured in Hofstede’s power distance index.
based on the theoretical formulation of Parsons They maintained, “the more general terms of
and Shils (1951), was achievement–ascription, conservatism, ascription, and particularism
which corresponds best with Hofstede’s concep- cover not just hierarchy but the overall basis of
tion of power distance. In achievement-oriented one’s obligations toward all other in-group
societies, each person is responsible for his or her members” (p. 261). Combining their first two
own fate; in ascriptive societies, status is granted dimensions with Hofstede’s power distance
by who you are and power status grows automat- index, they proposed that the highest power
ically with age and experience. distance would be associated with conservatism
Trompenaars determined that countries and loyal involvement, whereas the lowest
could be clustered on the basis of each of his power distance would be associated with egali-
dimensions, including achievement–ascription, tarianism and utilitarian involvement.
which yielded results similar to those of Smith and colleagues found power distance
Hofstede. Trompenaars’s initial findings were to be one of the important constructs that can
extended by Trompenaars and Hampden-Turner distinguish whether a society is collectively or
17-House.qxd 3/19/2004 5:01 PM Page 533
are manifested in terms of the power hierarchy for decentralized decision-making, learning,
(Cohen, 1976), and the analysis of power and corrective actions. On the whole, such
dynamics can yield fundamental insights into power distance reduction in the organizations
the ontology of the organizational culture. adds to the employee beliefs of self-efficacy and
In general, organizations tend to mirror the control in themselves, adds to their work satis-
culture of power distance practices and values faction, and enhances growth and productivity
in their society so that they can gain legitimacy (Huselid, 1995; Nadler, 1989).
and also appeal to the people from their host The use of teams that allow some informal
societies. Further, organizations often need to power sharing is one of the most prominent
respond to cultures of several different societies in approaches to power distance reduction in con-
which they conduct business. They tend to prefer temporary organizations, which also allows for a
business with societies that have similar cultures more strategic and need-based intervention of
of power distance and avoid societies in which authority. The team-based organization is at
a cultural gap is significant. Thus, for instance, the core of the high performance work systems
Confucian societies that have higher power (Nadler, 1989). Team systems support job rota-
distance are reluctant investors in the Nordic tion and total quality management and positively
nations, which tend to be highly egalitarian. influence the level and type of training offered to
However, sometimes the organizations may employees (Smith & Dowling, 2001). In a sample
seek to buffer themselves from the societal of 1,000 firms, Huselid (1995) maintained that the
practices and values of power distance. High magnitude of the returns on investments in high
power distance may be favored in military-type performance work practices is substantial. One
organizations, even if the societal culture is of standard deviation increase in such practices
low power distance. Similarly, power egalitari- was associated with $27,044 more sales, $18,641
anism may be favored in knowledge-intensive more market value, and $3,814 more profits (all in
organizations, in which team-based systems U.S. dollars) on a per employee basis, and a
and empowerment are of considerable interest. 7.05% decline in employee turnover.
In the contemporary learning- and knowl- In summary, although some degree of power
edge-driven environment, the mantra in several distance is implicit in the concept of hierarchy
Western cultures is that lower power distance and may be essential for organizational coordi-
is more effective in organizational settings. nation and control, substantial gains can be
Power sharing, empowerment, and, hence, obtained by reducing the level of power dis-
lower power distance is the prescribed norm. tance within an organization. Reduced power
The concept of empowerment and semi- distance can contribute to the flexibility of the
autonomous teams has proved particularly organization and enhance competence building
striking in the societies at the forefront of the and learning. The gains from empowerment
information technology revolution (Gupta, may be particularly great in high power distance
1998). In the United States, information tech- cultures, in which workforces otherwise may
nology has allowed firms to delay the monitor- not feel free or encouraged to make suggestions
ing and control role of middle managers, and and be committed to the organizational mission.
instead assign a more learning and knowledge At the same time, it may be more difficult to ini-
development role to them. In Europe, firms have tiate and sustain egalitarian power practices in
used information technology to make knowl- such high power distance cultures, in which
edge about the inputs, outputs, and process sup- workforces have been accustomed to depend on
port available to teams of employees, who are their supervisors for direction and decision mak-
then expected to work semiautonomously on ing. Unless appropriate technology and tech-
internal assignments and accomplishment of niques are adopted that facilitate the introduction
their role obligations. In Japan, technology has and incremental refinement of empowerment,
always been applied with creative techniques, power distance reduction may be dysfunctional.
such as kanban (a manufacturing strategy Installment of empowerment practices, when
whereby supplies and production demands are effective, will allow supervisors to focus more
operationally linked) and job rotation, to allow on strategic developmental initiatives.
17-House.qxd 3/19/2004 5:01 PM Page 535
Leadership and Power Distance his or her subordinate. Accordingly, the leader’s
choice of whether to increase or decrease power
The use of power as a source of influence distance may also be conditioned by specific
by leaders, which was studied extensively in relationships with individual subordinates. This
the 20th century, particularly in the context theory questioned the concept that members of
of political or military environments, is beyond an organizational unit are homogeneous, and
the scope of this chapter. Within the organiza- that a superior behaves in essentially the same
tional leadership domain, much of the literature way toward each of the subordinates. Instead,
flows from French and Raven’s (1959) typol- they considered two vertical exchange tech-
ogy of sources of power, which was discussed niques that might occur: Either supervision or
previously in this chapter. At one end of the leadership. Under the higher power distance
spectrum, from a power distance perspective, is style of supervision, there is virtually no rapport
the use of coercive power, which focuses on the between the leader and subordinate. Leaders will
threat or application of punishments to enforce rely almost exclusively on the formal contract
the leader’s wishes. Organizational examples with subordinates to achieve organizational
include the power to reprimand, suspend, goals, with the organization compensating the
demote, fine, or dismiss an employee. We tend employee through financial remuneration. In the
to associate a domineering, autocratic leadership leadership mode, which corresponds closely
style with this leadership type. with lower power distance, the leader works with
Various leadership models have incorporated followers by directly influencing their behavior
autocratic leadership in a dimension concerned through techniques such as offering job latitude,
with decision making, expressing the relative allowing them to influence decision making,
degree of participation that the subordinate is encouraging open and honest discussions, and
allowed (Heller & Yukl, 1969; Strauss, 1977; demonstrating confidence in the group members.
Tannenbaum & Schmidt, 1958; Vroom & Yetton, Which type of power utilization will make a
1973). This continuum generally starts with leader more effective? This would depend on
autocratic decision making followed by some how effectiveness is measured; the situational
form of consultation, then joint decision making, context; and the societal, industrial, and organi-
and finally delegation (Yukl, 2002). Hofstede’s zation cultures in which a given style of leader-
perceived and preferred manager questions for ship is assessed. On the basis of Western
measuring power distance index are also based research, Yukl (2002) concluded that the use of
on a similar typology of autocratic, paternalistic, coercive (autocratic) power is very unlikely to
democratic, or consultative decision-making style create commitment among subordinates; com-
of the leaders. pliance will possibly occur if it is used in a help-
As several prominent researchers have ful, nonpunitive way, and resistance is likely if
pointed out, most leaders use a variety of decision- coercion is used in a hostile and manipulative
making styles, either concurrently or at different way. However, Hersey and Blanchard’s (1977)
times. Their decision-making style will vary situational leadership theory suggested that a
with contingent factors such as the degree of leader should be very directive and autocratic in
formalization of the situation, external threat defining subordinates’ roles, objectives, stan-
faced by the leader and the group, the organiza- dards, and procedures if subordinates are very
tional structure, sophistication of the subordi- immature in relation to their task within the
nates (Heller & Yukl, 1969; Vroom & Yetton, organization. Bass (1985), Vroom and Yetton
1973), and the degree to which the specific situ- (1973), and Yukl (2002) determined that leaders
ational context might give the leader power, tend to be more directive and autocratic if sub-
control, or influence (Fiedler, 1967). ordinates have relatively structured tasks with
The leader–member exchange theory pro- low complexity, low variability, low uncertainty,
posed by Dansereau, Grean, and Haga (1975), and low difficulty.
and further developed in many iterations However, with the growth of empower-
(e.g., Wayne & Green, 1993), addresses the ment and team-based organizations discussed
role-making process between a supervisor and earlier, there is currently a strong trend toward
17-House.qxd 3/19/2004 5:01 PM Page 536
Table 17.2 Higher Power Distance Societies Versus Lower Power Distance Societies
1. Social inequities Society differentiated into Society has large middle class
classes on several criteria
2. Power bases Power bases are stable and Power bases are transient and
scarce (e.g., land ownership) sharable (e.g., skill, knowledge)
4. Social mobility Limited upward social mobility High upward social mobility
6. Governance Different groups (e.g., women) All the groups enjoy equal
have different involvement, and involvement, and democracy
democracy does not ensure ensures parity in opportunities
equal opportunities and development for all
8. Civil freedom Civil liberties are weak and Civil liberties are strong and
public corruption high public corruption low
9. Resources and capabilities Only a few people have access Mass availability of tools,
to resources, skills, and resources, and capabilities for
capabilities, contributing to low independent and entrepreneurial
human development and life initiatives, as reflected in wide
expectancies educational enrolment
11. Technology Mass use of technology, which Need for specialized technology,
supports general power distance adapted to each user
reduction
transformational leadership, in which the leader domineering, and arrogant use of power tend
inspires subordinates to exert themselves to to engender resistance (Yukl, 2002). In socie-
achieve the organization’s goals (Bass, 1985, ties that do not promote this use of power,
1997; House, 1977; House, Wright, & Aditya, power distance practices would clearly be per-
1997). Often, effective leaders use power in ceptible to all and the value placed on power
a subtle manner that diminishes status differ- distance would be low.
entials and avoids threats to the target person’s
self-esteem (Yukl, 2002). In societies that
Summary
emphasize this use of power, power distance
may not be perceptible in practice and therefore Table 17.2 summarizes the differences
would not be overtly devalued. On the other between higher and lower power distance cultures
hand, leaders who engage in a manipulative, expected in the GLOBE Project.
17-House.qxd 3/19/2004 5:01 PM Page 537
2. In this organization, rank and position in the hierarchy should have special privileges: (reverse scored)
Strongly Neither agree Strongly
agree nor disagree disagree
1 2 3 4 5 6 7
establishment and maintenance of dominance and We will now discuss the GLOBE findings
control of the less powerful by the more powerful. using the grand means of reported Power
To differentiate GLOBE’s Power Distance and Distance in societies and organizations, along
Uncertainty Avoidance cultural dimensions, with an analysis of the broad patterns of power
items concerning standardization, rules, proce- distance in clusters of societies around the world.
dures, and other forms of formalization were
reserved for the Uncertainty Avoidance construct,
Overall Findings
unless they implied the exercise of control over
individuals, groups, or organizations.
on Societal Power Distance
Descriptive Statistics
Of all nine dimensions of societal culture
GLOBE FINDINGS ON POWER DISTANCE practices studied in GLOBE, Power Distance
practice has the greatest mean of 5.17. In con-
GLOBE society-level means of Power Distance trast, of all nine dimensions of societal culture
are given in Table 17.4a for practices and values, Power Distance value has the least mean
Table 17.4b for values.1 The worldwide sample of 2.75. Table 17.5a presents the grand mean for
is comprised of 62 societies. The societies are GLOBE’s society Power Distance practices and
classified into five bands of practices scores and values scales. The grand means represent the
four bands of values scores on the basis of their worldwide sample averages of the aggregated
aggregated means for Society As Is and Society society scores. All GLOBE societies, except for
Should Be dimensions, respectively. Denmark (3.89), have Power Distance practices
17-House.qxd 3/19/2004 5:01 PM Page 539
Band
A B C D
a Germany (East): Former GDR *Higher scores indicate greater power distance.
b Germany (West): Former FRG Our response bias correction procedure identified response
c South Africa (White sample) bias in some countries for this scale (see endnotes).
d Switzerland (French-speaking)
e Canada (English-speaking)
f South Africa (Black sample)
scores that exceed the midpoint of 4.0 on the a strong existence of power structures that are
scale of 1 to 7. On the other hand, in all of the universally disliked.
societies, the Power Distance values average is
less than 4.0. Power Distance is thus reported to
Correlation Between As Is
be the least desirable, but the most prominent,
(Practices) and Should Be (Values)
feature of social practices in countries around
the world. The strong perceptions of the prac- The literature review suggested that power
tices of Power Distance, and the general con- has two elements: An influence and a command-
sensus against it in value terms, are indicative of oriented element. The influence element is the
17-House.qxd 3/19/2004 5:01 PM Page 540
Band
A B C D
Country Score Country Score Country Score Country Score Country Score
South Africaa 3.65 Hong Kong 3.24 Namibia 2.86 Hungary 2.49 Colombia 2.04
New Zealand 3.53 Egypt 3.24 Thailand 2.86 Italy 2.47
Albania 3.52 Qatar 3.23 Japan 2.86 Netherlands 2.45
Bolivia 3.41 Kuwait 3.17 U.S.A. 2.85 Austria 2.44
Kazakhstan 3.15 Mexico 2.85 Switzerland 2.44
Poland 3.12 Georgia 2.84 Zambia 2.43
Morocco 3.11 Iran 2.80 Turkey 2.41
China 3.10 Switzerlandb 2.80 Greece 2.39
Taiwan 3.09 England 2.80 Portugal 2.38
Singapore 3.04 Australia 2.78 Brazil 2.35
Malaysia 2.97 Denmark 2.76 Guatemala 2.35
France 2.76 Argentina 2.33
Philippines 2.72 Ecuador 2.30
Israel 2.72 Venezuela 2.29
Ireland 2.71 Spain 2.26
Sweden 2.70 Finland 2.19
Canadac 2.70
Nigeria 2.69
Germanyd 2.69
Indonesia 2.69
El Salvador 2.68
Zimbabwe 2.67
India 2.64
South Africae 2.64
Russia 2.62
Costa Rica 2.58
Slovenia 2.57
South Korea 2.55
Germanyf 2.54
a South Africa (Black sample) *Higher scores indicate greater power distance.
b Switzerland (French-speaking) Our response bias correction procedure identified response
c Canada (English-speaking) bias in some countries for this scale (see endnotes).
d Germany (East): Former GDR
e South Africa (White sample)
f Germany (West): Former FRG
Standard
Power Distance N Minimum Maximum Mean Deviation
Standard
Power Distance N Minimum Maximum Mean Deviation
Table 17.6 Correlation Between GLOBE and Hofstede Power Distance Measures
Common N 47 16 25 21
linear models (HLMs) were conducted to test Distance practices and values with Hofstede’s
these hypotheses more rigorously for organiza- power distance index, using Hofstede’s IBM
tional Power Distance practices and values. We sample and the three replication samples.
tested the GLOBE hypothesis regarding the Hoppe’s (1993) survey was conducted in
effect of societal culture on organizational culture 1981–1982, Shane’s (1994) survey was in
by conducting HLM analyses in which organiza- 1991–1992, and Helmreich and Merritt’s (1998)
tional Power Distance was predicted by societal survey was in 1995–1997.
Power Distance. These analyses supported our The GLOBE Power Distance practices
hypotheses that societal Power Distance prac- measure is correlated positively with Hofstede’s
tices have a significant and strong positive rela- measure in all the four studies, and the corre-
tionship with organizational Power Distance lation is significant in all studies except that
practices ( p < .01). We found a similar signifi- of Shane (1994) (r = .57**, r = .84**, r = .18ns,
cant and even stronger relationship between soci- r = .83**). On the other hand, the GLOBE
etal Power Distance values and organizational Power Distance values measure is not corre-
Power Distance values (p < .01). Both analyses lated with Hofstede’s power distance index for
support a principal proposition in the GLOBE the IBM data. The same holds true for the
theoretical model (i.e., Proposition 3, Figure 2.1, multinational organizations surveyed by Shane
Chapter 2, by House & Javidan): societal cul- (1994), and for the commercial airline pilots
tural values and practices affect organizational surveyed by Helmreich and Merritt (1998).
cultural values and practices.2 However, for the elite sample of Hoppe’s
(1993) study, Hofstede’s measure of power
distance index is significantly and negatively
GLOBE POWER DISTANCE AND correlated with the GLOBE Power Distance
HOFSTEDE’S POWER DISTANCE INDEX values measure (r = − .54).
In summary, it appears that Hofstede’s
Because both GLOBE and Hofstede’s measures original power distance index reflects more
seek to evaluate the construct of power distance, of societal Power Distance practices than
a significant correlation between the two may values, at least in terms of the GLOBE mea-
be expected. Hofstede (2001), in fact, holds that sures. In fact, Hofstede (2001) also recognizes
his measure would provide a valid comparative that two of the three items (“employees afraid”
scoring of societies for at least the next 100 and “perceived managerial behavior”) used to
years. One can use the findings of Hofstede and measure the power distance dimension reflect
GLOBE to confirm the nature of correlation perceptions, which are likely guided more by
between the two measures. Table 17.6 shows situational factors than by personal dispositions
the correlation of GLOBE societal Power or cultural values.
17-House.qxd 3/19/2004 5:01 PM Page 544
Table 17.7 Correlations of Societal Power Distance With Other Cultural Dimensions of Society
Table 17.8a Descriptive Statistics for GLOBE Industry-Level Societal Power Distance
Standard
Variable N Minimum Maximum Mean Deviation
Table 17.9 Descriptive Statistics for GLOBE Industry-Level Organization Power Distance
Standard
Variable n Minimum Maximum Mean Deviation
organizational practices scores are significantly put least value on Power Distance. In the food
higher than the mean Power Distance organiza- sector, the mean score for organizational Power
tional values scores (p < .01). Power Distance Distance practices is 3.88, which is less than the
practices scores are highest in telecommunica- midpoint of 4.0. The food processing organiza-
tions organizations, but these organizations also tions appear to use more egalitarian approaches
17-House.qxd 3/19/2004 5:01 PM Page 547
Total 9.98
df = 60
Total 7.31
df = 60
not based on positional ranks, privileges, and values and practices for every cultural dimension
power. In the telecommunications sector, and to among the 62 societal cultures. In this section
a lesser extent in the financial sector, power dif- we examine the grouping of these 62 sampled
fusion appears to be more difficult to practice, societies into 10 regional clusters. An Analysis
possibly due to a more volatile and uncertain of Variance (ANOVA) test was conducted to
environmental situation and a consequent need to determine if there exist cross-regional differ-
control, coordinate, and audit the actions of front- ences of Power Distance values and practices.
line employees. The results for the ANOVA test are summarized
On the whole, our analysis shows that in Table 17.10a. There are no significant cross-
GLOBE’s Power Distance practices and values regional differences in Power Distance societal
scores are similar across industries. But there do practices (F9, 51 = 1.74, p > .10). Recall that in
exist interindustry differences in Power Distance Table 17.4a that the GLOBE societies fall into
organizational scores. Specifically, organiza- four bands on the basis of their Power Distance
tional practices tend to be quite egalitarian in the societal practices scores. Only three societies
food sector, where stable environment condi- were in band D: the Netherlands, South Africa
tions are more conducive to diffusion of control. (Black sample), and Denmark. Each of these
Power hierarchy practices are most pervasive in belongs to a different regional cluster:
the telecommunications sector, which has an Germanic Europe, Sub-Saharan Africa, and
emergent and volatile environmental context Nordic Europe, respectively. Similarly, there
subject to high costs of technology development. were only four societies in band C (Qatar, Israel,
Albania, and Bolivia) and each of these also
GEOGRAPHICAL REGIONS belongs to a different regional cluster: Middle
AND POWER DISTANCE East, Latin Europe, Eastern Europe, and Latin
America, respectively. The other 54 societies in
the GLOBE sample are distributed across bands
Geographical Regions
A and B, with no clear relationship with any
and Societal Power Distance
regional cluster. Thus, these findings indicate a
As discussed in Chapter 8 by Hanges and similarity among all regions in the practice of
Dickson, there are significant differences in hierarchical control in their societies.
17-House.qxd 3/19/2004 5:01 PM Page 548
Table 17.10b Descriptive Statistics for GLOBE Geographical Regions: Society Power Distance-Values
On the other hand, there do exist significant Put differently, in Continental Europe and
cross-regional differences in Power Distance Latin America, the movement toward a wider
societal values (F9, 51 = 2.24, p < .05). For distribution of power is valued most.
instance, all Confucian Asia societies were
banded in B or C with respect to Power Distance Geographical Regions
societal values, whereas all Nordic Europe
and Organizational Power Distance
clusters were in C and D bands. We found that
regional differences account for 28% of the Geographical region accounts for 21.1%
variation in Power Distance societal values, of the variation in organizational Power
whereas 72% represent idiosyncratic societal Distance practices (F9, 266 = 7.90, p < .01)
differences. and 51.4% of the variation in organizational
The nature of cross-regional differences for Power Distance values (F9, 266 = 31.27,
Power Distance societal values can be seen in p < .01). Thus, regional cluster explains half of
Table 17.10b. Power Distance values are the global differences in organizational Power
reported to be greatest for the Middle East clus- Distance values, and about a fifth of the global
ter (mean = 3.03). On a scale of 1 to 7, this differences in organizational Power Distance
reflects aspirations for a modest degree of practices.
power differentiation in the society. Power The regional pattern of organizational Power
Distance values are reported to be least for Distance practices is presented in Table 17.11.
the Germanic Europe (mean = 2.51) and Latin The regional scores range from 3.47 to 4.41,
America (mean = 2.52) clusters. In general, which is around the midpoint 4.0 on a scale
the Germanic Europe, Nordic Europe, Latin of 1 to 7. We used analysis of variance contrast
America, and Latin Europe clusters have signif- test to identify the clusters with significantly
icantly lower aspirations for societal Power higher or lower scores on the organizational
Distance than do the other regions (value of Power Distance practices scale. The means for
contrast = .34, t = 3.97, df = 51, p < .01). organizational Power Distance practices are
17-House.qxd 3/19/2004 5:01 PM Page 549
significantly higher in Southern Asia (contrast = in Latin American organizations to engage their
.48, p < .01), Eastern Europe (contrast = .27, employees using a variety of teamwork initia-
p < .05), Anglo (contrast = .27, p < .05), and tives. In fact, as also discussed in Chapter 21 by
Confucian Asia (contrast = .23, p < .10) clusters. Dorfman and colleagues, the Latin America
In these clusters with high organizational Power cluster reports higher mean scores (5.96) on
Distance practices scores, respondents report a the effectiveness of team-oriented leadership
strong orientation toward obeying and respect- than any other cultural cluster (contrast = .25,
ing authority. In Anglo organizations, for p < .01).
instance, historical application of scientific As shown in Table 17.11, Middle Eastern
management principles often resulted in a con- organizations stand out, compared with other
siderable gap in the skills, compensation, and clusters, as reporting the highest values of
power of those in higher positions, as compared Power Distance at 4.04. Put differently, respon-
with those in lower positions. dents from Middle Eastern organizations are
The means for organizational Power most neutral in their values of Power Distance,
Distance practices are significantly lower in and are open to a moderate amount of power
Germanic Europe (contrast = −.56, p <.01), distance. Respondents from organizations in the
Nordic Europe (contrast = −.39, p < .01), and other clusters seem to desire a lower level of
Latin America (contrast = −.26, p < .05) clus- Power Distance.
ters. In general, the Northern European and
Latin American managers report practices
Summary
emphasizing subordinate participation, open
expression of disagreement with those in For both practices and values, the Asia-
authority, and emphasis on people’s potential Pacific region reports higher societal and orga-
for the organization. The literature, indeed, nizational Power Distance. Middle Eastern
identifies the Northern European organizations organizations report highest values of Power
as pioneers of the semiautonomous team sys- Distance in any region, while Sub-Saharan
tems in which autonomy and power are granted African societies report moderate practices and
to the members of the team (see Gupta, 1998). values of Power Distance. In contrast, the
Formal training and professional qualification Nordic Europe and Germanic Europe clusters
of the members is emphasized, and monitoring report lower societal and organizational Power
of performance is done through information Distance. Although the Anglo cluster also
technology as opposed to formal supervision reports strong organizational practices of Power
(Gupta, 1998). Consequently, greater power Distance, it has low organizational and societal
decentralization can be, and appears to be, prac- Power Distance values. Finally, the Latin
ticed in these organizations. American and Latin European clusters report
Latin American organizations, on the other lower organizational Power Distance practices,
hand, are usually characterized in the literature and also lower societal and organizational
as having low employee participation and a Power Distance values.
high deference to authority (Hofstede, 2001).
GLOBE’s findings are therefore surprising, but
they may be capturing some recent trends in CLIMATIC REGIONS AND
Latin America. Traditionally, power holders in SOCIETAL POWER DISTANCE
Latin America often reserved key positions for
their actual or adopted family members (under Hofstede (2001) identified physical climate as a
the Godfather system, in which power holders key factor shaping power distance of societies.
accept their loyal employees as godchildren). He observed that in colder, nontropical climates
The control of power at the top has created there has been a greater need to perform and to
considerable distrust and disinterest among the develop technology for surviving against nature.
employees and has made it difficult for the orga- As a result, there was a greater need for educa-
nizations to adopt international best practices tion for those of lower socioeconomic status and
(Altschul, 2003). There may be a growing trend also for greater social mobility and development
17-House.qxd 3/19/2004 5:01 PM Page 550
Table 17.11 Descriptive Statistics for Geographical Regions: GLOBE Organization Power Distance
Practices and Values
of a middle class. More federal and negotiated technology, and agriculture provided sufficient
forms of political representation resulted, and means for survival. As a result, there was less
children learned the things that their parents social mobility, and wealth and political power
never did, on the whole keeping the power were concentrated in a few hands. Children were
differences very low. But in warmer, tropical, dependent on the experience and learning of
and subtropical climates there was less need for their parents, and authority was not questioned.
17-House.qxd 3/19/2004 5:01 PM Page 551
To examine the relationship between own initiative. On the other hand, Charismatic/
physical climate and societal Power Distance, Value-Based and Participative leadership can
we rely on a more systematic classification prove to be quite effective in engaging the
of physical climates. In Chapter 10, Gupta efforts of employees in egalitarian cultures. So,
and Hanges classified GLOBE societies into in low power distance cultures, Charismatic/
seven clusters of physical climates: tropical Value-Based and Participative leadership styles
humid, tropical wet and dry (savanna), sub- may be quite effective.
tropical humid, subtropical wet and dry
(Mediterranean), continental, and marine west
Findings on the Effect of Power
coast (maritime). We conducted Analysis of
Variance tests to examine if societal Power
Distance on Culturally Endorsed
Distance practices and values vary across the Implicit Theories of Leadership
seven physical climate clusters. Physical cli- We tested for the relationship between
mate clusters accounted for 24.8% of the culture and the culturally endorsed implicit
between-society differences in Power Distance theories of leadership (CLT) dimensions by
practices (F = 2.97, p < .05), and only 14.7% of using hierarchical linear modeling (HLM). As
the between-society differences in Power stated in Chapter 11 by Hanges, Dickson, and
Distance values (F = 1.55, p > .10). Further Sipe, HLM can be thought of as a multistep
analysis revealed that only two climatic clus- process designed to test the significance of
ters differed significantly in the level of their relationships between independent and depen-
societal Power Distance practices. Savanna dent variables at multiple levels of analyses. An
climates have significantly higher practices overview of HLM analyses and a detailed dis-
of Power Distance (mean = 5.58, contrast = cussion of how these analyses were conducted
.43, p < .01), whereas continental climates are provided in Appendix C by Hanges and
have significantly lower practices of Power colleagues. This chapter presents relationships
Distance (mean = 4.90, contrast = − .33, p < between one culture dimension (Power
.01). We found no clear pattern among the Distance) and the six CLTs. Competitive tests
countries in these two clusters. Neither Power of all culture dimensions and CLTs are pre-
Distance practices nor Power Distance values sented in Chapter 21 by Dorfman and col-
were significantly correlated with mean tem- leagues.3 In general, we expect that societal
perature, mean rainfall, mean humidity, or and organizational values will be more strongly
mean number of rain days. Thus, contrary to related to CLT leadership dimensions than
Hofstede’s thesis, climate or temperature by societal and organizational practices. As indi-
itself does not appear to account for any over- cated previously, our notions of values and
whelming differences in practices or values of CLT leadership dimensions represent idealized
Power Distance. concepts of how the world should be in contrast
to practices that represent the world as is. As
you read through the results discussed below,
POWER DISTANCE AND it may be helpful to view Figure 17.1 for a
CULTURALLY ENDORSED LEADERSHIP visual summary. The figure, however, only
shows results regarding cultural values, not
How much Power Distance is practiced and practices. (All HLM coefficients are presented
valued should be an important factor in the in Table 21.10 of Chapter 21 by Dorfman and
leadership styles endorsed in various cultures. colleagues.)
In general, the concept of leadership itself may In the present analysis we examine the simul-
be supported more strongly in high power dis- taneous predictive power of organizational
tance cultures because high power distance and societal Power Distance values and prac-
practices and values likely enhance the legiti- tices on CLTs. The total amount of orga-
macy of a leader’s authority. Strong power dis- nizational and societal variance explained by
tance may also relieve frustration among those Power Distance scales ranged from 1% to
people who would find it difficult to act on their 38.6%. Power Distance was found to be a better
17-House.qxd 3/19/2004 5:01 PM Page 552
• Position
important + Humane (S)
• Obeying Self-Protective (S,O)
leaders
• Deferring to
elders
• Inequality Charismatic (S)
okay
− Participative (S,O)
Figure 17.1 Power Distance Cultural Values as Drivers for CLT Leadership Dimensions
* Only statistically significant relationships are shown (p < .05; see Table 21.10, Chapter 21 by Dorfman et al.). The most
important leadership CLT relationships are bolded (i.e., relationship is significant at both society and organization levels of
analyses or highest HLM coefficient within each level of analysis).
O = Organizational level
S = Societal level
predictor for some of the CLTs (e.g., Self- • Team-Oriented leadership. Power Distance
Protective leadership) than for others (e.g., practices scores were significantly related to
Autonomous leadership). the Team-Oriented leadership dimension and
If organizational level and societal level explained a total of 7.5% of organizational and
Power Distance values and practices were con- societal variance for this dimension. All of this
sidered, significant relationships were found variance was associated with forces operating at
with the following: the societal level of analysis. The societal Power
Distance cultural practices scores were positively
• Charismatic/Value-Based leadership. related ( p < .05) to the Team-Oriented leadership
Power Distance cultural values scores were dimension. Team-Oriented leadership is more
significantly related to the Charismatic/Value- likely to be a part of the shared leadership belief
Based leadership dimension and explained a system in organizations reported to have high
total of 17.7% of organizational and societal Power Distance practices.
variance for this dimension. All of this
explained variance was associated with forces • Participative leadership. Power Distance
operating at the societal level of analysis. The cultural practices and values scores were
societal Power Distance cultural values scores significantly related to the Participative leader-
were negatively related ( p < .01) to the ship dimension and explained a total of
Charismatic/Value-Based leadership dimen- 29.9% of organizational and societal variance
sion. Charismatic/Value-Based leadership is for this dimension. Approximately 5.2% of this
less likely to be a part of the shared leadership explained variance was associated with forces
belief system in organizations reported to operating at the organizational level of analysis.
espouse Power Distance values. The majority of the explained variance (94.8%)
17-House.qxd 3/19/2004 5:01 PM Page 553
was associated with forces operating at the • Self-Protective leadership. Power Distance
societal level of analysis. cultural practices and values scores were sig-
nificantly related to the Self-Protective leader-
The organizational Power Distance cultural ship dimension and explained a total of 38.6%
values scores were negatively related ( p < .01) of organizational and societal variance for this
to the Participative leadership dimension. dimension. Approximately 3.5% of this explai-
Participative leadership is less likely to be a ned variance was associated with forces oper-
part of the shared leadership belief system in ating at the organizational level of analysis. The
organizations reported to espouse Power remaining portion of the explained variance
Distance values. (96.5%) was associated with forces operating at
The societal Power Distance cultural prac- the societal level of analysis.
tices scores were negatively related ( p < .01) to
the Participative leadership dimension, as were The organizational Power Distance cultural
the societal Power Distance cultural values values scores ( p < .01) were positively related
scores ( p < .01). Participative leadership is less to the Self-Protective leadership dimension.
likely to be a part of the shared leadership belief Self-Protective leadership is more likely to be a
system in societies reported to have Power part of a shared leadership belief system in
Distance practices, and it is less likely to be a organizations reported to espouse Power
part of the shared leadership belief systems in Distance values.
societies reported to espouse Power Distance The societal Power Distance cultural prac-
values. tices scores were positively related ( p < .01) to
the Self-Protective leadership dimension, as
• Humane-Oriented leadership. Power were the societal Power Distance cultural values
Distance cultural values scores were signifi- scores ( p < .01). Self-Protective leadership is
cantly related to the Humane-Oriented leader- more likely to be a part of the shared leadership
ship dimension and explained a total of 5.5% of belief system in societies reported to have
organizational and societal variance for this Power Distance practices and Power Distance
dimension. All of the explained variance was values.
associated with forces operating at the societal In summary, Power Distance cultural values
level of analysis. The societal Power Distance and practices were statistically significant pre-
cultural values scores were positively related dictors of all six CLT dimensions. In particular,
( p < .05) to the Humane-Oriented leadership if a society’s Power Distance values increase,
dimension. Humane-Oriented leadership is the more likely Humane-Oriented and Self-
more likely to be a part of the shared leadership Protective leadership, and the less likely
belief system in organizations reported to Charismatic/Value-Based and Participative
espouse Power Distance values. leadership, are seen as effective. If a society’s
Power Distance practices increase, the more
• Autonomous leadership. Power Distance likely it is that Team-Oriented and Self-
cultural practices scores were significantly Protective leadership, and the less likely it is
related to the Autonomous leadership dimen- that Participative leadership, will be seen as
sion but explained a total of just 1.1% of orga- effective. At the organizational level, if employ-
nizational and societal variance for this ees describe the organization as holding Power
dimension. All of this explained variance was Distance values, the more likely it is that Self-
associated with forces operating at the organiza- Protective leadership, and the less likely that
tional level of analysis. The organizational Participative leadership, will be seen as effec-
Power Distance cultural practices scores were tive. Finally, if employees describe the organi-
positively related ( p < .05) to the Autonomous zation as having high Power Distance practices,
leadership dimension. Autonomous leadership the more likely that Autonomous leadership will
is more likely to be part of a shared leadership be seen as effective. The implications of these
belief system in organizations reported to have findings are discussed separately below for
Power Distance practices. societal and organizational Power Distance.
17-House.qxd 3/19/2004 5:01 PM Page 554
Societal Power Distance and Leadership likely to be very difficult to implement. From a
historical perspective, General Abdel Nasser in
HLM results suggest that societal Power Egypt is counted as an outstanding Egyptian
Distance practices are associated with lower leader (Bill & Leiden, 1979). Nasser and other
scores for Participative leadership. In the outstanding Arab leaders share patrimonial and
GLOBE dataset, Nordic Europe, Germanic paternalistic orientations in which the leader
Europe, and Anglo clusters score low on retains ideological and strategic centrality
societal Power Distance practices, but have for designing and administrating societal pro-
comparatively stronger scores on Participative grams (Kabasakal & Bodur, 2002). In-Group
leadership. These three clusters share Protes- Collectivism forms the basis for leadership in
tant reformation as a major influence. The all the societal domains, with leaders relying on
reformed clusters profess the direct relationship their close circle—with whom they share emo-
of God and humans without the mediation of tional ties—for advice and protection against
the Church or clerics, and thus nurtured a vision being seen as autocratic and dictatorial.
of humans as responsible persons. The institu- Alternatively, low societal Power Distance
tional leadership in these clusters is thus ori- values are associated with stronger Participative
ented toward encouraging participation at the and Charismatic/Value-Based leadership, and
grassroots levels, and empowering the masses weaker Humane-Oriented and Self-Protective
to share their ideas and express their indepen- leadership. These findings are indicative of a
dent perspectives in a formalized participatory more impersonal, task-oriented focus of leader-
forum. In contrast, societies such as Turkey ship in nonegalitarian societies—that is,
score high on societal practices of Power societies in which power is unequally shared.
Distance but have weak scores on the effective- Thus, for example, the Nordic Europe cluster,
ness of Participative leadership. Previous which strongly values equitable power in all
research by Kozan (1993) suggests that in spheres, has also been at the forefront of semi-
Turkey it is common to use a domineering autonomous approaches at the team level (as
style for handling differences with subordi- opposed to the leadership level). These
nates, and such style is perceived to be an effec- approaches, designed to give self-managing
tive method by superiors and subordinates autonomy to teams, are termed sociotechnical
alike. In general, compared with other nations, work systems. Also, if an egalitarian power dis-
Turkish managers fall toward the lower end in tribution is expected within a society, there is
sharing information and objectives, participa- fundamental distrust of any individual trying to
tion, and internal control. enact a leadership role. House and colleagues
On the other hand, high societal Power (1999), for instance, noted that “The Dutch
Distance values are associated with stronger place emphasis on egalitarianism and are skep-
Self-Protective and Humane-Oriented leader- tical about the value of leadership. Terms like
ship, and weaker Charismatic/Value-Based and leader and manager carry a stigma. If a father is
Participative leadership. These findings possi- employed as a manager, Dutch children will not
bly reflect an emphasis on authoritarian and admit it to their schoolmates” (p. 171). In such
paternalistic modes of leadership, involving cases, universal value-based orientation is criti-
harmony, face saving, benevolence, and in- cal for the leaders to enact their role of helping
group cooperation, in societies that value Power to preserve and promote egalitarianism.
Distance. An illustrative nation from the
GLOBE database is Egypt. Previous research by
Organizational Power
Leila, Yassin, and Palmer (1985) shows that
Distance and Leadership
Egyptian employees have a high degree of apa-
thy, low concern for productivity, and a strong High organizational Power Distance prac-
emphasis on job security. Under these condi- tices are associated with stronger Autonomous
tions, Humane-Oriented and Self-Protective leadership. In hierarchical organizational cul-
leadership are favored, whereas Participative tures, leaders may be effective only by follow-
and Charismatic/Value-Based leadership are ing an independent style and making decisions
17-House.qxd 3/19/2004 5:01 PM Page 555
on the basis of their best judgment. Illustrative leadership by prior research on the use of power.
nations from the GLOBE database with Gardner (1990), for instance, observed that
Autonomous leadership include former East
Germany and Greece. Greece was at the center In our democratic society we make grants of
of the Eastern Orthodox theology that devel- power to people for specified purposes. If for
oped in 1054 C.E. under the influence of Islam. ideological or temperamental reasons they refuse
This split Western Christianity from the to exercise the power granted, we must turn to
Christianity in Eastern Europe. Still, clerics others. To say a leader is preoccupied with power
continued to have an autonomous leadership is like saying that a tennis player is preoccupied
role as the blessed messengers of God, helping with making shots his opponent cannot return. Of
to connect their followers with the transcenden- course leaders are preoccupied with power! (cited
tal power of God (Bakacsi, Sandor, Andras, & in Pfeffer, 1994, p. 12)
Viktor, 2002). The effective leaders in the
Greek context are authority-principled and
Thus, endorsement of power distance in the
paternalistic, who carry a God-like sanctity and
society carries an implicit theory that effective
thus tend to take independent decisions within a
leaders would be ones who make autonomous
hierarchical organization.
decisions, are able to protect themselves from
On the other hand, high organizational
the acts of criticism and corruption, and respond
Power Distance values are associated with
to humane considerations.
stronger Self-Protective leadership and weaker
Similarly with respect to the use of power
Participative leadership. Illustrative nations are
in organizations, and the effectiveness of
included in the Middle East cluster, in which
Self-Protective leadership, Pfeffer (1994)
firms endorse hierarchical power and leadership
emphasized that
that is Self-Protective and not very Partici-
pative. In most of the Middle Eastern organi-
zations, societal effects have an important By pretending that power and influence don’t
influence. As such, they might find it difficult to exist, or at least shouldn’t exist, we contribute to
gain the trust of the populace, which is essential what I and some others (such as John Gardner) see
for the effectiveness of Participative leadership. as the major problem facing many corporations
today, particularly in the United States—the
almost trained or produced incapacity of anyone
Summary except the highest-level managers to take action
In summary, organizational Power Distance and get things accomplished. (p. 10)
values have a positive association with Self-
Protective leadership, and societal Power If Power Distance is valued in organizations,
Distance values are positively related with Self- it thus becomes critical for the leadership to take
Protective and Humane-Oriented leadership. steps for being protected from the likely criti-
The Self-Protective leadership dimension is cism of arbitrary decisions and oversight, as
composed of items that reflect being status and illustrated recently, for instance, by the case
class conscious, ritualistic, procedural, norma- of leadership in the Enron Corporation in the
tive, secretive, evasive, indirect, self-centered, United States.
and asocial. If a society operates in a very diffi- Leaders are presented with a different type of
cult internal environment or is focused on the challenge in cultures that do not value Power
dependent and powerless groups, leaders may Distance. Here, it is important not to appear
need to make autonomous and paternalistic bossy or dictatorial, but instead to delegate and
decisions when responding to humane needs recognize egalitarian and collectivist interests
and be self-protective so that they are not made and to focus on the big picture rather than
into scapegoats for political ends. micromanage. Charismatic/Value-Based leader-
Further light is shed on the relationship ship approaches that encourage team-based
between the societal values of Power Distance engagement and participation at the grassroots
and Self-Protective and Humane-Oriented level therefore become most effective.
17-House.qxd 3/19/2004 5:01 PM Page 556
Public
Societal Sector
Support for Support Success Human
Power Economic Economic for Competitiveness in Basic Life Development
Distance Prosperity Competitiveness Prosperity Index Science Expectancy Index (HDI)
NOTE: Each correlation remained significant, even when per capita GNP was controlled.
* Correlation is significant at the .05 level (2-tailed).
** Correlation is significant at the .01 level (2-tailed).
Respect
Strength for
of Family Lack Role
Power Societal General Family and of of Gender
Distance Health Satisfaction Ties Friends Passiveness Voice Government Equality
NOTE: The correlation of Power Distance practices with strength of family ties, passiveness, lack of voice, role of
government, and gender equality became nonsignificant when per capita GNP was controlled.
** Correlation is significant at the .01 level (2-tailed).
* Correlation is significant at the .05 level (2-tailed).
empowerment of others. Such leaders often need devotion to family and religion that supports the
to enact their power by motivating, preparing, development of each member of the society.
training, and helping others to be empowered. We are now in the position to examine the
Using additional archival measures not dis- validity of Hofstede’s thesis that power distance
cussed by Javidan and Hauser, we found that is associated with less affluent and highly popu-
higher GLOBE Power Distance societal values lated low-latitude conditions. These conditions
predict greater corruption (r = .36, p = 0 < .01) are munificent and nonthreatening to individu-
and lower civil liberties (r = .38, p < .01). The als or organizations in that people can live off
Corruption Index is a measure published online natural resources easily. Hofstede (1980) identi-
by Transparency International, an organization fied power distance as one of the most critical
that assesses perceptions of corruption in public defining attributes of societies. He supported his
services (www.transparency.org). The Civil argument using the findings of stepwise regres-
Liberties Index is a measure published by sion that showed that external data such as the
Freedom House (Freedom House Surveys, 1991) type of political system are most strongly
that assesses freedom of demonstration, media, related to the dimension of power distance, as
movement, religious institutions, and business compared with three other dimensions: individ-
and private organizations. We interpret these ualism, masculinity, and uncertainty avoidance.
findings to mean that under conditions in which Our findings show that, consistent with
Power Distance is highly valued, corrupt behavior Hofstede’s (1980), GLOBE’s measure of soci-
is legitimated as a privilege of position. Further- etal Power Distance practices is negatively
more, in such societies higher status people are correlated with national wealth as measured by
allowed to restrict the liberties of others. GNP per capita (r = −.40, p < .01, N = 62), and
Thus, the focus of managerial intervention with latitude (−.28, p < .05, N = 62). Despite
in societies that practice or value power distance the fact that in Chapter 7 Javidan and Hauser
must also be based on the desire of people to caution against controlling for GNP, to be con-
accomplish results, not just on developing a sistent with Hofstede we followed his practice
select few into leadership positions that need of partialing out GNP per capita. Our analysis
to be protected. A greater sense of performance showed that latitude did not explain any incre-
orientation would endow a self-generated mental correlation with the societal practices of
sense of empowerment, contribute to a vibrant Power Distance; thus our analysis did not
socioeconomic milieu, and facilitate a healthy support Hofstede’s position. We might note that
17-House.qxd 3/19/2004 5:01 PM Page 559
benevolent while being conscious of status and 3. Results between the single HLM and multiple
privilege. At the same time people in these HLM tests will likely differ somewhat. The differ-
societies do not expect the leaders to allow for ences between the results of the multiple HLMs and
participation or to be accountable for results. single HLMs are conceptually similar to the differ-
Leaders in such societies are treated with such ences between a multiple regression analysis and a
a level of deference and respect that they are correlation coefficient. Table 21.10 in Chapter 21 by
not expected to be performance oriented or Dorfman and colleagues presents both single and
visionary. multiple HLM coefficients. In addition, the relation-
ships for all culture dimension values are summarized
in Chapter 3 by Javidan and House.
ENDNOTES
Pfeffer, J. (1994). Managing with power: Politics Trompenaars, F. (1993). Riding the waves of culture.
and influence in organizations. Boston, MA: London: Nicholas Brealey.
Harvard Business School Press. Trompenaars, F., & Hampden-Turner, C. (1998).
Pugh, D. S. (1976). The “Aston” approach to Riding the waves of culture: Understanding
the study of organizations. In G. Hofstede & diversity in global business (2nd ed.). New York:
M. S. Kassem (Eds.), European contributions McGraw-Hill.
to organization theory (pp. 62–78). Assen, The United Nations Development Programme. (1998).
Netherlands: Van Gorcum. Human development report. Geneva: UNDP.
Ronk, L. A. (1993, May). Gender gaps within man- van de Vliert, E., & Van Yperen, N. W. (1996). Why
agement. Nursing Management, 24, 65–67. cross-national differences in role overload?
Rosen, B. C., & D’Andrade, R. G. (1959). The psy- Don’t overlook ambient temperature! Academy
chosocial origins of achievement motivation. of Management Journal, 39, 986–1004.
Sociometry, 22, 185–218. Veblen, T. (1919). The vested interests and the common
Rousseau, J. J. (1972). Du contrat social [From the man. New York: B. W. Huebschy.
social contract]. Oxford, UK: Clarendon. Veblen, T. (1965). The theory of business enterprise.
Runciman, W. G. (1969). The three dimensions NY: Sentry Press. (Original work published 1904)
of social inequality. In A. Beteille (Ed.), Social Verderber, R. F., & Verderber, K. S. (1992).
inequality: Selected readings (pp. 45–63). Inter-act—using interpersonal communication
Middlesex, UK: Penguin. skills. Belmont, CA: Wadsworth.
Schwartz, S. H. (1999). A theory of cultural values and Vroom, V. H., & Yetton, P. W. (1973). Leadership
some implications for work. Applied Psychology: and decision-making. Pittsburgh, PA: University
An International Review, 48(1), 23–47. of Pittsburgh Press.
Selznick, P. (1949). TVA and the grassroots. Wayne, S. J., & Green, S. A. (1993). The effects of
Berkeley: University of California Press. leader-member exchange on employee citizen-
Shane, S. (1994). Cultural values and the champi- ship and impression management. Human
oning process. Entrepreneurship Theory & Relations, 46(12), 1431–1441.
Practice, 18(4), 25–41. Weber, M. (1976). The Protestant ethic and spirit of
Smith, A., & Dowling, P. J. (2001). Analyzing firm capitalism. London: George Allen & Unwin.
training: Five propositions for future research. Whaba, M., & Bridwell, L. (1976). Maslow reconsid-
Human Resource Development Quarterly, 12(2), ered: A review of research on the need hierarchy
147–167. theory. Organizational Behavior and Human
Smith, P. A., & Smits, S. J. (1994, February). The Performance, 15, 212–240.
feminization of leadership. Training and Develop- Whyte, W. F. (1969). Culture and work. In R. A.
ment, pp. 43–46. Webber (Ed.), Culture and management
Smith, P. B., Dugan, S., & Trompenaars, F. (1996). (pp. 30–39). Homewood, IL: Irwin.
National culture and the values of organizational Winter, D. G. (1973). The power motive. New York:
employees: A dimensional analysis across 43 Free Press.
nations. Journal of Cross-Cultural Psychology, Winter, D. G., & Stewart, A. J. (1977). Content analy-
27(2), 231–264. sis as a technique for assessing political leaders.
Stogdill, R. M. (1974). Handbook of leadership: A In M. G. Hermann (Ed.), A psychological exami-
survey of the literature. New York: Free Press. nation of political leaders (pp. 27–61). New
Strauss, G. (1977). Managerial practices. In J. R. York: Free Press.
Hackman & J. L. Suttle (Eds.), Improving life at Winterbottom, M. R. (1958). The relation of
work (pp. 297–362). Santa Monica, CA: Goodyear. need for achievement to learning experiences in
Tannenbaum, R., & Schmidt, W. H. (1958, independence and mastery. In J. W. Atkinson
March–April). How to choose a leadership (Ed.), Motives in fantasy, action and society
pattern. Harvard Business Review, 36, 95–101. (pp. 453–478). Princeton, NJ: Van Nostrand.
Triandis, H. C. (1995). Individualism and collectivism. Yukl, G. A. (2002). Leadership in organizations (5th
Boulder, CO: Westview. ed.). Upper Saddle River, NJ: Prentice Hall.
18-House.qxd 3/19/2004 5:05 PM Page 564
18
HUMANE ORIENTATION IN
SOCIETIES, ORGANIZATIONS,
AND LEADER ATTRIBUTES
HAYAT KABASAKAL
MUZAFFER BODUR
his chapter begins with a discussion other economic and social indicators, and other
AUTHORS’ NOTE: The authors wish to thank Idil Evcimen, Fahri Karakas, Aslihan Nasir, and Murat
Sahtiyanci for their help and assistance.
564
18-House.qxd 3/19/2004 5:05 PM Page 565
involves promoting self-interest and self- cross-cultural study conducted in seven countries
gratification, and can be interpreted as being less and eight communities—Canberra, Brisbane,
humane oriented. Winnipeg, Phoenix, Berlin, Hong Kong, Taipei,
Hofstede (1980, 2001) identifies toughness and Osaka—achieved similar findings of the
and tenderness as a component of his masculin- importance of family for the development of
ity versus femininity dimension, as measured children’s personalities (Scott, Scott, & McCabe,
with his MAS Index. On the other hand, he uses 1991). In all the communities that were included
this dimension to refer to the distribution of in the study, children’s self-esteem was associ-
emotional roles between genders, asserting ated positively with family harmony and parental
that his MAS Index also measures materialistic nurturance, and children’s hostility was corre-
orientation and success striving. Although the lated with parental punitiveness. There was gen-
index includes and is confounded by several eral uniformity over cultures in magnitudes of
cultural variables, it is similar to GLOBE’s correlations. These findings indicate that there
Humane Orientation dimension in that cultures would be more psychologically healthy individu-
that score low on the MAS Index are consi- als in societies that promote parental nurturance
dered to be as relationship oriented as in high and minimize punitiveness in the family.
humane-oriented societies. The associated cross-cultural comparisons
Paternalism, a form of benevolence, is also reflect the influence of societal context on per-
a dimension of social norms that differentiate ception of social support. In societies located in
cultures from each other (James, Chen, & North America and Europe where child-rearing
Cropanzano, 1996; Kanungo & Aycan, 1997). norms emphasize autonomy rather than close
In paternalistic societies, people in authority are control, children perceive close parental super-
expected to act like a parent and take care of vision as rejection and lack of love.
subordinates’ and employees’ families. Eastern One of the few cross-cultural studies of
cultures, such as China, India, Turkey, Pakistan, families was conducted in the mid-1970s in
and Taiwan, and most Latin-American societies Indonesia, Germany, Korea, the Philippines,
demonstrate more paternalistic values. In Singapore, Taiwan, Thailand, Turkey, and the
societies that score low in paternalism, there is a United States (Bulatao, 1979; Darroch, Meyer, &
limited interest in workers’ problems, which is Singarimbun, 1981; Fawcett, 1983; Hoffman,
restricted to job-related issues. 1987; Kagitçibasi, 1982a, 1982b, 1982c). This
In societies that lack formal welfare insti- study attempted to find the motives for having
tutions, where resources are very unevenly children and the kinds of value attributed to
distributed and where political power is often children in these societies. Cross-cultural com-
unstable, a system of patronage based on rela- parisons indicated the presence of two types of
tionships of family and friends emerges to values that are attributed to children: Economic
fulfill some needs of individuals (Wolf, 1966). and psychological. Economic value of children
Organizations in these societies tend to choose refers to a child physically helping the family
employees on the basis of the individual’s and serving as a guarantee of positive treatment
relationship to the employer or patron, rather in the old age of parents. In developing societies,
than the organization’s needs or the skills of to serve the purpose of economic value, obe-
the individual (Kiray, 1997; Ong, 1987). dience rather than autonomy is valued in children
as a desirable trait. An autonomous child may
Family Practices, Norms, place self-interest above the interests of the family;
consequently, autonomy in children is discour-
and Humane Orientation
aged. In developing countries, particularly in the
Many studies conducted in English-speaking rural areas, the normal way of life is such that
cultures point to the influence of family relations children take care of their parents and provide
on children’s well-being. There is strong evidence material help in their old age.
that parental attention and warmth produce high In some developed countries, such as Japan,
self-esteem (Adams & Jones, 1983; Hoelter & close parental control continues to be perceived
Harper, 1987; Peterson & Kellam, 1977). A as a sign of attention, acceptance, and affection
18-House.qxd 3/19/2004 5:05 PM Page 567
because of the persistence of traditional values, Collective and communistic societies are
even if there is no economic necessity for communal, status-based, and governed accord-
children to take care of aging parents. Inter- ing to tradition. As Donnelly (1989) describes,
dependence among family members continues there is generally no notion of autonomous
even without economic necessity. In Japan, and individuals in such societies, and censorship is
most Western, developed countries, children are sometimes practiced (Wolf, 1994).
viewed as sources of psychological satisfaction
rather than economic security. In these societies The Role of Government
children provide psychological value to their
and Humane Orientation
parents and are regarded as sources of pleasure,
friendship, and companionship. In many industrialized nations, a welfare
state replaces informal and family relationships
that are needed for the survival of members
Human Rights, Discrimination,
of society. According to Briggs (1961), a wel-
and Humane Orientation
fare state is one in which organizational power
Human rights are recognized by almost all is used through politics and administration to
nations. Moreover, half of the world’s states modify the market forces so that individuals
have accepted legal obligations to enforce and families are guaranteed a minimum income
“these rights by becoming parties to the Inter- irrespective of the market value of their prop-
national Human Rights Covenants and almost erty. Jessop (1991) views such an effort by the
all the other nations have either signed but not state as narrowing the extent of insecurity and
yet ratified the Covenants or otherwise expressed ensuring that all citizens are offered a certain
approval of and commitment to their content” range of social services to meet contingencies
(Donelly, 1989, p. 2). such as illness, old age, and unemployment.
The Universal Declaration of Human Rights, Esping-Anderson (1985) discusses three
as adopted by the United Nations General forms of welfare state in a typology, namely the
Assembly in 1948, covers a wide range of issues social democratic model, the conservative
including the right to life (i.e., protection from model, and the liberal model. Among the three
infliction of death), rationality, recognition types of welfare states, the social democratic
before law, protection against torture, protection model is argued to be more humanely oriented
against discrimination on such bases as race and in the sense that the personal worth of all
sex, and the right to education and to participate individuals is honored by the recognition of
in the cultural life of the community. These universal welfare rights and the concept of
standards reflect what are considered the mini- social citizenship prevails. Financing comes
mum conditions for a dignified life. from general taxation and there is a commit-
Prior to the creation of capitalist market ment to egalitarian distribution of resources and
economies and modern nation states, the prob- high standards of public provision. This type of
lems that human rights seek to address, in par- welfare state is best illustrated in Scandinavian
ticular violations of human dignity as Donnelly countries such as Sweden and Norway where
(1989) points out, either did not exist or were the working class is unionized.
not widely perceived as crucial social problems. In the conservative model, there are differen-
Multicultural societies and communities that tiated welfare rights that are linked to occupation
stand for the freedom and equality of all people and status, and social welfare does not have egal-
rest upon mutual respect for intellectual and cul- itarian aims. Austria, Germany, France, Italy,
tural differences. Failing to recognize or respect and Belgium in Europe, and Japan in East Asia
the particular identities of different cultural have this type of a welfare state. Finally, in the
groups is under severe criticism, especially in liberal model, limited welfare rights are pro-
liberal democracies in which commitment to the vided to individuals by the state, and state expen-
principle of equal representation is the highest, ditures are made only for the benefits that are
especially with regard to the rights and needs of means-tested. Britain, the United States, Canada,
members of specific cultural groups. and Australia depict this type of a welfare state.
18-House.qxd 3/19/2004 5:05 PM Page 568
In general, states in all societies have been arrangements and a reserve labor force. Under
attributed a positive role in intervening to these circumstances, the states’ role turned into
correct market failures, such as shortages in promoting flexibility to compensate for defi-
capital, foreign exchange, or some other inputs. ciencies in the market and making compliance
Neoclassical development theory proposes that with state measures voluntary through “third
the ideal role of the state is to provide property sector” forces (Jessop, 1991).
rights and basic infrastructural services. As best In general, in the more economically devel-
demonstrated by the Japanese experience, the oped nations of the world, the state takes
state plays a positive and socially constructive on the role of at least a minimum standard
role in the development process by maintaining of guaranteeing well-being of members of
a mixture of autonomy and cooperation. Thus, societies. In low humane-oriented societies,
the state performs more of a strategic role for sus- which are economically more developed, the
taining long-term social reforms by disengaging state is actively involved in social protection,
from the intervening role in favor of facilitating providing business incentives, and improving
organizational links that connect business inter- working conditions.
ests to the state through such practices as pri- Participation of children under age 15 in
vatization of state enterprises. Theoretically, as a the workforce differs from country to country.
result, direct public provisions are reduced, and Rates increase with age and are higher for boys
profit-making investments and joint business–state in the case of wage work or work in the house-
ventures are encouraged. hold enterprise, whereas girls participate more
The equal opportunities policy perspective in in domestic activities (International Labor
European social protection systems calls for the Organization, 1992).
need to integrate child care into the system as a According to a study by Grootaert and
social risk just like the risk of growing old. As Kanbur (1995), in less-developed countries the
described by Brouwer and Wierda (1998), the variables of larger household size, low parent
basic feature of this scenario is that both women income and landholdings, and failure in educa-
and men are able to fulfill child care responsi- tional systems play a significant role as determi-
bilities by combining paid labor tasks and nants of child labor. Developing countries, such
unpaid care tasks in their “private” time. In this as India and Egypt, are implementing policies to
model, social security rights are not individu- reduce child labor; however, the welfare states
alized but “familized.” of the developed countries have introduced
With the internationalization of capital that more radical measures to attack the child labor
came with world market integration, production problem.
was decentralized to exploit locational advan- In addition to their more controlling charac-
tages in sourcing. New technologies in telecom- ter, most European governments take a more
munication and computer-aided manufacturing supportive role toward organizations in compar-
systems generated more flexibility in adapting to ison to the United States by use of vocational
consumer demand by specialization. As mobility training conducted in cooperation with state
of white-collar workers increased, a need for agencies (Pieper, 1990), allocation of a fairly
flexibility in labor markets also became appar- high percentage of gross domestic product to
ent. Consequently, recruitment rules and wage public expenditure on labor market programs,
rates started to require more heterogeneity. retraining, and job transition support for the
When flexibility is viewed as a central part of unemployed (Brewster, 1995).
international economic modernization, states
often choose a neo-liberal model to restructure
Economics and Humane Orientation
their economy. As put forth by Hirsch (1991),
when a neo-liberal model is adopted, the state is In less-developed societies, individuals rather
reoriented to revitalize the market by promoting than governments are expected to offer both
small business, reforming environmental pro- material and psychological support. Privileged
tection measures, and by liberation of wage and members of these societies are expected to share
working conditions such as short-time working their money, material possessions, and other
18-House.qxd 3/19/2004 5:05 PM Page 569
resources. When people are living through solidarity and love relationships are also likely
difficult times, such as death, illness, or financial to suffer” (Allardt, 1993, p. 91). Although this is
problems, friends, neighbors, and family members the only study relating humane social relation-
spend time with them, bring cooked food, and ships to quality of life, we speculate that it is
visit them often. likely that this finding represents a universal
Table 18.1 summarizes major connotations relationship between these two variables.
and variations of the humane orientation construct
at the societal level, on the basis of the above
literature review of societal practices and country- Social Support and Stress
specific manifestations of humane orientation.
In a review of research on social support,
Güngör (1997) identified five basic sources of
COMPARISON OF HIGH VERSUS LOW social support: (a) instrumental support, which
HUMANE ORIENTATION SOCIETIES involves material, financial, and service support;
(b) social companionship, which is spending
GLOBE’s concept of the Humane Orientation leisure time together; (c) emotional support based
cultural dimension is defined as the degree to on extending empathy and love; (d) esteem sup-
which an organization or society encourages port whereby the individual is accepted together
and rewards individuals for being fair, altruistic, with problems and deficiencies; and (e) informa-
friendly, generous, caring, and kind to others tional support, which involves supplying relevant
(House et al., 1999). This dimension is mani- information and the necessary skills to cope
fested in the way people treat one another and in with problems. Güngör (1997) concluded that
the social programs institutionalized within all types of social support from supervisors,
each society. Research indicates that the way in peers, subordinates, family members, and friends
which people treat one another, in the terms contribute to psychological and physical health.
defined above, varies by culture. Social support has a positive effect on
The GLOBE study implicitly relates quality of individual well-being by alleviating the nega-
life and social relations. Other studies have explic- tive effects of stress (Maslach & Jackson,
itly associated quality of life and resources, both 1984; Russell, Altmaier, & Van Velzen, 1987).
tangible and intangible. In 1970, national studies Social support can directly enhance well-being
measuring the level of welfare in societies were because it satisfies important human needs such
conducted in all Scandinavian countries includ- as security, social contact, approval, belonging,
ing Sweden, Denmark, Norway, and Finland and affection. In addition, social support mod-
(Allardt, 1993; Erikson, 1993). The Scandinavian erates the relationship between stressors and
approach assumed that one’s standard of living burnout through the use of adaptive coping
is associated with the resources by which individ- strategies (Cordes & Dougherty, 1993).
uals can master and control their lives. Resources
were defined to include money, possessions, Overview of Prior Cross-Cultural Studies
knowledge, mental and physical energy, security,
on the Concept of Humane Orientation
recreation, and political resources, as well as social
relations. Allardt (1993) notes that it was agreed As described in the review of large sample
that an emphasis on resources to the exclusion of cross-cultural studies in Chapter 2 by House
social support would in practice lead to a one-sided and Javidan, Schwartz and Bilsky (1987)
focus on material conditions. derived and confirmed seven distinctive mot-
A general finding in the Scandinavian study ivational domains of values that they claim
was that the amount and strength of social to be universal. These motivational domains
relations were not correlated with material were derived from a literature review of needs,
possessions. However, we might speculate the social motives, institutional demands, and func-
lack of correlation is a statistical finding that tional requirements of social groups. One of
applies to a normal Scandinavian. “As soon the domains is labeled the prosocial domain.
as material conditions markedly deteriorate This positive social requirement is expressed in
18-House.qxd 3/19/2004 5:05 PM Page 570
Table 18.1 Higher Humane Orientation Societies Versus Lower Humane Orientation Societies
Fewer psychological and pathological problems. More psychological and pathological problems.
Values of altruism, benevolence, kindness, love, Values of pleasure, comfort, self-enjoyment have
and generosity have high priority. high priority.
Need for belonging and affiliation motivate people. Power and material possessions motivate people.
Personal and family relationships induce Welfare state guarantees social and economic
protection for the individuals. protection of individuals.
Close circle receives material, financial, and Lack of support for others; predominance of
social support; concern extends to all people and self-enhancement.
nature.
Members of society are responsible for State provides social and economic support for
promoting well-being of others: The state is not individuals’ well-being.
actively involved.
The state supports the private sector and The state sponsors public provisions and sectors.
maintains a balance between public and private
domains.
Public policymakers establish sanctions against Public policymakers consider child labor
child labor practices. practices as a somewhat less-important issue.
Members of society are urged to be sensitive to Members of society are not sensitive to all forms
all forms of racial discrimination. of racial discrimination.
People are expected to promote paternalistic Formal welfare institutions replace paternalistic
norms and patronage relationships. norms and patronage relationships.
People are urged to provide social support to People are expected to solve personal problems
each other. on their own.
The children of less-developed societies are The children of more-developed societies are not
expected to give material support to their expected to give material support to their parents
parents in their old age. in their old age.
The children of less-developed societies can The children of more developed societies are not
participate in the labor force to help out their expected to participate in labor force to help out
families. their families.
Parents should closely control their children. Family members are independent.
such values as altruism, benevolence, kindness, that a value preference for promoting other’s
or love. To test the structure of the values, welfare (prosocial) contradicts both valuing per-
the researchers asked a sample of Israeli tea- sonal success (achievement) and valuing one’s
chers (N = 455) and German college students own pleasure and comfort (enjoyment).
(N = 331) to rank the 36 Rokeach values in order Schwartz and Bilsky (1990) further confirmed
of importance. Schwartz and Bilsky (1987) found the universality of the seven distinct motivational
18-House.qxd 3/19/2004 5:05 PM Page 571
Table 18.2a Sample Questions: Humane Orientation: Society Practices (As Is)
types of values with data from Australia, Pakistan, and India, Kanungo and Aycan (1997)
Finland, Hong Kong, Spain, and the United included paternalism as one of the four socio-
States. Moreover, the opposition between pro- cultural dimensions of societies. Paternalism
social and achievement values held across all was assessed with statements such as “In our
seven samples, indicating that the value conflict society, the ideal boss is like a parent,” and
between the two motivational domains is likely “People in authority in our society should take
to be universal. care of their subordinates as they would take
In another cross-cultural study using factor care of their children.” Eastern cultures were
analysis conducted in 12 nations (Bigoness & found to carry significantly more paternalis-
Blakely, 1996), responses of 567 managers tic values compared with Western countries.
yielded four factors in which the first factor Thus, paternalism might reflect a different form
included the values of forgiving, helpful, lov- of humane orientation that is dominant in
ing, and cheerful, corresponding to the humane Eastern societies.
orientation construct. However, compared with
the other three value dimensions, managers
ranked humane orientation as least important GLOBE MEASURES OF
in directing their lives, providing an interesting HUMANE ORIENTATION
insight into the values of managers as an occu-
pational group. Five questionnaire items in the GLOBE study
Schwartz (1992) developed an instrument were used to operationalize the societal Humane
that represented the basic universal human Orientation construct: Being concerned, sensi-
values, with a procedure that avoided a Western tive toward others, friendly, tolerant of mis-
imposed bias. Data were collected from takes, and generous. Chapter 8 by Hanges and
44,000 schoolteachers and university students Dickson describes the development, validation,
in 54 countries. Results showed that value and format of the items. Tables 18.2a and 18.2b
types are organized in two bipolar dimen- show sample questionnaire items that are used
sions, one being the self-transcendence (uni- for the societal Humane Orientation construct.
versalism and benevolence) and the other the As stated in the introduction to this section,
self-enhancement (power and achievement) the organizational As Is (practices) questions are
dimension, discussed in an earlier section. The designed to measure the respondents’ assess-
duality between self-transcendence and self- ment of the extent to which an organization
enhancement reflects the opposition between engages in Humane Orientation practices. The
humane orientation and task concerns. At the Should Be (values) questions measure the val-
same time, however, humane orientation was ues of respondents regarding how humane-
confirmed as a basic value that is prevalent in orientated they would like their organizations to
all of the regions of the world. be. Tables 18.3a and 18.3b show sample ques-
In a cross cultural study of Canada, the tions that measure the organizational Humane
United States, Romania, Turkey, China, Orientation practices and values, respectively.
18-House.qxd 3/19/2004 5:05 PM Page 572
Table 18.2b Sample Questions: Humane Orientation: Society Values (Should Be)
Table 18.3a Sample Questions: Humane Orientation: Organization Practices (As Is)
Table 18.3b Sample Questions: Humane Orientation: Organization Values (Should Be)
Table 18.4a Summary Statistics for Humane Orientation at the Societal Level
Band
A B C D
Band
A B C D
Country Score Country Score Country Score Country Score Country Score
Nigeria 6.09 Spain 5.69 Philippines 5.36 Costa Rica 4.99 New 4.49
Finland 5.81 Brazil 5.68 Albania 5.34 Zealand
Singapore 5.79 France 5.67 Hong Kong 5.32
Austria 5.76 South Africaa 5.65 China 5.32
Sweden 5.65 Portugal 5.31
Canadab 5.64 Venezuela 5.31
Switzerlandc 5.62 Qatar 5.30
Kazakhstan 5.62 Poland 5.30
Israel 5.62 India 5.28
Iran 5.61 Taiwan 5.26
Colombia 5.61 Ecuador 5.26
Georgia 5.60 Guatemala 5.26
South Korea 5.60 Slovenia 5.25
Russia 5.59 Greece 5.23
Italy 5.58 Netherlands 5.20
Australia 5.58 Zimbabwe 5.19
Argentina 5.58 Egypt 5.17
Switzerland 5.54 Indonesia 5.16
Zambia 5.53 Mexico 5.10
U.S. 5.53 South Africaf 5.07
Turkey 5.52 Bolivia 5.07
Malaysia 5.51 Kuwait 5.06
Morocco 5.51 Thailand 5.01
Hungary 5.48
Ireland 5.47
Germanyd 5.46
El Salvador 5.46
Denmark 5.45
Germanye 5.44
England 5.43
Japan 5.41
Namibia 5.40
a South Africa (White sample) * Higher scores indicate greater humane orientation. Our
b Canada (English-speaking) response bias correction procedure identified response bias in
c Switzerland (French-speaking) some countries for this scale (see endnotes).
d Germany (West): Former FRG
e Germany (East): Former GDR
f South Africa (Black sample)
aspire to higher humane orientation in their p < 01), Future Orientation (r = .26, p < .05),
societies. In societies in which humane orienta- Power Distance (r = .30, p < .05), and
tion practices are relatively low, members of Uncertainty Avoidance (r = .29, p < .05). These
society express a desire to reach higher humane findings point out that as the humane orientation
orientation relationships. It is interesting to note practices of a society increase, it tends to have
that the four highest-rated countries with respect more collectivist, performance-oriented, and
to Humane Orientation values measures (band nonassertive practices. Many of the humane
A: Nigeria, Finland, Singapore, Austria) are all practices found in societies at the same time
in the C and D bands of the Humane Orientation reflect practices aimed at maximizing collective
practices measures, illustrating the negative interests and achievement in a supportive,
relationship between societal practices and nonassertive way. These data show that social
values in Humane Orientation. support and solidarity go hand in hand with
As indicated in Chapter 2, the GLOBE theo- promoting collective interests in a supportive
retical model postulates that societal practices manner in societies across the globe. In addi-
and values affect organizational practices and tion, those societies that have more humane-
values. Two hierarchical linear models (HLMs) oriented practices value assertiveness, masculinity,
were conducted to test these hypotheses for power distance, future orientation, and avoiding
organizational Humane Orientation practices uncertainty to a greater extent. The negative
and values. We tested the GLOBE hypothesis relationship of Humane Orientation practices
regarding the effect of societal culture on orga- with the values of Gender Egalitarianism,
nizational culture by conducting HLM analyses and the positive relationship with the values
in which organizational Humane Orientation of Assertiveness, remained significant even
was predicted by societal Humane Orientation. when we controlled for the Humane Orientation
These analyses supported our hypotheses that values, suggesting that the results are not
societal Humane Orientation practices have a just a reflection of the negative relationships
significant and strong positive relationship with between Humane Orientation values and
organizational Humane Orientation practices practices. A possible explanation lies in the
(p < .01). We found a trend between societal emphasis on paternalistic behaviors, associated
Humane Orientation values and organizational with men, and with an assertive protection of
Humane Orientation values ( p < .10). Both the rights of the members in many societies that
analyses support a principal proposition in the practice humane orientation.
GLOBE theoretical model (i.e., Proposition 3, When the relationship between the societal
Figure 2.1, Chapter 2, by House & Javidan); Humane Orientation values and the other
societal cultural values and practices affect GLOBE societal constructs is considered, it is
organizational cultural values and practices.2 found that there is a significant negative corre-
lation between the Humane Orientation values
and the values of Power Distance (r = −.42,
Societal Humane Orientation and p < .01), and a significant positive correlation
Other GLOBE Culture Dimensions with assertive practices (r = .35, p < .01). These
Table 18.5 shows the correlations between findings show societies that value humane
the societal Humane Orientation dimension and orientation seek to reduce power distance and
the other GLOBE societal constructs. Societal use rather assertive practices, possibly to fight
Humane Orientation practices are correlated at a against injustice.
statistically significant level with the GLOBE
societal practices scores of Institutional
Collectivism (r = .43, p < .01), In-Group HUMANE ORIENTATION AND OTHER
Collectivism (r = .30, p < .05), Performance ECONOMIC AND SOCIAL INDICATORS
Orientation (r = .25, p < .05), and Assertive-
ness (r = −.42, p < .01) scales and with the In understanding the conceptual foundations for
GLOBE societal values of Assertiveness (r = .45, antecedents to humane orientation, a multidisci-
p < .01), Gender Egalitarianism (r = −.39, plinary perspective is required. In this section,
18-House.qxd 3/19/2004 5:05 PM Page 576
Table 18.5 Correlations of Humane Orientation as a Societal Construct with Other GLOBE Societal
Dimensions
we analyze the relationship between the GLOBE considers the capitalist system inhumane in the
Humane Orientation construct and other major sense that there is alienation and detachment of
cross-cultural data in various disciplines. As mechanical workers from workplace decisions.
explained in Chapter 7 by Javidan and Hauser, The political ideology of a society influences
four major ongoing reports produce the rele- the belief about distribution fairness among
vant data: the IMD’s Global Competitiveness its members. Deutsch (1975) identified three
Ranking, the United Nations Development basic norms concerning distribution fairness:
Program’s Human Development Report, and the Equity norm, equality norm, and need norm.
World Values Survey. In addition, 1995 U.S. The equity norm is expected to produce high
census data and data from other cross-cultural average group productivity. The equality norm
research were used to compare and contrast leads to the best group harmony and stability.
GLOBE Humane Orientation scores with other The need norm is the most humanistic and most
indicators of the human condition. helpful to individual well-being and personal
development (Kaicheng, 1998).
In Marxism the norm of allocation of
Political Systems and Humane Orientation
resources according to needs is regarded as ideal
Two alternative political ideologies, liberal although unrealistic because of the requirement
and socialist, claim to create a humane-oriented for material wealth in the society. Because of the
society (Macpherson, 1966). The liberal ideol- difficulties in achieving distributive justice based
ogy concentrates on rationality, self-interest, on the need norm, the Chinese perspective is
and property rights. The socialist ideology based on egalitarianism. Kaicheng (1998) and
focuses on eliminating the exploitation of labor Mao (1929) pointed out that egalitarianism is an
by preventing concentration of means of illusion of peasants and small proprietors and
production in the hands of a few. As such, the even under socialism there can be no absolute
socialist ideology is more concerned with indi- equality. After Deng Xiaoping took control of
vidual welfare maximization and freedom. It China in the 1970s, he implemented a series of
18-House.qxd 3/19/2004 5:05 PM Page 577
decreases and the accompanying social control Family Values and Humane Orientation
among members of the society breaks down.
Correlations between the GLOBE Humane Based on the World Values Survey, no signif-
Orientation practices dimension and economic icant correlation between Humane Orientation
health indices such as economic prosperity, public practices and strength of family ties (r = .16, n = 38,
sector support for economic prosperity, societal p > .10) is found.
support for economic competitiveness, and over- With an attempt to scrutinize whether or not
all competitiveness ranking are not found to be there is a relationship between attributing an
statistically significant. However, as seen in Table economic value to children and GLOBE soci-
18.6, Humane Orientation practices scores are etal Humane Orientation scores, Spearman rank
negatively correlated with gross national product order correlations were conducted between the
(GNP) per capita (r = −.36, n = 54, p < .01). These two constructs for eight countries that appeared
results suggest that more industrially wealthy in both the GLOBE and the Value of Children
countries have less Humane Orientation practices datasets (Kagitçibasi, 1996). Results show that
in general. On the other hand, no relationship has there is a positive and statistically significant
been found between Humane Orientation values relationship between Humane Orientation prac-
and economic health indicators. tices of a society and economic value attributed
The Gini coefficient of a country (a measure to children (rs = .81, p < .01). Societies that
of income inequality) is not related to societal have relatively more helping behavior among
practices and values Humane Orientation scores. their members tend to perceive children as an
Although economic development is correlated economic value to a greater extent. On the other
with Humane Orientation, distribution of income hand, partial correlation analysis, which holds
within the country has no relation to it. GNP per capita constant, yields an insignificant
Behavioral norms in highly humane oriented, relationship between GLOBE societal Humane
collectivist, and economically less-developed Orientation practices and economic value attrib-
societies have some overlap. This overlap can uted to children. Whereas there seems to be no
be expected based on the statistically significant relationship between Humane Orientation and
correlation between GLOBE societal Humane economic value attributed to children if eco-
Orientation practices with the GLOBE societal nomic development is held constant, there is a
practices of Institutional Collectivism (r = .43, statistically significant negative correlation
p < .01), In-Group Collectivism (r = .30, p < .05), between GNP per capita and economic value
and with GNP per capita (r = −.36, p < .01). In attributed to children (rs = −.919, p < .01).
highly humane-oriented societies, the central Societies that have a lower level of economic
behavioral norms include helping others and development tend to be both more humane and
being generous and friendly. Further, the attribute more economic value to children.
Humane Orientation practices dimension was Our data show that Humane Orientation
negatively correlated with alcohol consumption practices in societies increase as economic
per capita (r = −.45, p < .05). development decreases. The negative corre-
In sum, in humane-oriented cultures, lation between GLOBE Humane Orientation
members of society are expected to provide practices and GNP per capita is still found
material and financial help, spend time together, to be statistically significant after the effect
extend empathy and love, and share information of GLOBE Institutional Collectivism is con-
that is required to solve problems effectively. trolled for by partial correlation analysis
In cultures reported in the GLOBE survey as (r = −.50, n = 58, p < .01), but the signifi-
having highly humane-oriented practices, if cance attenuates greatly after the effect of
individuals have a problem they are expected to GLOBE In-Group Collectivism is controlled
ask for help and others are expected to offer (r = −.22, n = 58, p < .10). It seems there is
help. On the other hand, in societies reporting a higher need for solidarity and help within
low levels of humane orientation, individuals the family and members of societies in less-
are expected to be more self-centered and pro- developed countries, which increases Humane
mote self-enjoyment and self-interests. Orientation practices at the societal level.
18-House.qxd 3/19/2004 5:05 PM Page 579
Table 18.6 Relationship Between Societal Humane Orientation Practices and Modernization Measures*
Urbanization
Human condition
Economic health
Thus, economic development, economic value “Convention on the Rights of the Child” are not
of children, expectations of obedience from correlated with the GLOBE societal Humane
children, and Humane Orientation seem to be Orientation scores. Data for membership statis-
related to each other. tics were obtained from The United Nations and
Human Rights, 1945–1995 (1995). Further, soci-
etal Humane Orientation values score was found
Human Rights and Humane Orientation
to be negatively correlated with prisoners per
The GLOBE societal Humane Orientation 100,000 people (r = −.56, p < .05), as given in
values dimension is found to be correlated the Human Development Report (United
with the number of years a society has been a Nations Development Program, 1998). The
member of the International Convention on the societies that are reported to value more Humane
Elimination of All Forms of Racial Discrimina- Orientation are less likely to keep a significant
tion (r = .34, p = .018). Those countries that proportion of their population in prison.
aspire to higher levels of Humane Orientation In summary, if GLOBE’s Humane Orienta-
seem to have longer membership in this tion findings are compared with the findings
convention. On the other hand, such societal from secondary data, those societies that scored
indicators as membership in “Convention higher in Humane Orientation practices have
Against Torture and Other Cruel, Inhuman, or lower life expectancy, score lower on the
Degrading Treatment or Punishment” and Human Development Index, are less developed
18-House.qxd 3/19/2004 5:05 PM Page 580
economically, and have lower preference for with climates that create hardship for survival
left-wing political ideology. In addition, those (e.g., tropical humid) have significantly higher
societies that score higher in Humane Humane Orientation practices. Table 18.7 shows
Orientation values are less inclined to imprison the mean scores for Humane Orientation prac-
people and are against societal discrimination. tices across climates.
Under exceptionally stressful times, when
physical conditions deteriorate markedly, sup-
HUMANE ORIENTATION AND port among members of a unit may diminish in
COUNTRY DEMOGRAPHICS many nonindustrial societies of tropical humid
climates. For instance, Shalinsky and Glascock
Physical Conditions, Climate, (1988) observe that when the Yanomamo of
northern Brazil perceive a societal threat, they
and Humane Orientation
may decide to kill expendable persons to stabi-
Parallel with the modernization hypothesis, lize their conditions. These societies clearly
in societies in which physical conditions and distinguish between the killing of infants and
climate create difficulties for survival there the aged and killing a fully developed human.
is likely to be a greater need for helping and The latter is considered homicide and society
solidarity among members of society. severely punishes those who execute a homi-
GLOBE countries were categorized by cide. Infanticide concerns only the family, espe-
climate into seven groups: tropical humid, cially the mother, and not any larger unit of
tropical wet and dry (savanna), desert, sub- society. These societies perceive benefit to the
tropical humid, subtropical wet and dry mother and the family from infanticide and
(Mediterranean), marine west coast (maritime), killing the aged.
and continental. These groups are discussed On the other hand, Shalinsky and Glascock
in detail in Chapter 10 by Gupta and Hanges. (1988) point to beliefs and practices in modern
One-way ANOVA tests were conducted to Western society that are parallel to the killing of
determine if there exist climatic differences in infants and the aged in nonindustrial societies:
the practices and values of societal Humane (a) the beliefs and practices regarding abortion
Orientation. There is modest evidence of and (b) active support of the suffering aged
climatic differences in the societal practices person’s right to die, and, at least in the United
F6, 54 = 2.18, p < .10), and no evidence of States, shunting off the decrepit aged to nursing
climatic differences in the societal values homes with extremely poor conditions and few
(F6, 54 = 0.81, p > .10) of Humane Orientation. humanitarian practices.
Climatic differences account for 20% of the
variation in societal Humane Orientation prac-
Humane Orientation and Region
tices scores, whereas 80% represent idio-
syncratic societal differences. Thus, climate has As discussed in Chapter 8 by Hanges and
a moderate influence on the observed societal Dickson, there are significant differences in val-
differences in Humane Orientation practices. ues and practices for every cultural dimension
We used Analysis of Variance contrast among the 62 societal cultures (i.e., recall that
tests to identify the climates that are reported to we use the term societal cultures rather than
have significantly different levels of Humane nations). As reported in Chapter 10 by Gupta
Orientation practices. Only one climatic region and Hanges, societies were grouped into 10
was found to be significant: Societies in tropical cultural clusters: (a) Eastern Europe, (b) Latin
humid climates tend to score higher for societal America, (c) Latin Europe, (d) Confucian Asia,
Humane Orientation practices in general (con- (e) Nordic Europe, (f) Anglo, (g) Sub-Saharan
trast = 0.39, p < .05). In tropical humid regions, Africa, (h) Southern Asia, (i) Germanic Europe,
the air is like that in a very hot sauna, and the and (j) the Middle East. One-way ANOVA tests
humidity causes a lot of perspiration and showed that there exist significant cross-
makes people tire quite easily. Thus, there is regional differences in the societal practices
some support for our conjecture that societies (F9, 51 = 5.06, p < .01), but not in societal
18-House.qxd 3/19/2004 5:05 PM Page 581
Climate N Mean SD
values (F9, 51 = 0.91, p > .10) of Humane people. Gratitude is a prized trait in Filipino
Orientation. Regional differences (country society and any help is valued as a debt of
clusters) account for 47% of the variation in honor. Most Germanic societies, on the other
societal Humane Orientation practices scores, hand, tend to view humane orientation in rather
whereas 53% represent idiosyncratic societal rational terms, often overburdened with formal
differences. In other words, there is substantial procedures such as those related to the condition
regional differentiation among societies with of labor.
respect to practices of humane orientation.
Table 18.8 illustrates the averages for each
Technology and Market
country cluster on Humane Orientation prac-
tices scores. We used Analysis of Variance
Development and Humane Orientation
contrast tests to identify the clusters with signif- The argument that development and indus-
icantly higher or lower practices of Humane trialization have produced processes that lead
Orientation. The Humane Orientatio practices to universal attitudes about work finds some
score is significantly higher in the Southern support in the literature (Harbison & Myers,
Asia cluster (mean = 4.71; contrast = 0.68, 1959; Kerr, Dunlop, Harbison, & Myers, 1960;
p < .01), followed by Sub-Saharan Africa (mean Ronen, 1986). According to the convergence
= 4.42, contrast = 0.35, p < .05). The Humane argument, industrialization affects all organi-
Orientation practices score is significantly zations the same way, and one outcome of
lower in Germanic Europe (mean = 3.55 con- economic development and technological
trast = 0.60, p < .01), followed by Latin Europe advancement is to produce organizational
(mean = 3.71, contrast = 0.43, p < .01). cultures that are similar and independent of
Gupta, Surie, Javidan, and Chhokar (2002) national culture.
observe that a group-oriented humane approach Organizations of industrialized Western
is the hallmark of Southern Asian societies and countries in the 20th century had predominantly
involves a delicate, nonassertive balancing of Taylorist or Fordist processes and structures.
power versus performance. They illustrate the The aim of the Taylorist-Fordist paradigm
humane orientation of Southern Asian societies was to limit the arbitrariness of the supervisors
using the example of the Philippines. In rural and employees on the shop floor by tying
Philippine, neighbors commonly offer to help if work rhythm to the pace of the assembly line.
a person is constructing a house. Filipinos love This paradigm created a uniform mode of
helping not just one another, but also other control over the labor process, which is heavily
18-House.qxd 3/19/2004 5:05 PM Page 582
Table 18.8 Regional Clusters and Societal Humane Orientation Practices Mean Scores**
criticized for its low humane orientation: Forced including the national institutions governing
pace and work pressure, physical strain, and labor policy.
psychological stress on workers (Friedman, Table 18.9 presents the summary statistics
1974; Widick, 1976). for societal Humane Orientation practices and
In the early 1980s, in response to the pres- values dimensions in three industries: financial
sures of global competition, the advantages of services, food processing, and telecommunica-
the Taylorist-Fordist control forms were increas- tions. As seen in Table 18.9, societal practices
ingly questioned. Leibenstein (1987) introduced and values scores for Humane Orientation in the
the concept of “X-inefficiency” to describe the finance industry have an average rating of 4.04
phenomenon of many incomprehensible ineffi- and 5.43, respectively. For the food industry,
ciencies in traditional mass production facilities. Humane Orientation practices score has an
There is a tendency toward a relative increase in average value of 4.07, and the mean values
skilled laborers and more flexible forms of work score is 5.42. In the case of telecommunication
design in high-technology areas in the postin- sector, mean Humane Orientation practices and
dustrial societies. Jürgens (1991) studied 17 values scores are 4.14 and 5.41, respectively.
assembly plants at three automobile companies Table 18.10 shows the correlations among
in the United States, Great Britain, and the societal Humane Orientation practices and values
Federal Republic of Germany. His results for finance, food, and telecommunication indus-
showed differing configurations of departure tries. The Humane Orientation values in the
from the traditional Tayloristic-Fordist mode of telecommunication industry correlate signifi-
work related to strategic choices based on the cantly with the Humane Orientation values in the
demands for adaptation recognized in the early finance (r = .48, p < .01) and the food sectors
1980s (p. 240). Further, he found that the tech- (r = .72, p < .01). Furthermore, the Humane
nological and human resource strategies chosen Orientation practices in the telecommunications
were strongly influenced by the national context, industry correlate significantly with the Humane
18-House.qxd 3/19/2004 5:05 PM Page 583
Table 18.9 Summary Statistics for Societal Humane Orientation for Three Industries
** p < .05.
** p < .01.
and standard wage rates forced on employers most efficient, if compared with feudal or
regardless of varying profitability and ability to patrimonial organizations. According to Marx,
pay (Martin, 1984). Nordic Europeans sought bureaucracy is the instrument of the capitalist
alternatives to replace the Taylorist paradigm, class and a means of exploitation.
which labor perceived as inhumane, and improve In bureaucratic organizations, human resource
their competitive position in the international practices are standardized and aim at the effi-
markets. cient running of the organization. On the other
Severe levels of dissatisfaction with work on hand, dominant paternalistic values in a society
the part of well-educated young Nordic people lead to practices in which concern for employ-
resulted in widespread absenteeism, tardiness, ees’ families and obligation to one’s commu-
turnover, and strikes in workplaces, indicating nity override systematic application of human
that private Nordic firms rather than the unions, resource practices.
the politicians, or the academics (Elam & Rather than the standard rules and regula-
Börjeson, 1991) were most active in seeking tions of a bureaucratic system, organizations
new ways of working that encouraged individ- in paternalistic cultures make exceptions for
ual involvement and growth of team spirit. In employees on the basis of their needs, and
many cases, codetermination agreements were managers and supervisors show concern for
signed, which incorporated demands for a more workers’ personal problems. In paternalistic
humane workplace in Nordic Europe. cultures, patronage relationships emerge as a
Jürgens (1991) further concludes that in the form of providing informal support, whereas in
process of dissolution of Taylorist-Fordist con- nonpaternalistic cultures mentoring relation-
trol system two models emerge: The German ships between more experienced and younger
model (or the Nordic-German model) of skilled employees may replace the patronage relation-
worker-oriented labor regulation, and the Japanese ship as a means of providing informal support.
model of group-oriented labor regulation. The According to Kanungo and Aycan (1997),
German model focuses on the technical aspect in paternalistic societies managers believe that
of skilled work, the concept of profession, and employees by nature seek more supervisory
increased responsibility and self-regulation in guidance. Hence, jobs are designed in a way that
carrying out the task. does not give much autonomy to employees.
The Japanese model likewise focuses on
skilled labor, but in its ideal form the Japanese
Employee Relations
worker was self-regulated under the pressure
of the assembly line and the machine pace. The concept of human resource management
Further, the Japanese system emphasizes soci- (HRM) was developed initially on the basis of
etal context and cultural means to reinforce American practices in the 1960s and 1970s.
the importance of group solidarity and sup- Central to the notion of HRM as currently
port in Japanese organizations. According to accepted in the literature is the notion of orga-
Jürgens (1991), American companies were nizational independence and autonomy, which
more inclined to adapt the Japanese concepts may not be prevalent in many parts of the
of group support than the German emphasis world. In highly humane-oriented societies,
on technical expertise and professional ethic in organizations are relatively autonomous in their
the de-Taylorization process. employee relations, whereas in the relatively
low humane-oriented societies there is greater
Bureaucratic Versus Organic control of organizations in their HRM practices.
The European system is characterized by
Organizations and Organization Design
the significant role of social partners in the
An ideally rational organization, in the employment relationship (Sparrow & Hiltrop,
Weberian sense, is an organization performing 1994). These influences include the role of trade
its tasks with maximum efficiency. The com- unions in the setting of HRM policies, collective
plexity and size of modern administrative tasks bargaining at the state and regional levels, and
make the bureaucratic type of organization the direct codetermination at the company level. In
18-House.qxd 3/19/2004 5:05 PM Page 585
many European countries union recognition for manufacturing strategies, HRM practices,
collective bargaining is required by law. employee relations, and societal culture.
Table 18.12 summarizes organizational
Humane Orientation for three industrial sectors.
Relationships to Stakeholders
Organizational Humane Orientation practices are
In the relatively low humane-oriented weakest in the financial services sector, in which
societies such as the European countries, human economic logic is expected to dominate more
social life is not viewed as an economic transac- strongly, than in the food and telecommunications
tion but rather from the sociological perspec- sectors. The same is true with respect to organiza-
tive, and there is lower emphasis on contractual tional Humane Orientation values also, though the
sale of labor by employees (Schneider & inter-industry differences here are much smaller.
Barsoux, 1997; Sparrow & Hiltrop, 1994). From To analyze whether organizational Humane
this perspective, organizations are not expected Orientation practices and values differ on
to make a profit without fulfilling social respon- the basis of region (i.e., cultural clusters),
sibility and giving serious consideration to an analysis of variance test was used. There
all stakeholders, not only the shareholders. are significant between-region differences in
Compared with the United States and Japan, in organizational Humane Orientation practices
Europe there is greater influence of trade unions (F9, 264 = 4.39, p < .01); cultural region
and the state on the business system. accounts for 13% of the between-organization
In the more humane-oriented societies of differences in Humane Orientation practices.
the world, satisfying shareholders and making Similarly, there are significant between-
a profit is the primary focus of organizations. region differences in organizational Humane
Making profits is a legitimate goal that is shared Orientation values (F9, 264 = 4.59, p < .01);
in organizations. Although social responsibility cultural region accounts for 13.5% of the
is left to the initiative of senior managers and between-organization differences in Humane
owners, some organizations apply the charity Orientation values.
and stewardship principles and corporate social In addition, a series of one-way analyses of
responsiveness (Frederick, 1987). variance were conducted for each industry.
Table 18.13 shows the regional breakdown
of organizational Humane Orientation scores for
Manufacturing Strategy
the three industries. There exist strong statisti-
Advanced technology highlighted the need cally significant differences in organizational
for flexible manufacturing systems such as Humane Orientation practices across regions in
just-in-time and computer-aided processes. the financial services industry (F9, 120 = 2.90,
Post-Taylorist and post-Fordist manufacturing p < .01, eta sq = 0.18) and telecommunications
processes and ideas integrated the skilled labor industry (F8, 44 = 3.00, p < .01, eta sq = 0.35),
and greater autonomy required by these sys- and weak differences in the food industry (F9, 81
tems, relieving workers from the physical and = 1.73, p < .10, eta sq = 0.16).
psychological strains of machine-paced mass When practices are investigated, there are
production systems. In this way they create two cultural clusters with significantly higher
more humane organizational workplaces. Humane Orientation scores in their organiza-
Ironically, although the societies of Western tions: Sub-Saharan Africa (mean = 4.83, contrast
Europe score relatively low on the dimension of = .38, p < .01), followed by Nordic Europe
Humane Orientation, these humane organiza- (mean = 4.72, contrast = .22, p < .05). Similarly,
tional structures are prevalent there. there are two cultural clusters with significantly
Table 18.11 summarizes the major connota- lower scores for Humane Orientation in their
tions and variations of the Humane Orientation organizations: Latin Europe (mean = 4.15, con-
dimension at the organizational level. It particu- trast = −.38, p < .01), followed by Eastern
larly addresses technology and development, Europe (mean = 4.21, contrast = −.31, p < .01).
economic context, organizational design, trust In the case of organizational values also, one-
versus control, relationships to stakeholders, way analysis of variance revealed statistically
18-House.qxd 3/19/2004 5:05 PM Page 586
Table 18.11 Summary of Major Connotations and Variations of the Humane Orientation Differences in
Terms of Organizational Practices and Values
Social control based on shared values Social control based on bureaucratic practices.
and norms.
Organizations are trusted more and are Organizations are controlled by legislation and
autonomous in human resource practices. unionization.
Organizations are relatively autonomous in their Organizations are restricted in their employee
employee relations. relations by the concept of social partners.
Less influence of trade unions and the state on Greater influence of trade unions and the state on
the business system. the business system.
Higher emphasis on contractual sale of labor. Lower emphasis on contractual sale of labor.
Organizational members prefer to work with others Organizational members prefer to be left alone to
to get jobs done. get jobs done.
Table 18.12 Summary Statistics for Organizational Humane Orientation for Three Industries
NOTE: N = number of organizations. Only identifiable organizations with at least 7 respondents are included in the analysis.
18-House.qxd 3/19/2004 5:05 PM Page 587
NOTE: Only organizations with at least 7 respondents are included in the analysis.
significant differences in Humane Orientation values. Latin European organizations also report
across regions in the financial services sector low Humane Orientation practices scores,
(F9, 120 = 2.69, p < .01, eta sq = 0.17), telecom- whereas the Germanic European and Middle
munications industries (F8, 44 = 2.35, p < .05, Eastern organizations scored low in Humane
eta sq = 0.30), as well as food industries (F9, Orientation values.
81 = 2.70, p < .01, eta sq = 0.23).
The cultural clusters with significantly
higher organizational Humane Orientation HUMANE ORIENTATION AS A
values scores are Nordic Europe (mean = 5.36, LEADERSHIP CHARACTERISTIC
contrast = .27, p < .01), followed by Sub-Saharan
Africa (mean = 5.24, contrast = 0.26, p < .05), Various studies have shown that satisfaction of
whereas Eastern Europe (mean = 4.78, contrast belongingness and social needs can be achieved
= −.25, p < .01), followed by Germanic Europe with considerate leadership. Although there are
(mean = 4.81, contrast = −.22, p < .01) and no empirical leadership studies that we know of
Middle East (mean = 4.83, contrast = −.19, that specifically use humane orientation as a
p < .05) have the significantly lower scores. leadership behavior, there are other leadership
In summary, Nordic European and Sub- constructs that come close. Much of the early
Saharan African organizations stand out in the research that reflects aspects of humane orienta-
GLOBE sample as having significantly higher tion leadership behaviors was conducted by
scores for practices as well as values for Ohio State University researchers in the United
Humane Orientation. On the other hand, Eastern States. Ohio State researchers found that subor-
European and Germanic European organizations dinates perceived their supervisors’ behavior
stand out as having significantly lower scores primarily in terms of two broadly defined
for Humane Orientation practices as well as categories, which are labeled consideration and
18-House.qxd 3/19/2004 5:05 PM Page 588
initiating structure (Fleishman, 1953; Halpin & Mexico, and the United States—Dorfman and
Winer, 1957; Hempill & Coons, 1957). A sec- colleagues (1997) found supportive leadership
ond major research program on humane- led to subordinate satisfaction in all samples.
oriented leadership behavior was carried out by The universality of leader supportiveness
researchers at the University of Michigan at the should not be surprising, because a leader
same time as the Ohio State leadership studies demonstrates supportiveness through a humane
(Katz, Maccoby, & Morse, 1950; Katz & Kahn, concern for followers. The effects of supportive
1952; Katz, Maccoby, Gurin, & Floor, 1951). leadership were strongest in Mexico and Japan.
Three types of leadership behaviors influence Interestingly, supportive leadership influenced
effectiveness of leaders: task-orientation, rela- job performance only in Mexico.
tion-orientation, and participative leadership From the above review, it appears that
(Likert, 1961, 1967). Relation-oriented behav- humane orientation of leaders through con-
iors correspond to consideration behaviors that siderate and supportive actions is a culturally
were used in the Ohio State studies. generalizable phenomenon. That is, across all
Perhaps one of the most robust research find- cultures reviewed in the above studies, leaders
ings in the field of leadership is the relationship can be compared in terms of the degree to which
between supportive or considerate leadership they enact humane behaviors toward subordi-
behaviors and employee satisfaction with super- nates and followers of leaders. This, however,
visors (Yukl, 2002). Fisher and Edwards’s does not imply that societies value humane
(1988) meta-analysis of studies found that the orientation to the same degree.
correlation between supervisors’ consideration
with overall job satisfaction of subordinates
Image of Leaders
and satisfaction with supervisor to be very high.
Schriesheim (1982) found that consideration by Various studies conducted across countries
itself accounted for most of the influence on show that business leaders often do not have
the satisfaction of employees; initiating struc- an image reflecting a humane orientation. In a
ture added only 2%–4% to satisfaction after the 12-country study, Bass and his associates (Bass,
effects of consideration were controlled in a Burger, Doktor, & Barrett, 1979) asked man-
multiple regression analysis. A meta-analysis agers to judge the importance of generosity as
by Wofford and Liska (1993) also supports the a requirement for lower, middle, and top man-
overall importance of supportive–considerate agement. In general, generosity was judged to be
leadership. However, it is important to note that of low importance. The faster climbing man-
most studies used in meta-analyses were based agers were less rejecting of generosity as an
on male United States samples. important trait compared with managers with
low rate of advancement. In almost all countries,
managers were less rejecting of generosity in top
Overview of Prior Cross-Cultural Studies
management than in middle and lower manage-
on Humanely Oriented Leadership
ment levels. These results indicate that more
Findings derived from a 30-year research successful managers seemed to attribute more
endeavor conducted in Japan indicate two fac- importance to being generous and that top man-
tors of leadership behavior as similar to the agement was more associated with generosity
Ohio State and Michigan studies of the United compared with lower management levels.
States (Misumi & Peterson, 1985). These stud- According to an eight-country cross-
ies, which were conducted in Japan, were for- cultural comparison of leadership prototypes
mulated around the performance-maintenance conducted by Gerstner and Day (1994), leaders
(PM) theory of leadership. The consideration are not perceived to have highly humane-
measures used in field studies in the United oriented traits. Subjects in this study were cho-
States were quite similar to the maintenance sen from American, Chinese, French, German,
measures used in Japan. Honduran, Indian, Japanese, and Taiwanese
In a study of five countries in the Asian- students enrolled in a graduate program at an
Pacific Basin—Japan, South Korea, Taiwan, American university. They were presented with
18-House.qxd 3/19/2004 5:05 PM Page 589
Table 18.14 Major Implications of Variations in Societal Humane Orientation for Humane-Oriented
Leadership
Duty orientation as a life-goal has high priority. Duty orientation as a life-goal has low priority.
Generous and compassionate leader attributes Generosity and being compassionate do not
contribute to leader effectiveness. contribute to leader effectiveness.
Holistic concern for the followers. Limited concern for the followers.
Maintenance behaviors involve less task Maintenance behaviors involve more task
orientation and consultation. orientation and consultation.
Relationships with subordinates are more Relationships with subordinates are more formal
informal and personal. and impersonal.
maintenance and performance functions in the involve more task orientation and consultation
Japanese case: Similar to high M supervisors, that is geared toward improving the task. Table
high P leadership behaviors included teaching 18.14 presents the major implications of variation
new job skills and discussing difficulties in the in the societal Humane Orientation dimension for
person’s absence. leadership.
The separation between the task and main-
tenance functions was most distinctive in the
United States case. In other cultures, task- Measures
oriented behaviors indicating consideration
As explained in Chapter 21 by Dorfman,
or considerate behaviors were used to facilitate
Hanges, and Brodbeck, one of the six second-
task orientation. On the other hand, the British
order leadership factors identified in the
supervisors who exhibited strong M behaviors
GLOBE study is Humane-Oriented leadership.
demonstrated more task-centered and consultative
Respondents indicated on a 7-point scale the
behaviors compared with the other three
degree to which attributes such as being gener-
countries. This finding fits the relatively low rank-
ous and compassionate contribute to or hinder
ing of Britain in terms of Humane Orientation
effective leadership, where 7 = contributes
societal practices scores. For example, high M-
greatly to outstanding leadership and 1 = greatly
oriented supervisors in Britain demonstrated or
inhibits outstanding leadership.
used equipment, consulted widely about neces-
sary changes, expected suggestions for work
improvements from subordinates, and responded
Findings
positively to new suggestions. Thus, it can be pro-
posed that in the less humane-oriented societies, Table 18.15a presents the summary statistics
maintenance-oriented leadership behaviors would for perceived effectiveness of Humane-Oriented
18-House.qxd 3/19/2004 5:05 PM Page 591
Humane-Oriented Standard
Leadership Mean Deviation Minimum Maximum Valid n
Table 18.15b Intercorrelations Among Humane Oriented Leadership Scores in Three Industries
Telecommunication industry
Table 18.16 Perceived Effectiveness Scores for Humane-Oriented Leadership Based on Regions (Cultural
Clusters)
those societies with high Humane Orientation 6.09. The mean score for the Humane-Oriented
practices scores, like Southern Asia and Sub- leadership dimension was 4.89 and the range
Saharan Africa, would endorse Humane was 3.82 to 5.75. The findings regarding the per-
Oriented leadership more. The regions with sig- ceived effectiveness of Humane-Oriented lead-
nificantly lower scores are Nordic Europe ership show that, in general, humane-oriented
(mean = 4.42, contrast = −.49, p < .05) and traits are perceived to contribute slightly to
Latin Europe (mean = 4.45, contrast = −.46, outstanding leadership; however, there are clear
p < .01). Table 18.16 presents the effectiveness differences among societal cultures and clusters
scores for Humane-Oriented leadership on the regarding this leadership dimension.
basis of clusters.
Charismatic/VB (O)
Cultural Values
Team Oriented (O)
Participative (S,O)
• Concerned +
about others Humane Oriented (O)
• Sensitive
toward others
• Being
friendly
Autonomous (O)
• Tolerating
mistakes Self-Protective (S)
−
Figure 18.1 Humane Orientation Cultural Values as Drivers for CLT Leadership Dimensions
* Only statistically significant relationships are shown (p < .05; see Table 21.10, Chapter 21 by Dorfman et al.). The most
important leadership CLT relationships are bolded (i.e., relationship is significant at both society and organization levels of
analyses or highest HLM coefficient within each level of analysis).
VB = Value-Based
O = Organizational level
S = Societal level
contrast to practices, which represent the world We tested for the relationship between
As Is. As described earlier in this volume, cul- culture and the CLT leadership dimensions by
ture is believed to have its effect on the content using hierarchical linear modeling (HLM).
of CLTs at multiple levels of analysis. Leader HLM is a multistep process designed to test the
attributes perceived to be effective might be a significance of relationships between indepen-
function of (a) societal cultural practices, dent and dependent variables at multiple levels
(b) societal cultural values, (c) organizational of analyses. An overview of HLM analyses and
cultural practices, and (d) organizational cultural a detailed discussion of how these analyses
values. In the present chapter we discuss the were conducted is provided in Chapter 11 by
results of statistical analyses examining the Hanges, Dickson, and Sipe. In the next section,
extent to which the cultural dimension of we report the HLM results exploring the rela-
Humane Orientation has an effect on CLTs. tionship between organizational and societal
Specifically, we examine the extent to which the culture and the content of CLT leadership
content of CLTs varies as a function of Humane dimensions.
Orientation values and practices in societies and
organizations within societies. As you read HLM Analysis: Organizational
through the results discussed below, it may be
and Societal Variation
helpful to view Figure 18.1 for a visual sum-
mary. The figure, however, shows only results In the present analysis we examine the asso-
regarding cultural values, not practices. (All ciation of organizational and societal Humane
HLM coefficients are presented in Table 21.10 Orientation values and practices with the six
of Chapter 21 by Dorfman et al.) GLOBE CLT dimensions. The total amount of
18-House.qxd 3/19/2004 5:05 PM Page 594
organizational and societal variance explained organizational and societal variance for this
by Humane Orientation ranged from 4.1% to dimension. Of the total explained variance,
27.6%. Humane Orientation was found to be a approximately 16.8% was associated with forces
better predictor for some of the CLTs (e.g., operating at the organizational level of analysis.
Humane-Oriented leadership) than for others The majority of the explained variance (83.2%)
(e.g., Autonomous leadership). was associated with forces operating at the soci-
When organizational level and societal level etal level of analysis.
Humane Orientation values and practices were The organizational Humane Orientation cul-
considered, significant relationships were found tural values scores were positively related
with the following CLTs: (p < .01) to the Participative leadership dimen-
sion. Participative leadership is more likely to
• Charismatic/Value-Based leadership. be a part of the shared leadership belief system
Humane Orientation cultural values scores in organizations reported to espouse Humane
were significantly related to the Charismatic/ Orientation values.
Value-Based leadership dimension and explained The societal Humane Orientation cultural
a total of 5.6% of organizational and societal values scores were positively related (p < .01)
variance for this dimension. All of the explained to the Participative leadership dimension.
variance was associated with forces operating Participative leadership is more likely to be a
at the organizational level of analysis; only part of the shared leadership belief system
the organizational Humane Orientation cultural in societies reported to espouse Humane
values scores were positively related (p < .01) to Orientation values.
the Charismatic/Value-Based leadership dimen-
sion. Charismatic/Value-Based leadership is • Humane-Oriented leadership. Humane
more likely to be a part of the shared leadership Orientation cultural practices and values scores
belief system in organizations reported to were significantly related to the Humane-
espouse Humane Orientation values. Oriented leadership dimension and explained a
total of 27.6% of organizational and societal vari-
• Team-Oriented leadership. Humane
ance for this dimension. Of the total explained
Orientation practices and values scores were
variance, approximately 70.9% was associated
significantly related to the Team-Oriented
with forces operating at the organizational level
leadership dimension and explained a total of
of analysis. The remaining portion of the
12.2% of the organizational and societal vari-
explained variance (29.1%) was associated with
ance for this dimension. All of this variance was
forces operating at the societal level of analysis.
associated with forces operating at the organiza-
The organizational Humane Orientation
tional level of analysis. The organizational
cultural practices scores were negatively related
Humane Orientation cultural practices scores
(p < .01) to the Humane-Oriented leadership
were negatively related (p < .01) to the Team-
dimension, but the organizational Humane
Oriented leadership dimension, whereas the
Orientation cultural values scores were posi-
organizational Humane Orientation cultural
tively related (p < .01) to the Humane-Oriented
values scores were positively related (p < .01)
leadership dimension. Humane-Oriented leader-
to the Team-Oriented leadership dimension.
ship is less likely to be a part of the shared lead-
Team-Oriented leadership is less likely to be
ership belief system in organizations reported to
a part of the shared leadership belief system
have Humane Orientation practices. This finding
in organizations reported to have Humane
may suggest that if organizational practices are
Orientation practices, but more likely to be in
highly humane, the need for humane-oriented
the leadership belief systems of organizations
leadership is less. Finally, Humane-Oriented
reported to espouse Humane Orientation values.
leadership is more likely to be in the shared lead-
• Participative leadership. Humane Orien- ership belief systems of organizations reported
tation cultural values scores were significantly to espouse Humane Orientation values.
related to the Participative leadership dimen- The societal Humane Orientation cultural
sion and explained a total of 19.3% of the practices scores were positively related (p < .01)
18-House.qxd 3/19/2004 5:05 PM Page 595
least Humane-Oriented clusters are Latin Europe in the control of organizations by promoting
and Nordic Europe. In societies scoring high in legislation, unionization, and codetermination
Humane Orientation, leaders have a holistic con- in employee relations and human resource prac-
cern for followers, and relations to subordinates tices. In the more humane-oriented societies,
are informal and personal. If the joint effect of organizations are granted more trust and power
organizational and societal culture is considered, and are more autonomous in their relations
societal values and practices had a weaker impact to employees and other institutions. Fourth,
on the evaluations of leadership attributes com- organizations reflect the culture (practices and
pared with organizational culture. If the influence values) in the society in which they are embed-
of societal culture is considered, the more a ded. Humane societies foster humane organi-
society values Humane Orientation, the higher zations. Fifth, Humane-Orientation practices
the members of that society rate Participative in a society influence perceived effectiveness
leadership as a desired attribute and the lower of Humane-Oriented leader behaviors. The
they rate Self-Protective leadership. As Humane more societies practice humaneness, the more
Orientation practices in a society increase, the Humane-Oriented leader behaviors are per-
likelihood of evaluating Humane-Oriented lead- ceived to be effective. Finally, leaders in the less
ership as an effective attribute increases. humane-oriented cultures are more formal and
Humane Orientation values and practices in impersonal in their relationships to their follow-
organizations relate significantly to all of the lead- ers because the organizations in which they
ership dimensions. In general, the less organiza- operate are more controlled by formalized poli-
tions practice Humane Orientation and the more cies, procedures, unionization, and legislation.
they value it, the more likely it is that Humane- The Humane Orientation construct has
Oriented, Team-Oriented, Charismatic/Value- implications for the new international scene. In
Based, and Participative leadership are perceived response to trends in globalization and regional-
as outstanding attributes, and the less likely ism, companies operating in world markets need
Autonomous leadership is rated as outstanding by to understand the level of humane orientation
the members of that organization. This overall that exists in various cultures. In practices of
finding likely reflects dissatisfaction with the cur- multinational corporations, strategic alliances,
rent state of affairs in organizations that do not use and joint ventures, the construct may influence
humane orientation practices as a way of organi- negotiations, employee relations, human resource
zational life. practices, and leader behaviors. Clusters of
Our study, on the basis of a through literature world markets based on humane orientation
review of secondary data and the original may reveal strategic guidelines for firms enter-
GLOBE data, presents several generalizations ing into different countries. Thus, Humane
about Humane Orientation at the societal, Orientation as a construct becomes a valuable
organizational, and leadership levels. First, analytical dimension for international business
Humane Orientation at the societal level seems as well as for organization theories.
to increase under more difficult economic,
physical, and climatic conditions. Help and
generosity among members of society become ENDNOTES
the necessary norms for life survival and well
being. As a result of social solidarity, the human 1. Our response bias correction procedure identi-
condition in the form of psychological and fied response bias in some countries for this scale.
physical health improves in highly humane- We recomputed the predicted response bias corrected
oriented societies. Second, because the norms scale score for each country. Response bias corrected
regarding solidarity and helping behavior are scores are:
less endorsed in less humane-oriented societies, Practices: Morocco, 4.52 (no change in band);
the state more often and more strongly inter- Qatar, 4.79 (moves from band B to band A); and
venes to protect and give social support and Taiwan, 3.82 (no change in band).
security to its members. Third, in the less Values: France, 5.91 (moves from band B to band
humane-oriented societies, the state intervenes A); New Zealand, 4.85 (moves from band E to band
18-House.qxd 3/19/2004 5:05 PM Page 598
D; furthermore, this move eliminates band E); and Bass, B. M., Burger, P. C., Doktor, R., & Barrett,
Nigeria, 5.71 (moves from band A to band B). G. V. (1979). Assessment of managers. New
For a complete discussion of this procedure and York: Free Press.
all response bias corrected scores, see Appendix B. Bigoness, W. J., & Blakely, G. L. (1996). A cross-
2. As reported in Chapter 20 by Brodbeck, national study of managerial values. Journal of
Hanges, Dickson, Gupta, and Dorfman, we found that International Business Studies, 27(4), 739–753.
all the cultural dimensions of organizational cultural Brewster, C. (1995). Towards a “European” model of
values and practices significantly differed across human resource management. Journal of
societies. Although important, this prior analysis did International Business Studies, 26(1), 1–23.
not identify the particular aspect of societal differ- Briggs, A. (1961). The welfare state in historical per-
ences that was related to organizational culture. In the spective. Archives Europeennes de Sociologie,
present chapter, we found that societal and organiza- 2, 221–258.
tional Humane Orientation practices were signifi- Brouwer, I., & Wierda, E. (1998). The combination
cantly related (R2 Total = 2.7%, R2 Societal = 37.0%, model: Child care and part time labour supply
p < .01). We also found a trend between societal and of men in the Dutch welfare state. In J. J.
organizational Humane Orientation values (R2 Total Schippers, J. J. Siegers, & J. de Jong-Gierveld
= 0.8%, R2 Societal = 10.5%, p < .10). As discussed (Eds.), Child care and female labour supply in
in Chapter 11 by Hanges, Dickson, and Sipe, the R2 the Netherlands: Facts, analyses, policies.
Total considers all levels of analysis (i.e., individual, Amsterdam: Thela Thesis.
organizational, and societal) whereas the R2 Societal Bulatao, R. A. (1979). On the nature of the transition
isolates the societal-level portion of the dependent in the value of children (Publication No. 60-A).
variable and indicates the percentage of variance Honolulu: East-West Center.
accounted for by the predictor at only this level. Cordes, C., & Dougherty, T. (1993). A review and an
Although we have primarily taken the conservative integration of research on job burnout. Academy
approach and reported the R2 Total in GLOBE, of Management Review, 18, 621–656.
several scholars suggest that R2 Societal provides a Darroch, R. K., Meyer, P. A., & Singarimbun, M.
more accurate description of aggregated relation- (1981). Two are not enough: The value of children
ships. For further discussion, see the paper by Lance to Javanese and Sundanese parents (Publication
and James (1999). No. 60-D). Honolulu: East-West Center.
3. Results between the single HLM and multiple Deutsch, M. (1975). Equity, equality, and need: What
HLM tests will likely differ somewhat. The differ- determines which values will be used as the
ences between the results of the multiple HLMs and basis of distributive justice? Journal of Social
single HLMs are conceptually similar to the differ- Issues, 31, 137-149.
ences between a multiple regression analysis and a Donnelly, J. (1989). Universal human rights in theory
correlation coefficient. Table 21.10 in Chapter 21 by and practice. Ithaca, NY: Cornell University Press.
Dorfman et al. presents both single and multiple Dorfman, P. W., Howell, J. P., Hibino, S., Lee, J. K.,
HLM coefficients. In addition, the relationships for Tate, U., & Bautista, A. (1997). Leadership
all culture dimension values are summarized in in Western and Asian countries: Common-
Chapter 3 by Javidan and House. alities and differences in effective leadership
processes across cultures. Leadership Quarterly,
8(3), 233–274.
REFERENCES Dover, K. J. (1980). Plato: Symposium. New York:
Cambridge University Press.
Adams, G. R., & Jones, R. M. (1983). Female adoles- Elam, M., & Börjeson, M. (1991). Workplace reform
cents’ identity development: Age comparisons and the stabilization of flexible production in
and perceived child-rearing experiences. Sweden. In B. Jessop, K. Nielsen, H. Kastendiek,
Developmental Psychology, 19, 249–256. & O. K. Pedersen (Eds.), The politics of flexibil-
Allardt, E. (1993). Having, loving, being: An alterna- ity (pp. 314-337). Cheltenham, UK: Edward
tive to the Swedish model of welfare research. In Elgar.
M. C. Nussbaum & A. Sen (Eds.), The quality of Erikson, R. (1993). Descriptions of inequality:
life (pp. 88–94). Oxford, UK: Clarendon. The Swedish approach to welfare research. In
18-House.qxd 3/19/2004 5:05 PM Page 599
M. C. Nussbaum & A. Sen (Eds.), The quality of Harbison, F., & Myers, C. A. (1959). Management in
life (pp. 67–83). Oxford, UK: Clarendon. the industrial world: An international study.
Esping-Andersen, G. (1985). Politics against mar- New York: McGraw-Hill.
kets. The social democratic road to power. Hardy, F. (Ed.). (1988). The religions of Asia.
Princeton, NJ: Princeton University Press. London: Routledge.
Fawcett, J. T. (1983). Perceptions of the value of Hempill, J. K., & Coons, A. E. (1957). Development
children: Satisfactions and costs. In R. A. of the leader behavior description questionnaire.
Bulatao, R. D. Lee, P. E. Hollerbach, & In R. M. Stogdill & A. E. Coons (Eds.), Leader
J. Bongaarts (Eds.), Determinants of fertility in behavior: Its description and measurement.
developing countries (Vol. 1, pp. 367-369). Columbus: Ohio State University, Bureau of
Washington DC: National Assembly Press. Business Research.
Ferrari, G. R. F. (1987). Listening to the cicadas: A Hirsch, J. (1991). From the Fordist to the Post Fordist
study of Plato’s Phaedrus. New York: Cambridge state. In B. Jessop H. Kastendiek, K. Nielsen, &
University Press. O. K. Pedersen, (Eds.), The politics of flexibil-
Fisher, B. M., & Edwards, J. E. (1988). Consideration, ity: Restructuring state and industry in Britain,
initiating structure and their relationships with Germany, and Scandinavia (pp. 67–81).
leader effectiveness: A meta-analysis. In Best Cheltenham, UK: Edward Elgar.
paper proceedings (pp. 201–205). Anaheim, CA: Hoelter, J., & Harper, L. (1987). Structural and
Academy of Management. interpersonal family influences on adolescent
Fleishman, E. A. (1953). The description of supervi- self-conception. Journal of Marriage and the
sory behavior. Personnel Psychology, 37, 1–6. Family, 49, 129–137.
Frederick, W. C. (1987). Corporate social responsibil- Hoffman, L. W. (1987). The value of children
ity and business ethics. In S. P. Sethi & C. M. to parents and child rearing patterns. In
Falbe (Eds.), Business and society (pp. 142–161). C. Kagitçibasi (Ed.), Growth and progress in
Lexington, MA: Lexington Books. cross-cultural psychology (pp. 159-170. Lisse,
Friedman, G. (1974). Industrial society: The emer- The Netherlands: Swetz and Zeitlinger.
gence of the human problems of automation. Hofstede, G. (1980). Culture’s consequences:
Toronto: Ayer. International differences in work-related values.
Gebert, D., & Steinkamp, T. (1991). Leadership style Beverly Hills, CA: Sage.
and economic success in Nigeria and Taiwan. Hofstede, G. (2001). Culture’s consequences:
Management International Review, 31, 161–171. Comparing values, behaviors, institutions, and
Gerstner, C. R., & Day, D. V. (1994). Cross-cultural organizations across nations (2nd ed.).
comparison of leadership prototypes. Leadership Thousand Oaks, CA: Sage.
Quarterly, 5(2), 121–134. House, R. J., Hanges, P. J., Ruiz-Quintanilla, S. A.,
Grootaert, C., & Kanbur, R. (1995). Child labor: An Dorfman, P. W., Javidan, M., Dickson, M. W.,
economic perspective. International Labor Gupta, V., et al. (1999). Cultural influences
Review, 134(2), 187–203. on leadership and organizations: Project GLOBE.
Güngör, S. (1997). Effects of job stressors and social In W. H. Mobley, M. J. Gessner, & V. Arnold
support on burnout in a Turkish sample. Unpub- (Eds.), Advances in global leadership
lished master’s thesis, Bogaziçi University, (pp. 171–233). Stamford, CT: JAI.
Istanbul, Turkey. House, R. J., Wright, N., & Aditya, R. N. (1997).
Gupta, V., Surie, G., Javidan, M., & Chhokar, J. Cross-cultural research on organizational
(2002). Southern Asia cluster: Where the old leadership: A critical analysis and a proposed
meets the new? Journal of World Business, 37, theory. In P. C. Earley & M. Erez (Eds.). New
16–27. perspectives on international/organizational
Halpin, A. W., & Winer, B. J. (1957). A factorial psychology (pp. 535–625). San Francisco: New
study of the leader behavior descriptions. In Lexington Press.
R. M. Stogdill & A. E. Coons (Eds.), Leader International Labor Organization. (1992). World
behavior: Its description and measurement. labour report 1992. Geneva: Author.
Columbus: Ohio State University, Bureau of James, K., Chen, D. L., & Cropanzano, R. (1996).
Business Research. Culture and leadership among Taiwanese and
18-House.qxd 3/19/2004 5:05 PM Page 600
U.S. workers: Do values influence leadership Katz, D., Maccoby, N., Gurin, G., & Floor, L. (1951).
ideals? In M. N. Ruderman, M. W. Hughes- Productivity, supervision, and morale among
James, & S. E. Jackson (Eds.), Selected research railroad workers. Ann Arbor: University of
on work team diversity (pp. 33–52). Greensboro, Michigan, Survey Research Center.
NC: Center for Creative Leadership. Katz, D., Maccoby, N., & Morse, N. C. (1950).
Jessop, B. (1991). The welfare state in the transition Productivity, supervision, and morale in an
from Fordism to post-Fordism. In B. Jessop, office situation. Ann Arbor: University of
H. Kastendiek, K. Nielsen, & O. K. Pedersen, Michigan, Institute for Social Research.
(Eds.), The politics of flexibility: Restructuring Kerr, C., Dunlop, J. T., Harbison, F., & Myers, C. A.
state and industry in Britain, Germany, and (1960). Industrialism and industrial man.
Scandinavia (pp. 82–105). Cheltenham, UK: Cambridge, MA: Harvard University Press.
Edward Elgar. Kiray, M. (1997). Abandonment of the land and
Jürgens, U. (1991). Departures from Taylorism and transformation to urban life. In Human develop-
Fordism: New forms of work in the automo- ment report: Turkey. Ankara, Turkey: The
bile industry. In B. Jessop, K. Nielsen, H. United Nations Development Programme.
Kastendiek, & O. K. Pedersen (Eds.), The pol- Korte, C. (1980). Urban-nonurban differences in social
itics of flexibility: Restructuring state and behavior and social psychological models of
industry in Britain, Germany, and Scandinavia urban impact. Journal of Social Issues, 36, 29–51.
(pp. 233–247). Cheltenham, UK: Edward Lance, C. E., & James, L. R. (1999). A proportional
Elgar. variance accounted for index for some cross-
Kabasakal, H., & Bodur, M. (1998). Leadership and level and person-situation research designs.
culture in Turkey: A multifaceted phenomena. Organizational Research Methods, 2, 395–418.
Unpublished manuscript. Leibenstein, H. (1987). Inside the firm: The inefficien-
Kagitçibasi, C. (1982a). The changing value of cies of hierarchy. Cambridge, UK: Cambridge
children in Turkey (Publication No. 60-E). University Press.
Honolulu: East-West Center. Likert, R. (1961). New patterns of management.
Kagitçibasi, C. (1982b). Sex roles, value of children New York: McGraw-Hill.
and fertility in Turkey. In Ç. Kagitçibasi (Ed.), Likert, R. (1967). The human organization: Its man-
Sex roles, family and community in Turkey agement and value. New York: McGraw-Hill.
(pp. 151-180). Bloomington: Indiana University Macpherson, C. B. (1966). The real world of democ-
Press. racy. New York: Oxford University Press.
Kagitçibasi, C. (1982c). Old-age security value of Mao, Z. (1929). On correcting mistaken ideas in the
children and development. Journal of Cross- party. Selected works of Mao Zedong (Vol. 1,
Cultural Psychology, 13, 29–42. 1st ed.). Beijing: Foreign Language Press.
Kagitçibasi, C. (1996). Insan, aile, kültür [Person, Martin, A. (1984). The erosion of the Swedish model.
family, culture]. Istanbul: Remzi Kitabevi. In P. Gourevitch (Ed.), Unions and economic
Kaicheng, Y. (1998). Chinese employees’ perceptions crisis: Britain, West Germany and Sweden.
of distributive fairness. In A. M. Francesco & London: Allen and Unwin.
B. A. Gold (Eds.), International organizational Maslach, C., & Jackson, S. E. (1984). Burnout in
behavior (pp. 302-313). Englewood Cliffs, NJ: organizational settings. In S. Oskamp (Ed.),
Prentice Hall. Applied social psychology annual: Applications
Kanungo, R. N., & Aycan, Z. (1997, June 11–14). in organizational settings (Vol. 5, pp. 133–153).
Organizational culture and human resource Beverly Hills, CA: Sage.
practices from a cross-cultural perspective. Meindl, J. R., Yu, K. C., & Lu, J. (1987). Distributive
Paper presented at the annual meeting of the justice in the workplace: Preliminary data on
Canadian Psychological Association, Toronto. managerial preferences in the PRC. In J. B. Beak
Katz, D., & Kahn, R. L. (1952). Some recent findings (Ed.), Proceedings of the International Conference
in human relations research. In E. Swanson, on Personnel and Human Resources Management,
T. Newcomb, & E. Hartley (Eds.), Readings in Hong Kong.
social psychology. New York: Holt, Rinehart & Misumi, J., & Peterson, M. F. (1985). The perfor-
Winston. mance-maintenance (PM) theory of leadership:
18-House.qxd 3/19/2004 5:05 PM Page 601
19
UNCERTAINTY AVOIDANCE
MARY SULLY DE LUQUE
MANSOUR JAVIDAN
AUTHORS’ NOTE: The authors are indebted to Robert House for his painstaking review and detailed com-
ments on previous drafts of this chapter. The authors gratefully acknowledge the dedicated assistance provided
by Narda Quigley, Danielle Rizk, Tom Spies, Melanie Morris, Jana Heard, Chris Morgan, Robin Broyles, Arul
Rajendran, Chad Law, and James Louis at various times during the preparation of this chapter.
602
19-House.qxd 3/19/2004 5:34 PM Page 603
1950). Similarly, assessing the limits of logic, that relates to neuroticism, anxiety, stress,
Godel’s incompleteness theorem claims that a uncertainty avoidance, or whatever we want
variety of statements about natural numbers, to call it, that differentiates among modern
although true, may be not be provable and thus nations and affected IBM employees [his origi-
would be considered “undecidable” (Dawson, nal sample] as much as anyone else” (p. 156).
1999). Ultimately, even Einstein found the There may be distinctions among these various
deterministic laws of the universe to be cloaked terms; however, the terms appear to be of the
in mystery (Clark, 1984). same class of concepts (Clampitt & de Koch,
Many scholars consider it a fundamental 2002). The focus of this section is on identifing
need to reduce uncertainty in our lives, to commonalities and interrelationships between
explain our world, and establish predictability these concepts and uncertainty avoidance. We
(Berger & Calabrese, 1975). More recently, sev- will discuss some of the early origins of the
eral scholars suggest that certainty or true pre- notion of uncertainty and its influence on the
dictability are illusions, perpetuated by misguided construct of uncertainty avoidance.
Western thought, which advocates the control
of nature (Bradac, 2001). Influenced by Zen
Buddhism, this stream of thought suggests that Uncertainty Avoidance
we live in an earthly impermanence and we at the Individual Level
show naïveté in trying to go beyond our reality.
Tolerance for Ambiguity
Whether the world is fundamentally predictable
or essentially waiting to be predicted, however, Humans cognitively possess some varying
the notion that uncertainty exists remains. level of uncertainty avoidance. Frenkel-
Perhaps conceptualizing uncertainty and cer- Brunswik (1949), in one of the earliest known
tainty on a continuum, instead of as an either– citations of ambiguity tolerance, described a
or proposition, is more useful (Clampitt & de study that assessed respondents’ attitudes regard-
Koch, 2002). On the continuum, there are ing ethnic prejudice, in which ambiguity toler-
degrees of uncertainty, and the degree to which ance was noted as a personality variable related
people accept uncertainty depicts their tolerance to social orientation. Research in ambiguity
level. The effects of organizational and societal and uncertainty may have been hastened by
influences on a broader cultural continuum are psychologists’ interest in the study of authoritar-
the focus of the research that follows. ianism, which was prompted by the need to
understand the origins of World War II (Adorno,
Frenkel-Brunswick, Levinson, & Sanford, 1950).
CONCEPT OF UNCERTAINTY AVOIDANCE Defining the concept, Frenkel-Brunswik (1949)
posited, “ambiguity tolerance generalizes to
As defined by GLOBE, Uncertainty Avoidance the entire emotional and cognitive functioning
refers to the extent to which members of col- of the individual, characterizing cognitive style,
lectives seek orderliness, consistency, structure, belief and attitude systems, interpersonal and
formalized procedures, and laws to cover situa- social functioning, and problem solving behav-
tions in their daily lives. Known by many labels, ior” (Furnham & Ribchester, 1995, p. 180).
the study of uncertainty avoidance has been According to one of the more noted researchers
researched throughout many social science on the subject, being intolerant of ambiguity does
disciplines for decades. Originally used by not necessitate an increase of rules. More accu-
Cyert and March (1963) as an organizational rately, an increase in rules is a function of being
phenomenon, Hofstede’s (1980) publication, intolerant of ambiguity (Budner, 1962).
Culture’s Consequences, brought the term uncer- As mentioned earlier, although not fully syn-
tainty avoidance to prominence in explaining the onymous, intolerance of ambiguity (an individ-
behavior of societies. Discussing the relationship ual cognitive state) and uncertainty avoidance
between uncertainty avoidance and one of its (a behavioral phenomenon) are concepts that
many correlates, Hofstede (1980) states that are likely to be positively related (Furnham &
the concept represents “a national syndrome Ribchester, 1995). Determining ambiguity
19-House.qxd 3/19/2004 5:34 PM Page 604
intolerance has been the focus of many and happiness (Diener, Diener, & Diener,
researchers (Budner, 1962; Norton, 1975; 1995). Aggregated individual measures of
Rydell & Rosen, 1966) as well as cognitive uncertainty avoidance showed a significant rela-
tests (Bochner, 1965; Kreitler, Maguen, & tionship with subjective well-being (Arrindell
Kreitler, 1975). et al., 1997), with low uncertainty avoidance
related to high levels of subjective well-being.
Further, Steel and Ones (2002) found that
Anxiety, Stress, and Neuroticism
aggregated personality traits explain variance in
The notion of anxiety and stress is a common subjective well-being, with neuroticism–stress
theme in uncertainty and ambiguity research. In accounting for substantial variance in national
one compelling study, tolerance of ambiguity subjective well-being clearly above other factors
demonstrated a close association to anxieties such as national wealth.
(Furnham, 1994). Studies linking anxiety and
stress at the psychological level have been pro-
Feedback Seeking
lific, with many connecting stress to personality
dimensions of neurotic anxiety, extraversion, In organizational research, uncertainty is a
flexibility, and security (Kahn, Wolfe, Quinn, notable issue and is an implicit reason for dis-
Snoek, & Rosenthal, 1964). Differences in seminating feedback (Kluger & DeNisi, 1996).
the effects of uncertainty were reported from Uncertainty is a pivotal catalyst in feedback-
a psychological survey in five countries seeking behavior (Ashford & Cummings, 1983);
(Tannenbaum, Kavcic, Rosner, Vianello, & if individuals experience uncertainty, they will
Wieser, 1974), including such effects as depres- be motivated to enact an information search
sion, resentment, and low self-esteem. In another (Ashford, 1986; Morrison, 2002). Individuals
study, the relationship of the individual-level less tolerant of ambiguity have been shown to be
personality test, the Eysenck Personality more likely to seek feedback than those more
Questionnaire (Eysenck & Eysenck, 1975), was tolerant of ambiguity (Bennett, Herold, &
correlated with various medical-related indexes Ashford, 1990), especially in high uncertainty
across 37 countries (Lynn & Martin, 1995). The situations (Ashford & Cummings, 1985).
researchers found the national-level norms dif- Tolerance for ambiguity was shown to moderate
fered along the three personality dimensions,1 feedback-seeking behaviors and uncertainty
with an especially strong relationship noted regarding an individual’s role contingencies
between Hofstede’s (2001, p. 156) uncertainty (Ashford & Cummings, 1985). Research shows
avoidance and neuroticism versus emotional sta- that in order to manage uncertainty, individuals
bility. This supports analyses at the ecological will engage in strategies for seeking feedback
level that have shown uncertainty avoidance to from a variety of sources (superiors, peers, sub-
be strongly related with national neuroticism– ordinates), either through asking questions
anxiety and stress levels (Hofstede, 1980, 2001). (inquiry) or observing (monitoring).
The recent focus on positive psychology Over the past two decades research on
(Seligman 1999; Seligman & Csikszentmihalyi, feedback seeking behavior has focused on
2000; see also Aspinwall & Staudinger, 2003, three motives (Ashford, Blatt, & VandeWalle,
and Snyder & Lopez, 2002) has elicited in press): instrumental motive for achieving
research on a wide variety of constructs in the goals (Brown, Ganesan, & Challagalla, 2001;
social sciences. One such topic is the notion of Renn & Fedor, 2001), ego-based motive for ego
subjective well-being (SWB), commonly protection, and image-based motive for protect-
defined as people’s cognitive and affective ing and enhancing one’s image in the organiza-
assessment of their lives (Diener, 2000). tion (Bailey, Chen, & Dou, 1997; Heine et al.,
Subjective well-being has been suggested as a 2001). These motives may be affected by the
societal-level manifestation (Arrindell et al., context in which feedback seeking takes place;
1997; Hofstede, 1980, 2001; Schyns, 1998) in a context filled with uncertainty, the instru-
related to uncertainty avoidance (Arrindell mental motive for seeking feedback may domi-
et al., 1997), neuroticism (Steel & Ones, 2002), nate (Ashford et al., in press). The situations
19-House.qxd 3/19/2004 5:34 PM Page 605
that encourage the instrumental motive are Levels of analysis issues become particularly
exemplified by contextual uncertainty (Ashford salient when examining a construct studied
& Cummings, 1985), change (Ashford, 1988), so thoroughly at the individual level (Judge &
and novelty (Callister, Kramer, & Turban, 1999; Bono, 2001), such as the research on individual-
Morrison, 1993; Wanberg & Kammeyer- level uncertainty avoidance differences across
Mueller, 2000). In uncertain situations, people cultures (Basabe et al., 2002; Gudykunst &
seek feedback more frequently if feedback has Nishida, 2001). Urging the use of subjective
particularly high instrumental value. This is well-being (SWB) as a national-level index,
especially true in situations such as new job or Diener (2000) states that great progress has been
organization entry (Ashford, 1986; Ashford & made in deciphering the components, the under-
Cummings, 1985; Brett, Feldman, & Weingart, pinnings, and the cultural influences of SWB.
1990; Callister et al., 1999; Miller & Jablin, Methodological refinements are now used in
1991; Morrison, 1993). Further, information the study of individual-level constructs such as
seeking has been associated specifically with SWB, which allow its use as a national indica-
uncertainty avoidance across cultures (Earley, tor. Steel and Ones (2002) convincingly argue
1997; Sully de Luque & Sommer, 2000). that aggregated data in applied psychology
research commonly show that constructs of
personality can have robust relationships, with
Communication
stronger correlations found at the national level
In communication, uncertainty has been a than at the individual level. A similar argument
focus of study. The phenomenon of uncertainty has been advanced for mood and emotion con-
has been associated with uncertainty reduction tagion in work teams (Kelly & Barsade, 2001).
theory (Berger & Bradac, 1982; Berger & The effects may be greater at the group level
Calabrese, 1975), problematic integration than at the individual level.
theory (Babrow, 1992) and uncertainty manage-
ment theory (Babrow, Hines, & Kasch, 2000;
Uncertainty Avoidance as an
Ford, Babrow, & Stohl, 1996). Additionally,
Organizational Phenomenon
Festinger (1957) alludes to the concept of
uncertainty in his work on cognitive dissonance. Cyert and March (1963, p. 116) originally
Uncertainty reduction has been found to be a coined the term uncertainty avoidance, to refer
common element in interpersonal relationships, to an organizational phenomenon, one of the
and has been found to vary (Gudykunst, 1983; four major relational concepts in their behav-
Gudykunst, Nishida, & Schmidt, 1989) across ioral theory of the firm. Addressing Gordon’s
cultures (Hall, 1976; Nakane, 1974). Uncertainty (1948) observation of a wide discrepancy
has been described as a cognitive experience, between organizations’ reaction to uncertainty
with predictive uncertainty being related to and theoretical predictions in economic theory,
the estimation of others’ feelings, attitudes, and Cyert and March (1963) posited that organiza-
behaviors (Berger & Calabrese, 1975). Anxiety tions adopt procedures that minimize the need
is the emotional equivalent of uncertainty, and it for predicting uncertain events in the future
is derived from feeling worried, uncomfortable, (pp. 13, 102).
or apprehensive (Gudykunst & Nishida, 2001). Uncertainty avoidance may occur within
Avoidance is a behavioral effect of anxiety various levels of the organization. To illustrate,
(Stephan & Stephan, 1985), and there is a Beatty and Gordon (1988) have used the term
reciprocal positive relationship suggested bet- uncertainty avoidance to describe a human barrier
ween these concepts (Demerath, 1993; Turner, to the successful implementation of Computer
1988). Anxiety/uncertainty management theory Aided Design/Computer Aided Manufacturing
proposes that communication effectiveness is (CAD/CAM) systems in manufacturing firms.
influenced by individuals’ capacity to manage In this context, uncertainty avoidance refers to
uncertainty and anxiety (Gudykunst, 1988, an individual phenomenon at various levels in
1995), and Hofstede (2001) recognizes its the organization. At senior levels of management
relationship with uncertainty avoidance. in the organization, especially where technical
19-House.qxd 3/19/2004 5:34 PM Page 606
expertise on proposed implementation of new in their relationships with the market. Internal
projects is lacking, there may be a reluctance to budgeting and planning procedures (in contrast to
take on new projects with uncertain outcomes. sales plans and forecasts) act as contracts within
At the same time, in the middle management the organization among the various departments
levels, uncertainties relating to the mastery of so that the uncertainty of the immediate envi-
new tools and ways of doing work (on which ronment of a division of the organization can be
managers may be evaluated) constitutes a barrier controlled (McNally, 1980).
to new projects such as CAD/CAM.
Planning
Feedback
The relationship between planning and orga-
Short-term feedback is one proactive method nizational success, especially in small business
that organizations use to avoid the uncertainty development, has been related to uncertainty
associated with outcomes (Ilgen, Fisher, & Taylor, management (Nakata & Sivakumar, 1996). The
1979; Kluger & DeNisi, 1996). Acceptance and literature on small firm success assumes that to
enforcement of standardized decision rules also reduce uncertainty, planning is a central factor
serve the same purpose. Internal biases in orga- (Ryans, 1997), although the support for this
nizations (such as sales departments’ estimates has been equivocal. Although some studies have
of sales) serve to increase pressure on the firm shown no relationship between planning and suc-
to develop decision heuristics, or “rules of thumb.” cess (Lumpkin, Shrader, & Hills, 1998), others
Organizational rules are suggested as a way of have shown that thorough planning, in the initial
maintaining a stable environment given that the stages of creating a small-scale business, is asso-
behavior of organizational members may be ciated with organizational success (Ackelsberg &
unpredictable (Perrow, 1972). Arlow, 1985; Bracket, Keats, & Pearson, 1988).
Short-term reactions to short-term feedback, Planning is thought to control uncertainty
in Cyert and March’s behavioral theory of the within the organization, yet the connection
firm, help to explain how organizations avoid between planning and return on investment can
the problem of correctly predicting future be tenuous. In a study assessing formal plan-
events. For instance, production decisions are ning in small business environments, relation-
usually dictated by daily and weekly feedback ships were revealed between formal planning
from inventory and sales departments rather and the growth rate of sales, although no rela-
than by long-term sales forecasts. In a similar tionships were found on return on equity or
vein, in the behavioral theory of the firm return on assets (Lyle, Baird, Orns, & Kuratko,
(Cyert & March, 1963), the strategy of negoti- 1995). Although the focus and findings of this
ated environments helps to eliminate uncertain- research have been multifaceted, a primary
ties. This is in contrast to the classical models of issue in internal planning is the management of
oligopoly in which firms make predictions uncertainty.
about the behavior of the environment, such as
competition, consumer behavior, and supplier
Innovation
dynamics. Control over the environment is a
more proactive way of dealing with uncer- The uncertainty created by the introduction of
tainty than is prediction of the environment. innovative activities has been the implicit focus
The establishment of customary practices in an of research in organizations. In conventional
industry acts as a control on competitive behav- organizations, innovation is frequently difficult
ior. Examples of such practices include product to develop because employees usually want to
pricing, codes of ethics, and budgeting. Product hold to established procedures on which past
pricing on the basis of industry-established con- organizational legitimacy has been based
ventions ensures predictability and stabilization (Dimaggio & Powell, 1983). Indeed, employees
of consumer demand. Fair business practice often rely on the traditional hierarchical organi-
codes in an industry regulate the behavior of zation if they are presented with information that
organizations in competing with one another and challenges their norms and beliefs (Williamson,
19-House.qxd 3/19/2004 5:34 PM Page 607
1975), even if this approach is inefficient in a markets and plans. Technologies developed to
changing environment (Venkataraman, Shane, handle uncertainty may include medical devices,
McGrath, & MacMillan, 1993). Innovation tends security systems, and even military armament
to introduce unanticipated changes for the or capital goods industries, such as machinery
employees and cause uncertainty that may lead to and transport equipment. Laws are enacted in
resistance to innovation (Shane, Venkataraman, & societies to cope with uncertainty by providing
MacMillan, 1995; Van de Ven, 1986). both informal and formal rules to guide behavior.
The GLOBE research has furthered insight The legal system in a society sets forth guidelines
into the topic of uncertainty avoidance. for managing the consequences when laws
Specifically, GLOBE developed scales to eval- are not followed. Uncertainties that exist outside
uate uncertainty avoidance at the organizational technological and legal resolutions are often
level of analysis, using organizational referents coped with through religion, often in the form of
in its measures. In addition, GLOBE research rites and rituals. Hofstede (2001) argues rituals
specifically examined the association between serve the function of allowing societal members
organizational and societal uncertainty avoidance, to carry on with their lives while confronting
extending the understanding of the multilevel intolerable levels of anxiety. Often revealing a
complexity of the construct. philosophical understanding of the nature of exis-
tence, religion imparts a sense of acceptance of
uncertainties.
Uncertainty Avoidance as a
Technology, rules, policies, and rituals are all
Dimension of National Culture
means used by organizations to deal with uncer-
Initially interpreting the concept of uncer- tainty. Organizations confront uncertainty in their
tainty avoidance as explaining organizational environments, through the behavior of both
occurrences, Hofstede (1980)2 used it to describe employees and stakeholders, as well as the exter-
how people in societies accept uncertainty in nal agencies with which the organizations inter-
everyday life. Through this reconceptualization, act (Hofstede, 2001). Rules and regulations serve
uncertainty avoidance shifted from being viewed to bring about more predictable behavior.
as a behavioral organization-level variable in Hofstede (2001) notes that there are both good
Cyert and March’s (1963) theory to being and bad rules: Good rules lead to desired out-
thought of as a societal value orientation. comes and concur with peoples’ values; bad rules
People in societies create coping mechanisms are formed when there are value differences
to handle the anxiety produced by excessive between those who generate the rules and those
uncertainty (Hofstede, 2001). All individuals who follow them. Organizational rituals include
are affected by anxiety in different ways, depend- events such as business meetings, which have
ing on their psychological, physical, spiritual both acceptable and proscribed behaviors, as well
and philosophical dispositions. Hofstede (2001) as management training programs, which are
notes that the primary mechanisms through often initiation rites for potential organizational
which societies cope with uncertainty are tech- leaders. Many additional uncertainty-avoiding
nology, law, and religion. He convincingly rituals may be observed in organizations,
presents the roles of the three mechanisms by Hofstede (2001) notes, including “the writing
stating, “Technology has helped us to defend and filing of memos and reports, accounting,
ourselves against uncertainties caused by nature; planning and control systems, computer simula-
law, to defend against uncertainties in the behav- tions, and the nomination of experts as persons
ior of others; religion, to accept the uncertainties who are beyond uncertainty” (p. 148). At the
we cannot defend ourselves against” (Hofstede, national cultural level, Hofstede (2001) suggests
2001, p. 146). The simplicity of these mecha- that traditionalism and ethnocentrism, intoler-
nisms is translated into a complexity of practices ance of differing opinions, inflexibility and dog-
in societies and organizations. matism, tendencies toward racism, and prejudice
In societies, uncertainty-reducing technolo- are associated with the norm for intolerance for
gies may take the form of a service such as prod- ambiguity, which he measures through the uncer-
uct warranties, insurance policies or investment tainty avoidance index (p. 146).
19-House.qxd 3/19/2004 5:34 PM Page 608
originally proposed to measure organization-level their tolerance for uncertainty. The UAI may also
or individual-level constructs (Hofstede, 1980, possibly reflect labor market opportunities or
2001). As a composite of three items, the uncer- collective orientation of the society (see Chapter
tainty avoidance scale showed great variability 16 by Gelfand, Bhawuk, Nishii, & Bechtold).
across countries. From this, the index for uncer- Even with the employment stability question
tainty avoidance (UAI) was calculated through limitations, it is possible for two countries to
responses to these three items (Hofstede, 1980, obtain the same UAI scores through very differ-
2001), labeled as stress, rule orientation, and ent profiles. Rule orientation, for instance, may
employment stability, respectively. Stemming be preferred for reasons other than coping with
from the conceptual link of mean level of anxi- uncertainty. In a country whose dominant
ety in a country, Hofstede accounted for the religion ensures a philosophical approach to
correlation among these three scores. According managing uncertainty, respondents may have
to Hofstede, high scores on the UAI represent very low stress levels even though they are rule
higher levels of uncertainty avoidance. oriented. Again, in certain societies a mean
The first question in the UAI scale asked higher rate of national or occupational employ-
respondents to rate their agreement with the ment security and longer average tenure on
statement that rules should never be broken, even jobs is likely to be associated with making
if it is in the best interests of the company to do individuals feel fairly secure and, consequently,
so. Higher rule orientation (agreement with the less stressed in their jobs, although this state of
statement) aggregated at the societal level raised affairs is by no means indicative of how they
the UAI. The second item asked respondents to would react if their jobs were not secure. These
estimate the amount of time they planned to stay profiles are compensatory so that, ultimately,
in the employment of their company. The longer their UAI scores will be similar.
they planned to stay, the higher their desire for The relationship between anxiety and rule ori-
employment stability, and the higher their pre- entation was proposed initially by Van Gunsteren
sumed uncertainty avoidance. The percentage of (1976). Through this, it is suggested that when
respondents planning to spend more than 5 years anxiety levels are high, people experience more
in their present company contributed directly to stress. Through the formulation of rules and
the societal UAI. The third item asked respon- structure, stress and anxiety are often managed,
dents to indicate how often they felt nervous or enhancing the sense of security. Hofstede (2001)
tense at work. The higher this frequency was notes that the research suggesting that anxiety is
reported at the societal level, the higher the UAI. a construct with societal-level implications has
Revisiting the construct two decades later, been rigorous and compelling, as outlined earlier
Hofstede (2001, p. 148) distinguished uncertainty in this chapter. In assessing societal shifts in
avoidance from risk avoidance. stress levels, Hofstede (2001) states that shifts
Some features of Hofstede’s UAI measure toward higher stress have occurred in both high
are noteworthy. The item in his measure that tar- and low UAI cultures (p. 182). However, these
geted employment longevity may be a function significant worldwide shifts did not occur for
of not only the respondent’s desire for employ- either high or low UAI cultures on the other ques-
ment stability, but also of the amount of time tions (rule orientation and employment stability).
already spent in the company and forecasts of Thus Hofstede (2001) acknowledges that using
employment stability. Thus, a person who has the UAI composite measure as an index for lon-
spent 10 years in the company at the time of gitudinal culture change is not advisable. This
survey who answers that she plans to move admission, together with importance of the
on within the next couple of years would serve “stress” component shift and the remarkable
to lower the UAI; another employee who has association between stress–anxiety measures and
been in the company for less than a year who UAI, creates the impression that UAI may be a
responds that he plans to continue in the com- more likely measure of stress than an overarching
pany for another 10 years would raise the UAI. measure of uncertainty avoidance.
Yet, according to Hofstede, they are both opera- In spite of explicit acknowledgment of
tionally in the same category with regard to the constraints under which the uncertainty
19-House.qxd 3/19/2004 5:34 PM Page 610
avoidance items came to constitute UAI validity of the measures. The fact that Hofstede
(Hofstede, 1984, 2001), scholars have made few suggests that the employment stability question
attempts to probe deeper into the construct as a may be unique to his sample is an example of
cultural dimension until the GLOBE study and this. Recognizing the limitations of his measures,
the subsequent revision of the UAI measure in Hofstede (1984, 2001) acknowledges the restric-
the values survey module (VSM 94; Hofstede, ted scope of his uncertainty avoidance measure:
2001, p. 186).4 However, several scholars have the choice of items for UAI was restricted to the
noted the limitations of the UAI in measuring items available in archived data, from surveys
societal culture (for further discussion, see used for a different purpose. Thus, the authors
Fernandez, Carlson, Stepina, & Nicholson, here suggest the items have questionable face
1997; Hansen, 2001; Tayeb, 1994). Arguing in validity. Further, findings related to UAI may
favor of the contribution that an ethnographic reflect the competitive hypothesis that UAI is
approach can make to the clarification of con- more a measure of stress than it is of Uncertainty
structs in cross-cultural organizational research, Avoidance values or practices.
d’Iribarne (1997) highlights two major limita- As evidenced by the confluence of literature
tions in Hofstede’s measures: the issue of neu- from a variety of disciplines, uncertainty has an
trality of items and the question of interpretation. influence in societies. In Project GLOBE,
Neutrality refers to the fact that two cultures Uncertainty Avoidance was defined in terms of
may have the same level of the attribute—in this a tendency toward orderliness and consistency,
case the UAI score—but may manifest different structured lifestyles, clear specification of social
ways of coping with uncertainty within the expectations, and rules and laws to regulate
framework of their cultures. The question of uncertain situations. In addition, built into the
interpretation is a concern voiced by many GLOBE design was a validation study based on
cross-cultural researchers. For instance, d’Iribarne an independent unobtrusive measure of societal-
(1997) explains how the French take the letter of level practices, and independently collected
the rule less seriously than do Americans. This data reported by Inglehart (1997) to measure
would lead to differences in responses to the societal-level values (see Chapter 9 by Gupta,
item in Hofstede’s survey that deals with rule Sully de Luque, & House).
breaking (rule orientation). As a result, France
ranks high in Hofstede’s UAI, whereas the Uncertainty Avoidance
United States ranks low, in contrast to what Correlates in Large-Scale Studies
d’Iribarne finds in the two cultures. The French
and Meta-Analysis of Hofstede
presumably express strong agreement with the
statement that rules should not be broken— The impact of the Hofstede (1980) study on
except that the qualifiers to the statement are the concept of uncertainty avoidance is substan-
taken for granted. Americans, on the other hand, tial and the contribution is notable. Business
are more likely not to assume qualifying condi- textbooks often include a discussion of the
tions unless explicitly stated in the item. dimensions of culture put forth in his seminal
More recently, Hofstede (2001) does allow work. Hofstede (2001) notes that in 140 further
for estimating reliability estimates for his studies comparing from 5 to 39 countries, one or
Uncertainty Avoidance scale. He then justifies more of the four (later five) dimensions were
the validity of the construct by the strong corre- found to be correlated with his original rankings.
lations he finds with other measures, interpret- Assessing the replications and applications
ing these in ways consistent with the construct of the Hofstede study, Søndergaard (1994)
definition. Many correlations can be interpreted conducted an analysis of published and unpub-
in two or more ways by using culturally specific lished work spanning the years 1980–1993. He
manifestations as examples. In the GLOBE found four distinctive means through which
study, we argue that constructs first need to be Hofstede’s work was being applied: citations,
of good quality, supported by rigorous method- reviews, empirical replications, and as a para-
ology. As a result of rigorous, theory-driven digm.5 At the time of his analysis, 61 replications
research, we may have more assurance of the had been conducted.6 Among studies that found
19-House.qxd 3/19/2004 5:34 PM Page 611
full confirmation with all of Hofstede’s (1980) country rankings of the Hofstede dimensions
dimensions (Hoppe, 1990; Punnett & Withaney, (Ashkanini, 1984; Chew & Putti, 1993; Chow,
1988; Shackleton & Ali, 1990), the Hoppe study Shields, & Chan, 1991; Dunphy & Shi, 1986;
was the first broad-based attempt to replicate the Fidalgo, 1993; Forss, 1989; Huo & Randall, 1991;
Hofstede results. This study of 19 countries7, Maldonado, 1983; Pooyam, 1984; Punnett &
using an elite sample from the Salzburg Seminar Whitaney, 1988; Westwood & Everett, 1987;
Alumni Study (SSAS), revealed a high degree of Yeh, 1988). Although many studies confirmed
agreement between the UAI scores of the SSAS Hofstede’s rankings, confirmation was not
respondents and the scores of Hofstede UAI found for uncertainty avoidance in several other
respondents (r = .64, p < .01; rho = .63, p < .01).8 attempted replications (Fidalgo, 1993; Pooyan,
However, this comparison of the two studies 1984; Punnett & Withaney, 1988). Often, politi-
also revealed that Hoppe’s power distance index cal and environmental factors were cited as rea-
(PDI) was more strongly related to the Hofstede sons for these discrepancies. Of the scores of
UAI than to the Hofstede PDI (r = .71, p < .01; studies conducted based on the Hofstede find-
rho = .81, p < .001). It is important to note that ings, Lowe (1996) was one of few researchers to
Hofstede’s PDI and UAI were highly correlated use a population of respondents similar to the
(r = .78, p < .001; rho = .75, p < .001), while respondents in the Hofstede study. Interestingly,
Hoppe’s PDI and UAI were also highly related Lowe (1996) found similar ranking of all of the
(r = .65, p < .01; rho = .81, p < .001). As a con- dimensions except for uncertainty avoidance.
sequence of this multicollinearity, comparison The most frequent use of the Hofstede (1980)
of UAI and PDI is problematic. Speculating on dimensions is as a paradigm for describing
these results, Hofstede (2001) notes that the country differences (Søndergaard, 1994). In other
higher educational level of the Salzburg alumni words, the original four Hofstede dimensions are
“implied less rule orientation and less employ- used to categorize and rationalize (often in a
ment stability” (p. 154), which led to discrepant speculative way) the influence of culture on the
findings. Indeed, the Hoppe (1990) study played dependent variable of interest to the researchers.
an important role in Hofstede’s second version Decades after the original Hofstede study, these
of his value survey module (VSM 94). dimensions are quite often used as paradigmatic
Helmreich and Merritt (1998) conducted a reference points for research across disciplines.
large-scale study of commercial airline pilots Many scholars have cautioned researchers about
that included the original three UAI questions taking these dimensions, and the countries’ place-
used by Hofstede (1980). During the years 1993 ment on these dimensions, as taken-for-granted
and 1997, the researchers surveyed 15,000 com- assumptions (Baskerville, 2003; Tayeb, 1994).
mercial airline pilots across 23 companies from
36 countries. They reported a significant rela-
Correlates of Uncertainty Avoidance
tionship between the commercial airlines pilot
and Individual-Level Phenomena
respondents and the Hofstede (1980) respondents
from IBM (r = .49, p < .05; rho = .47, p < .05). In a study examining country-level medical
Interestingly, the airline pilots’ PDI scores also statistics, Lynn and Hampson (1975) found the
correlated with the IBM UAI scores (r = .46, notion of neuroticism–anxiety as the primary
p < .05; rho = .50, p < .05). Whereas the IBM factor (explaining 57% of the variance) of a
respondents shared a mutual employer but came factor analysis conducted for 18 developed
from different professions, the commercial airline countries (p. 125). Collected in 1960, their data
pilots had a common profession but not a com- show a strong correlation of neuroticism–
mon employer. Hofstede (2001) notes that the anxiety factor scores with Hofstede’s UAI
three uncertainty avoidance questions carried dif- and related stress scores (rho = 73, p < .01), and
ferent implications for the pilot respondents, who this relationship was quite compelling for
shared the same profession, than they did for the the neuroticism–anxiety construct. Hofstede’s
IBM respondents, who had a common employer. UAI also rank-correlated at .71 (p < .05) with
The Søndergaard (1994) review included Millendorfer’s stress rankings across 12 countries
several studies reflecting partial replications of (Millendorfer, 1976). In a later study assessing
19-House.qxd 3/19/2004 5:34 PM Page 612
national normative scores on the EPQ personality management (interpreted as reducing risk for
inventory, the personality dimension “neuroti- the individual decision maker) and a tendency
cism versus emotional stability” was correlated toward McGregor’s (1960) theory X, involving
significantly (rho = .44, p < .05) with Hofstede’s less optimism about people’s capabilities and
UAI across the 25 countries common for each ambitions.
sample (Lynne & Martin, 1995). As such, Assessing the effect of uncertainty and
greater tendency to avoid uncertainty, as mea- anxiety on perceived effectiveness of communi-
sured by the Hofstede UAI scale, appears cation across relationships and cultures, stranger
strongly related to higher levels of neuroticism and close-friend relationships have been speci-
and stress. Supporting this, Arrindell and fically examined in the U.S. and Japan9
colleagues (1997) found that uncertainty avoid- (Gudykunst & Nishida, 2001) at the individual
ance is more predictive of happiness (inverse level.10 The results show a moderate connection
of stress–neuroticism) than wealth (t = −1.38, between anxiety and uncertainty with perceived
p < .10, one-tailed). communication effectiveness, across both
Hofstede examined McClelland’s (1961) arti- close-friend and stranger relationships. For
cle, which used data obtained in 1925, to assess perceived communication effectiveness in the
need for achievement (nAch), need for affiliation United States, anxiety had a substantially
(nAff), and need for power (nP). He found across greater effect than uncertainty (reverse scored)
22 countries a strong negative correlation between (r = −.59, p < .001) for stranger interaction,
uncertainty avoidance and nAch (rho = −.64, whereas uncertainty had a greater effect than
p < .01), suggesting that societies low in tolerance anxiety (r = .73, p < .001) for close-friend inter-
for uncertainty tended to be less achievement action. In Japan, anxiety had a greater effect
oriented. This relationship is consistent with the than uncertainty on perceived communication
idea that willingness to take risks may be a pre- effectiveness in both stranger (r = −.54, p < .001)
requisite to extraordinary levels of achievement. and friend (r = −.56, p < .001) interactions.
Thus, willingness to take risks may also be nega- Among the characteristics of uncertainty
tively related to one’s level of anxiety. avoiding societies in Hofstede’s analysis are
In another study, a survey of consumption higher anxiety levels in the population, higher job
patterns and attitudes was conducted measuring stress, more resistance to change, higher average
younger and older generations (Reader’s Digest, age in higher-level jobs, fear of failure, less risk
1970). Hofstede (1980, p. 128) found a strong taking, lower ambition for individual advance-
negative correlation (rho = −.77, p < .01) across ment, preference for clear requirements and
15 countries between favorable attitudes toward instructions, strong values against breaking rules,
the younger generation and the UAI. Consis- lower tolerance for ambiguity in perceiving
tent with this result, Hofstede also found that others, and lower readiness to compromise (for
ages of leaders in 11 European countries corre- details see Hofstede, 1984, pp. 132–133). All of
lated strongly with the UAI (rho = .75, p < .05) these associations between the UAI countries’
(de Bettignies & Evans, 1977). scores, with the exception of higher average age
Hofstede also examined Least Preferred in higher-level jobs, are consistent with higher
Coworkers (LPC) (Fiedler, 1967) scores of levels of stress. Given that one of Hofstede’s
managers from 16 countries attending a course at items concerned level of stress, these findings are
International Management Development Institute consistent with the notion that the UAI index
in Switzerland IMEDE. He found a correlation of may be a more specific measure of stress rather
−.44 (p < .10) between LPC and UAI, consistent than a general measure of uncertainty avoidance.
with the notion that high LPC scores could be a This competitive hypothesis remains to be tested.
reflection of greater tolerance for uncertainty.
Analyzing data from another study of managers Correlates of Uncertainty Avoidance
in 19 countries (Haire, Ghiselli, & Porter, 1966),
and Organizational-Level Phenomena
he found that in countries characterized by high
uncertainty avoidance there exists a stronger The uncertainty avoidance norm has been
preference for group decisions and consultative examined conceptually and empirically in
19-House.qxd 3/19/2004 5:34 PM Page 613
recruitment, or internships and apprenticeships These results suggest that the less the tolerance
may be considered credible low-risk practices for uncertainty, the lower the number of verifi-
preferred in high uncertainty avoidance cul- cation methods used, and the fewer the extent of
tures (Jeanquart-Barone & Peluchette, 1999). verification. They also suggest that reported
Thus, choice of selection techniques such as uncertainty avoidance cultures tend to rely on
assessment centers’ structured interviews that interpersonal methods of investigation rather
are designed to reduce uncertainty (Dessler, than on formal methods.
1993) may be favored in societies high on
uncertainty avoidance. In a study comparing Accounting Practices. Studies in the field of
U.S. and German firms, Jeanquart-Barone and accounting have revealed that accounting
Peluchette (1999) found that German firms systems differ among the various parts of the
reported a greater use of internal recruiting than world. Gray (1988) has offered a theoretical
did U.S. firms for blue-collar (t = 1.77, p < .05) model of cultural influences on accounting
and white-collar (t = 2.46, p < .01) employees. practices. From a review of the literature, Gray
For both blue- and white-collar employees, (1988) identified four value dimensions in
German firms were also more likely than U.S. accounting: professionalism versus statutory
firms to hire their apprentices and interns control, uniformity versus flexibility, conser-
(t = 2.32, p < .05). At the white-collar level, vatism versus optimism, and secrecy versus
German firms used internships and apprentice- transparency. From this theoretical platform,
ships significantly more than did United States Gray advanced four propositions involving
firms (t = 2.54, p < .01), and German firms Hofstede’s dimensions of national culture. With
were found to invest a higher percentage of regard to uncertainty avoidance, Gray proposed
their total budget on training than did firms that societies high in this dimension will have
in the United States (t = 2.36, p < .01). Finally, greater preference for statutory controls as
the authors reported that for white-collar opposed to individual professionalism, have
recruitment, the German firms used more struc- greater uniformity in procedures across compa-
tured interviews (X2 = 11.57, p < .001) and nies, be more conservative in their measurement
assessment centers (X2 = 3.86, p < .05) than did of accounting and finance variables, and have
U.S. firms. more stringent confidentiality and disclosure
Ryan, McFarland, Baron, and Page (1999) policies.
examined the influence of national culture on Gray’s (1988) theoretical postulations were
personnel selection practices using responses operationalized and tested by Salter and
from 959 organizations, mostly from developed Niswander (1995). They collected data from 29
countries. The authors found evidence for an countries, involving all major stock markets
association between the societal norm of uncer- and comprising more than 84% of the 1989
tainty avoidance and a number of organizational global domestic product and more than 97%
selection practices. The selection practices con- of the 1990 stock market capitalization. Their
cerned number of verification methods used, sample contained a variety of cultures in terms
frequency of verification, and number of inter- of language, geographical location, colonial
views. For the cultural dimension scores, they antecedents, and economic development. A
adopted the indices reported by Hofstede (1984; total of nine dependent variables representing
cited in Ryan et al., 1999, p. 365). Hofstede’s Gray’s accounting value dimensions were
UAI correlated negatively with the number of developed for their ordinary least squares
verification methods used (r = −.72, p < .01) as regression analyses.
well as with extent of verification (frequency; Consistent with Gray’s (1988) first proposi-
r = −.61, p < .01), and positively with the tion, the investigators found that uncertainty
number of interviews (r = .57, p < .05). They avoidance (as measured by Hofstede’s UAI)
found significant corresponding correlations of was significantly negatively related to both
−.80, −.71, and .37 (N = 959) with number of practice and structure of subdivisions of profes-
verification methods used, frequency of verifi- sionalism. On the other hand, uncertainty avoid-
cation, and number of interviews, respectively. ing countries tend to have more prescriptive
19-House.qxd 3/19/2004 5:34 PM Page 615
legal and statutory requirements regarding degree of legal uniformity, pessimism (one of
accounting procedures.11 With regard to the sec- their measures of conservatism), and secrecy.
ond proposition, one of the measures assessing Higher marginal tax rates would achieve the
uniformity of procedures, that is “the extent to opposite effect on professionalism, legal unifor-
which a country’s legal system is code-law mity, conservatism, and secrecy. They found
based” (legal uniformity), was significantly pos- support for the hypothesized relationships
itively related to Hofstede’s UAI. The other between market capitalization and profession-
measure of uniformity, assessing the actual alism, legal uniformity, conservatism, and
extent to which uniformity is achieved in prac- secrecy. With regard to marginal tax rates, they
tice (effective uniformity), revealed a signifi- found support for the hypothesized relationships
cant negative relationship to Hofstede’s UAI. to professionalism and conservatism. These
The first measure of uniformity was also signif- results, which extend the original hypotheses,
icantly negatively related to Hofstede’s mas- highlight the influences of factors other than
culinity scale and the second was significantly cultural in the accounting value dimensions the-
positively related. In reconciling these results, orized by Gray (1988). Overall, Gray’s hypothe-
the researchers argue that the second uniformity sized association between accounting values and
measure is vulnerable to the comprehensiveness the uncertainty avoidance construct of culture
of rules dictating the legal accounting proce- are supported, with the exception of one mea-
dures, a characteristic of many societies high in sure of uniformity.
uncertainty avoidance. Therefore, in situations Following up on Salter and Niswander’s
where no rules exist, decisions are left to the (1995) study, Zarzeski (1996) modeled the influ-
individual accountant, giving the impression of ence of several organizational variables (sales,
flexibility when in fact it is simply a failure of debt ratio, and size of the firm) followed by
the high masculinity system. Thus, the negative Hofstede’s measures of uncertainty avoidance,
relationship must be seen as ambiguous rather individualism, masculinity, and power distance
than unequivocally valid. on accounting disclosure practices, using data
With respect to the third proposition of Gray from published sources in seven industrialized
(1988), Salter and Niswander (1995) had two countries. In a regression model that accounted
measures of conservatism. The first measure, for 48% of the variance in disclosure, Hofstede’s
assessing the “aggressive conservatism designed UAI yielded a negative beta coefficient of −.18,
to reduce income whenever possible,” failed (p < .001), coming in fourth after debt ratio, firm
to reveal any significant relationship. The sec- size, and individualism. This negative relation-
ond (labeled pessimism), addressing a generally ship between uncertainty avoidance and disclo-
conservative approach, revealed a significant sure lends further support for Gray’s hypothesis
positive relationship to the UAI, suggesting that regarding secrecy.
more uncertainty avoidance societies tend to be One must remember, however, that
more conservative. Finally, the fourth proposi- accounting firms, like other firms, also have
tion was supported by both of their measures global operations. Soeters and Schreuder
of secrecy—informative annual reports and (1988) found significant differences based on
extent of disclosure of accounting practices cultural characteristics between three Dutch
(both reverse coded)—with a relatively strong accounting firms and three international Big
relationship. Eight accounting firms operating in the
In an interesting extension of Gray’s (1988) Netherlands. These differences were particu-
theory, Salter and Niswander (1994) posited the larly noticeable in uncertainty avoidance. The
influence of two economic, as opposed to cul- authors, however, are careful to point out that
tural, factors: The state of development of a these differences are probably due to a process
nation’s capital markets and the country’s mar- of self-selection rather than socialization of
ginal tax rate. They hypothesized that the more employees resulting in any internalization of
developed a country’s capital market, the higher other societal cultural values by the Dutch.
would be the extent of professionalism and Whatever the case, the implication is that
effective uniformity, and the lower would be the data collection from accounting firms should
19-House.qxd 3/19/2004 5:34 PM Page 616
the United States, the Netherlands, Hong Kong, other form of interference by the organization.
and Taiwan. Their findings are interesting: These champions have been found to manage
Whereas positive outcomes dampened percep- this through two methods. In one approach, the
tions of risk to a greater degree among the two champion creates a greater level of autonomy
Western samples than it did for Taiwan (Hong for the innovator, who is then free to develop
Kong being intermediate), negative outcomes and suggest creative solutions (Burgelman,
were more complex in their effects. The 1983; Howell & Higgins, 1990; Schon, 1963).
Western samples were more influenced by the In another approach, the autonomy champion
probability of a loss; Hong Kong and Taiwan, encourages the innovator to create ideas that
on the other hand, were influenced more by the work well within the strategy and culture of the
magnitude of the loss. Bontempo and colleagues organization. The champion develops an envi-
(1997) found that the direction and interpreta- ronment compatible with the innovator’s norms,
tion of the cultural differences in their results procedure, and rules.
were consistent with Hofstede’s UAI. Shane (1994) found that uncertainty avoid-
ance was significantly related to a proclivity for
Innovation Championing Strategies. Whereas champions who go against rules, procedures,
uncertainty avoidance can be theoretically associ- and norms of the organization (rho = − .35,
ated with the societal level variable of innovation p < .05).15 This implies that in higher uncer-
per se, the championing of innovation is an orga- tainty-avoiding societies, the preferred cham-
nization-level variable that may also be theoreti- pion strategies are those that operate within
cally linked with national culture. Shane (1994) the procedures, rules, and norms of the organi-
investigated the association between national zation. In similar research, Shane and col-
culture, using Hofstede’s measures, and national leagues (1995) examined the influence of
preference for innovation-championing strategies. national culture on innovation championing
Shane conducted a study of 24 national offices strategy preferences. This study of 1,228 indi-
with a financial services company, surveying 937 viduals in 30 countries who were members
managers engaged in accounting and consulting, of four organizations16 found that Hofstede’s
to assess this phenomenon. UAI scales were significantly negatively
The study of champions in the promotion of related to autonomy from organizational norms
innovations in organizations has gained momen- and procedures when controlling for demo-
tum since the sixties (Schon, 1963). Broadly, graphic difference (t = − 3.96, p < .001), organi-
champions are defined as people who go beyond zational differences (t = − 3.23, p < .001), and
formally assigned roles in organizations to pro- economic differences (t = − 3.28, p < .001).
mote a product or venture (Chakrabarti, 1974; Shane and colleagues (1995) state that “the
Shane, 1994). Of six innovation-championing more uncertainty accepting a society is, the
strategies addressed by Shane (1994), persuasion more people in it prefer champions to over-
and autonomy are relevant to the present dis- come organizational inertia to innovation by
cussion. Through the strategy of persuasion, the violating organizational norms, rules and pro-
champion rationally presents the value of inno- cedures” (p. 945). In societies reporting high
vation through quantitative approaches. Thus, UAI scores, employees in organizations pre-
the champion convinces managers of the need ferred innovation champions who followed
for change through such notions as cost–benefit budgetary processes, justifying their decision
analysis, budgets, presentation models, and on the basis of financial rationale.
financial projections (Howell & Higgins, 1991). This is not to say that innovators in high
Societies with higher levels of uncertainty avoid- uncertainty avoidance cultures would not
ance are thought to prefer champions who achieve as much as in low uncertainty avoid-
acquire backing for new ideas through formal ance cultures. As Hofstede (1984) points out,
mechanisms such as plans and forecasts. there is some evidence (e.g., Inkson, Schwitter,
Through the strategy of autonomy, the Pheysey, & Hickson, 1970; Kohn, 1971) that
champion “protects” the innovator from the working with the structure of rules and proce-
constraining rules, policy, systems, and any dures within a culture can actually promote
19-House.qxd 3/19/2004 5:34 PM Page 618
Table 19.1 Higher Uncertainty Avoidance Societies Versus Lower Uncertainty Avoidance Societies
• Have a tendency toward formalizing their • Have a tendency to be more informal in their
interactions with others interactions with others
• Document agreements in legal contracts • Rely on the word of others they trust rather
than contractual arrangements
• Be orderly, keeping meticulous records, • Be less concerned with orderliness and the
documenting conclusions drawn in meetings maintenance of records, often do not document
the conclusions drawn in meetings
• Rely on formalized policies and procedures, • Rely on informal interactions and informal
establishing and following rules, verifying norms rather than formalized policies,
communications in writing procedures and rules
• Take more moderate calculated risks • Be less calculating when taking risks
• Inhibit new product development but • Facilitate the new product development
facilitate the implementation stage through especially in the initiation phase, through
risk aversion and tight controls higher risk taking and minimal planning or
controls
• Show stronger resistance to change • Show less resistance to change
• Show stronger desire to establish rules • Show less desire to establish rules to dictate
allowing predictability of behavior behavior
• Show less tolerance for breaking rules • Show more tolerance for breaking rules
innovation. It is possible that people in more showing cultural attributes that tend to cluster
formalized organizations can be more intellec- together; thus societies may be expected to have
tually flexible than their counterparts in less a mixture of the extreme attributes presented.
formalized ones.
1. In this society, orderliness and consistency are stressed, even at the expense of experimentation and
innovation. (reverse scored)
Strongly Neither agree nor Strongly
agree disagree disagree
1 2 3 4 5 6 7
2. In this society, societal requirements and instructions are spelled out in detail so citizens know what
they are expected to do. (reverse scored)
Strongly Neither agree nor Strongly
agree disagree disagree
1 2 3 4 5 6 7
1. I believe that orderliness and consistency should be stressed, even at the expense of experimentation
and innovation. (reverse scored)
Strongly Neither agree nor Strongly
agree disagree disagree
1 2 3 4 5 6 7
2. I believe that societal requirements and instructions should be spelled out in detail so citizens know
what they are expected to do.
Strongly Neither agree nor Strongly
agree disagree disagree
1 2 3 4 5 6 7
was determined in each case by the psychometric Avoidance, respectively, with regard to values.
analyses reported in Chapter 8. The items were coded so that strong agreement
Four items measured current GLOBE- with the statements resulted in a high Uncertainty
defined Uncertainty Avoidance practices at the Avoidance scale of scores. Details of scale vali-
societal level. Two sample items are shown in dation are provided in Chapter 8.
Table 19.2. The questions address the extent to The manifestation of Uncertainty Avoidance
which life is structured, predictable, orderly, and in values systems, as well as in current prac-
consistent, and rules and regulations to regulate tices, underlies the rationale for two separate
societal practices are emphasized. Table 19.4 measures of this construct in the GLOBE pro-
shows sample questionnaire items designed to ject. In the following sections, the findings from
measure existing organizational practices with the GLOBE study are presented and set against
regard to Uncertainty Avoidance. Three of the the backdrop of other empirical evidence from
items in this scale are isomorphic; the corre- cross-cultural research. Uncertainty Avoidance
sponding societal scale may be seen in Table is examined as an aspect of practices and values
19.3. Tables 19.4 and 19.5 show the items that at both societal and organizational levels and
are isomorphic with those in the practices scales the implications of this cultural dimension for
for societal and organizational Uncertainty organizational leadership are explored.
19-House.qxd 3/19/2004 5:34 PM Page 620
1. In this organization, orderliness and consistency are stressed, even at the expense of experimentation
and innovation. (reverse scored)
Strongly Neither agree nor Strongly
agree disagree disagree
1 2 3 4 5 6 7
2. In this organization, job requirements and instructions are spelled out in detail so employees know
what they are expected to do. (reverse scored)
Strongly Neither agree nor Strongly
agree disagree disagree
1 2 3 4 5 6 7
1. In this organization, orderliness and consistency should be stressed, even at the expense of
experimentation and innovation. (reverse scored)
Strongly Neither agree nor Strongly
agree disagree disagree
1 2 3 4 5 6 7
2. In this organization, job requirements and instructions should be spelled out in detail so employees
know what they are expected to do. (reverse scored)
Strongly Neither agree nor Strongly
agree disagree disagree
1 2 3 4 5 6 7
Standard
Variable Mean Deviation Minimum Maximum Valid N
N = 61 cultures.
The scores ranged from 2.88 (Russia) to 5.37 Uncertainty Avoidance only modestly in their
(Switzerland). Also shown are country groupings, values tend to exhibit this cultural characteristic
ranked similarly in descending order of scale rather strongly in their practices. There is an
values. Society groupings were carried out using exception to this that is notable. GLOBE
two statistical procedures: Scheffe’s post hoc respondents from China reported highest band
test, and the standard error of difference banding on both values and practices of Uncertainty
procedure (see Chapter 8, by Hanges & Dickson). Avoidance, as reported in Tables 19.7 and 19.8.
The two results were very similar; the grouping An examination of Table 19.7 shows that most
shown here represents the somewhat clearer of the countries with high reported uncertainty
results from the second procedure. The groups’ avoidance practices are technologically developed
ranks imply that countries shown within a band and those with low reported practices are techno-
were not meaningfully different in their scale logically developing countries. Table 19.8 shows
mean scores, but were significantly different that the opposite is true with regard to values.
from countries of a different band. Table 19.8 Technology seems to be the underlying ele-
gives societal values (Should Be) scores, with ment of distinction between the two groups. All
societies rank-ordered from highest to lowest things considered, it appears that the difference
in the GLOBE Uncertainty Avoidance cultural between the practices and values scores is a
dimension.18 The scores range from 3.16 reflection of the current national level of uncer-
(Switzerland) to 5.61 (Thailand). tainty caused by unresolved issues. For devel-
oping nations, the positive difference would
imply that members of these societies perceive
The Correlation Between
themselves as being on the road to, but not quite
Overall Practices and Values Scores
accomplished in, reducing the uncertainties in
The Pearson correlation between GLOBE soci- their lives. As they acquire more sophisticated
etal Uncertainty Avoidance practices and technologies, over the long term, this difference
GLOBE societal Uncertainty Avoidance values between their practices and values scores may
was negative (r = −.62, p < .01) across the 61 grow smaller. For most of the developed
GLOBE cultures used in the analysis. The grand nations, however, the negative difference would
mean for societal practices was 4.16, whereas imply that members of these societies see exces-
for societal values it was 4.62, suggesting that sive structure in their environment, leading
the societies reflect Uncertainty Avoidance perhaps to boredom and a resulting desire to
more in their values than in their practices have less structure in their lives. In the work-
(difference statistically significant t(60) = 3.30, place, we may see this in the use of such things
p < .01). These findings suggest that cultures as flextime. Employees in Germany and
reported in the GLOBE survey as having strong Switzerland have embraced flextime (see
Uncertainty Avoidance values have weaker Hofstede, 2001, p. 168). The respondents in the
Uncertainty Avoidance practices, with only a few GLOBE surveys from both of these countries
exceptions. In contrast, the cultures reporting ranked very low in Uncertainty Avoidance
19-House.qxd 3/19/2004 5:34 PM Page 622
Band
A B C D
a Germany (West): Former FRG * Higher scores indicate greater uncertainty avoidance.
b Germany (East): Former GDR Our response bias correction procedure identified response
c Switzerland (French-speaking) bias in some countries for this scale (see endnotes).
d South Africa (Black sample)
e Canada (English-speaking)
f South Africa (White sample)
values, indicating their desire for more flexible As can be seen from the table, there is much
schedules. In the leisure market, we see the correspondence among industries in the reported
emergence of sporting activities that test the societal Uncertainty Avoidance scores for both
limits of human endurance, abilities, and skills, the current practices and values dimensions.
serving to enhance the uncertainty of outcomes
for individuals within industrialized societies.
Correlation With Other
Overall Scores and Industry Scores GLOBE Society Culture Dimensions
Each culture in the GLOBE study sampled up In this section we review the relationships
to three industries: Financial services, food pro- between managerial responses to the GLOBE
cessing, and telecommunications. Table 19.9 societal-level Uncertainty Avoidance practices
shows the correlations among industries on both scale and responses to other scales. The correla-
societal values and practices scores across cultures. tions of societal Uncertainty Avoidance practices
19-House.qxd 3/19/2004 5:34 PM Page 623
Band
A B C D E
Country Score Country Score Country Score Country Score Country Score
Thailand 5.61 Slovenia 4.99 Kazakhstan 4.42 Switzerlandd 3.83 Netherlands 3.24
Nigeria 5.60 Brazil 4.99 Israel 4.38 Denmark 3.82 Switzerland 3.16
Albania 5.37 Colombia 4.98 Japan 4.33 Canadae 3.75
Iran 5.36 Malaysia 4.88 France 4.26 Austria 3.66
Egypt 5.36 Guatemala 4.88 Singapore 4.22 Sweden 3.60
El Salvador 5.32 Qatar 4.82 England 4.11 Germanyf 3.32
Morocco 5.32 South Africaa 4.79 New Zealand 4.10
Taiwan 5.31 Kuwait 4.77 Ireland 4.02
China 5.28 Spain 4.76 U.S. 4.00
Venezuela 5.26 Zimbabwe 4.73 Australia 3.98
Mexico 5.26 India 4.73 Germanyc 3.94
Georgia 5.24 Poland 4.71 Finland 3.85
Indonesia 5.23 Bolivia 4.70
Ecuador 5.16 Turkey 4.67
Philippines 5.14 Zambia 4.67
Namibia 5.13 South Korea 4.67
Greece 5.09 South Africab 4.67
Russia 5.07 Hungary 4.66
Argentina 4.66
Hong Kong 4.63
Costa Rica 4.58
Italy 4.47
Portugal 4.43
a South Africa (Black sample) * Higher scores indicate greater uncertainty avoidance.
b South Africa (White sample) Our response bias correction procedure identified response
c Germany (East): Former GDR bias in some countries for this scale (see endnotes).
d Switzerland (French-speaking)
e Canada (English-speaking)
f Germany (West): Former FRG
and values scores with the other eight dimensions GLOBE Uncertainty Avoidance practices,
of culture assessed in the GLOBE program are uncertainties are reduced through institu-
presented in Table 19.10. tional collectives; however, with less reliance
on in-group collective practices. This tendency
toward collective institution interests helps
Uncertainty Avoidance
manage technology and information, as well as
Practices and Other GLOBE Dimensions
skills and knowledge. The managing of uncer-
Respondents to the GLOBE questionnaire tainty and risk creates an outlook that pro-
show that Uncertainty Avoidance practices are motes long-term thinking. This in turn allows
positively correlated with the practices of Future societies to focus on performance, with less of
Orientation, Institutional Collectivism, and an inclination for societal power differential
Performance Orientation; these same practices status. In addition, Uncertainty Avoidance
are negatively correlated with Power Distance practices scores are inversely correlated with
and In-Group Collectivism. These relation- values scores for Future Orientation and both
ships suggest that, in societies scoring high on Institutional and In-Group Collectivism. Many
19-House.qxd 3/19/2004 5:34 PM Page 624
Table 19.9 Correlation Between Industries on Uncertainty Avoidance Societal Values and Practices
Scores Across Cultures
Values
Finance .84** .90**
(41) (30)
Food .77**
(22)
Practices
Finance .94** .78**
(41) (30)
Food .82**
(22)
Table 19.10 Correlations of Societal Uncertainty Avoidance With Other Societal Cultural Dimensions
N = 61
* Significant at p < .05.
** Significant at p < .01.
19-House.qxd 3/19/2004 5:34 PM Page 625
societies practicing uncertainty avoidance seem an overall strong preference for managing
more assured that their collectives are handling uncertainty to create a secure environment. For
uncertainty, thus they value more flexibility and example, in a study comparing the United States
are more open to change. with China, text analysis revealed that achieving
mutual understanding, mutual benefit, and
mutual trust, and long-term cooperation is
Uncertainty Avoidance
important in reaching agreement on business
Values and Other GLOBE Dimensions
standards if negotiating in China (Zhao, 2000).
Respondents to the GLOBE questionnaire In a study of Hong Kong-based small companies
show that Uncertainty Avoidance values are conducting business in China, researchers
positively correlated with the values of Future reported that social meetings in restaurants and
Orientation and both Institutional and In-Group sending gifts were the two most common ways
Collectivism; these same values of Uncertainty of relationship building (guanxi), and that these
Avoidance are negatively correlated with relationships were essential for business success
Gender Egalitarianism. We spectulate that if (Leung & Yeung, 1995). This development of
societies desire more management of uncer- trust in relationships reduces the uncertainty in
tainty, members may be motivated to look to the society. A common saying from the philoso-
institutional and in-group collectives to accom- phy of Confucius, which is the foundation of
plish this. As a desire for uncertainty manage- much of the Chinese society, dictates that sons
ment increases, women may play a part within should obey their fathers, and wives should obey
the group and institutional collective, but only husbands. Another common saying is that the
through more traditional female roles. In middle road is best. By conforming to this philos-
societies desiring more management of uncer- ophy, society members maintain harmony and
tainty, members may be encouraged to use such respect for structures of the state, which in turn
techniques as long-term planning, which in turn reduces ambiguity and allows predictability in
leads to better management of technology and the society. These examples illustrate the
information. In addition, Uncertainty Avoidance importance of managing certainties through val-
values scores are positively related to Power ues and practices in a society.
Distance, In-Group Collectivism, and Humane
Orientation practices and inversely related to
The Relationship of
practices scores for Future Orientation. Members
Collectivism and Power
of many societies valuing uncertainty avoidance
Distance to Uncertainty Avoidance
may prefer more of a stratified distribution of
power within the society and the societal group The dimensions of national culture proposed
to which they belong, relying on such structures by Hofstede (1980) were framed on the basis of
(practices) to fulfill the desires (values) of man- relative theoretical and empirical independence.
aging uncertainty. Further, societies valuing A possible association between uncertainty
uncertainty avoidance tend to favor caring and avoidance and individualism–collectivism is
generous practices within these collectives. worth mentioning. On a logical basis, one may
However, societies hoping to manage uncer- speculate on collectivism as an uncertainty
tainty do not tend to favor techniques such as avoidance mechanism. The old adage “united
scenario planning to accomplish the reduction we stand, divided we fall” would help explain
of uncertainty. why a society may emphasize collective values
Interestingly, several societies’ rankings on to overcome the uncertainties associated with
the GLOBE scores do not vary between prac- threats to its existence. The empirical correla-
tices and values: China, where members practice tion of Hofstede’s (1984, p. 213) UAI with indi-
and value the strong avoidance of uncertainty, vidualism (r = −.35, p < .05, across 40 countries)
tends to practice strong Institutional and In- supports this line of reasoning. The GLOBE study
Group Collectivism, yet tends to be only mod- affords separate examination of Uncertainty
estly Future Oriented. It seems the institutional Avoidance practices and values. The correlation
and in-group collective practices combine with between GLOBE Uncertainty Avoidance
19-House.qxd 3/19/2004 5:34 PM Page 626
practices and Institutional Collectivism practices produced comparable data are by Schwartz
(r = .40, p < .01, across 61 cultures) supports (1999) and Inglehart and Baker (2000).
the conceptual relationship expressed above.
GLOBE Uncertainty Avoidance values correlate
Relation to Hofstede’s Rankings
−.19 with GLOBE Institutional Collectivism
practices, which was nonsignificant. It should be noted that there are inherent dif-
In like manner, it may be that the cultural ficulties comparing the GLOBE rankings with
dimension referred to as Power Distance is also those obtained by Hofstede (1980). As outlined
an outcome of Uncertainty Avoidance. What earlier in this chapter, Hofstede’s uncertainty
Lebas and Weigenstein (1986) observed with avoidance scale had three items measuring one
respect to organizations may also be extended to ecological variable (employment stability) and
societies. Groups or individuals derive power two individual-level variables (rule orientation
and status partly from the uncertainty inherent and stress) aggregated to the societal level. In
in the environment. The greater the uncertain- contrast, the GLOBE measure used four items
ties faced by that society, the greater the need to measuring individual-level variables on a 7-point
reduce them. Under such circumstances, the scale and emphasized consistency–orderliness,
greater is the power of those who control that structured lifestyles, explicit specification of
environment and the greater the public’s accep- societal requirements, and societal regulations.
tance of their authority. However, the actual Further, the GLOBE items formed two distinct
reduction of uncertainties in society is rarely measures—one reflecting respondents’ percep-
achieved by any single individual, but through tions of current societal practices and the other
an organization of forces and the creation of reflecting values. In contrast, one of Hofstede’s
structure and systems. Once these are in place, items measured a value orientation (rule orienta-
the perceived need for leadership may be dimin- tion), the second assessed an expectation of
ished. These arguments suggest a complex the- future outcome (the amount of time a respondent
oretical relationship between power distance expected to remain with the firm), and a third
and uncertainty avoidance. In the GLOBE data, expressed an outcome (the experience of stress).
a strong negative correlation between GLOBE The scores on these three items were formed
Uncertainty Avoidance practices and Power into a composite index so that separation of
Distance values (r = −.50, p < .01, across values from practices for the purpose of a clear
61 cultures) suggests that highly organized and comparison becomes impossible (see Hanges
structured societies no longer endorse power Appendix A).
hierarchies in their current practices. That the Therefore, a comparison of rankings from
acceptance of power hierarchies may be associ- the two studies is not easily interpretable.
ated with a concurrent need for the incor- That said, the GLOBE Uncertainty Avoidance
poration of structure, orderliness, and consistency practices scores exhibit the following correla-
is borne out by a strong positive correlation tions with Hofstede’s UAI: r = −.62, rho = −.60,
between Power Distance practices and Uncer- p < .01. At first glance, this significant neg-
tainty Avoidance values (r = .48, p < .01, across ative correlation may seem unexpected; how-
61 cultures). ever, it is consistent with the overall GLOBE
finding that cultural practices are negatively
correlated with cultural values scales. The
COMPARISON OF GLOBE GLOBE Uncertainty Avoidance values scores
FINDINGS WITH PREVIOUS STUDIES have the following correlations with Hofstede’s
UAI: r = .35, rho = .36, p < .05. This indicates a
It is interesting to compare the GLOBE findings moderate positive relationship between Hofstede
with other cross-cultural studies that have inves- UAI and GLOBE societal level values scores.
tigated the same or similar variables. The most Although modestly related, these comparisons
well-known study based on questionnaire data indicate that Hofstede’s measure of uncertainty
similar to the GLOBE study was conducted by avoidance is not the same as that measured in
Hofstede (1980, 2001). Other studies that have the GLOBE study.
19-House.qxd 3/19/2004 5:34 PM Page 627
Results from a variety of studies may serve long-term planning than would United States
to highlight these differences. In a qualitative companies, given the assumption that Japan was
study of Demark and France regarding their a higher uncertainty avoidant society (Ueno &
tendency toward rule orientation, Schramm- Sekaran, 1992). GLOBE respondents’ scores
Nielsen (2000) found disparities with the from Japan and the United States were not signifi-
Hofstede findings. Contrary to expectations, the cantly different on societal practices, and Japanese
French respondents did not report that they respondents ranked lower than American
refrain from bending or breaking company respondents on societal values.
rules; the Danish respondents were not averse to
structuring activities, and unlike the French,
The Schwartz Study
were more likely to actually obey the rules.
As stated above, the Hofstede UAI scale Schwartz (1999) identified three fundamental
has questionable face validity and can be inter- values relating to all cultures: (a) autonomy ver-
preted as a variety of outcomes, whereas the sus embeddedness, (b) mastery versus harmony,
GLOBE Uncertainty Avoidance scales have and (c) hierarchy versus egalitarianism. The
face validity in, and were validated against, concept most salient for Uncertainty Avoidance
unobtrusive and outcropping measures (see is autonomy versus embeddedness.
Chapter 9, by Gupta, Sully de Luque, & House) In autonomous cultures, individuals are
and within the context of a nomological net (see perceived as autonomous, bounded entities,
Chapter 8, by Hanges and Dickson). finding meaning in life by appreciating their
An important deviation from the Hofstede uniqueness. Two types of autonomy are distin-
data (1984, 2001) should be highlighted because guished in Schwartz’s framework. Intellectual
of the extant research that has been conducted autonomy is realized as individuals are encour-
on the basis of his ranking. GLOBE respon- aged to follow their own ideas and intellectual
dents from the country of Japan reported directions, with an emphasis on creativity,
Uncertainty Avoidance scores in the lower half curiosity, and broadmindedness. Affective auto-
of Uncertainty Avoidance practices, and moder- nomy is realized as people are encouraged to
ate levels of Uncertainty Avoidance values. find positive experiences for themselves, with an
Hofstede reported the Japanese society to rank emphasis on variety, excitement, and pleasure in
among the highest of his sample on UAI.19 life. Conversely, in embedded cultures people
GLOBE respondents from Japan reported sig- are perceived as compelled by the collectivity,
nificantly lower scores for both practices and finding life’s meaning through participating in
values than did respondents from the United the group and identifying with its goals. Respect
States, quite contrary to the Hofstede rankings. for tradition, social order, obedience, and family
It should be recognized that in Japan the con- security are some of the values emphasized.
cepts of ambiguity and uncertainty do not nec- Cultures embedded in relationships generally
essarily have undesirable connotations, as they encourage social order for security. On the other
do in the United States (Pascale & Athos, 1983). hand, autonomous cultures tend to encourage
Indeed, many cross-cultural researchers who contractual types of behavior that are less based
include Japan in their sample assume Japan to on social order (see Granovetter, 1985). Indeed, as
be a high uncertainty avoidance society in reported in Table 19.11, Schwartz’s autonomy-
accordance with the Hofstede (1980) paradigm. embedded scale (r = .42, p < .01) is positively
When confronted with contrary results, much related with GLOBE’s Uncertainty Avoidance
discussion usually follows as to why Japan practices scale and negatively related to the Uncer-
was not found to be a high uncertainty avoidant tainty Avoidance values scale (r = −.73, p < .01).
culture. For example, in a study comparing Further, GLOBE’s Uncertainty Avoidance
planning and control practices in Japan and practices correlate positively with affective and
the United States, no differences were found intellectual autonomy (r = .44, p < .01 and .30,
in planning time horizons between the two p < .05) and negatively with Schwartz’s embed-
societies, although it was predicted that Japanese dedness scale (r = −.41, p < .01). GLOBE’s
companies would engage in significantly higher Uncertainty Avoidance values scale is correlated
19-House.qxd 3/19/2004 5:34 PM Page 628
NOTE: We thank Shalom Schwartz and Lilach Sagiv for providing the above correlations.
N = 48.
* Correlation significant at p < .05.
** Correlation significant at p < .01.
negatively with affective and intellectual autonomy secular-rational authority, is connected with
(r = −.67, p < .01 and −.61, p < .01) and positively early industrialization and the ascendancy of the
with Schwartz’s embeddedness scale (r = .74, working class and reflects the process of mod-
p < 0.01). As such, Uncertainty Avoidance society ernization. Prominent in preindustrial and pre-
values scores share the belief in social order and modern societies, a traditional worldview
security engaging in long-term interaction reflects an emphasis on obedience to traditional
inherent in embeddedness, as opposed to behav- authority, particularly deference to God that
ior based on more contractual relationships goes with deference to the family and nation.
associated with autonomy. However, the Uncer- Conversely, a secular worldview is prominent in
tainty Avoidance society practices scores share industrial and modern societies, where authority
the importance of behavior based on more is legitimated by rational-legal norms, empha-
contractual relationships. sizing economic accumulation and individual
achievement. According to Inglehart and Baker
(2000), the societal development process is
The Inglehart and Baker Study
associated with an increased probability of a
The World Values Survey evaluates more shift first from traditional to secular rational
than 70 societies and includes responses across authority, and second from survival to self-
societies covering a range of economic, politi- expression values.
cal, and cultural issues, from approximately 350 As reported in Table 19.12, the GLOBE
questions on human values (Inglehart, Basanez, societal Uncertainty Avoidance practices scale
& Moreno, 1998). In the World Values Surveys, is positively and significantly correlated with
two dimensions surface along which cultures both the traditional–secular rational authority
appear to be differentiated (Inglehart & Baker, scale (r = .55, p < .01) and the survival–self-
2000, p. 21). The first, termed survival versus expression scale (r = .78, p < .01). Respondents
self-expression values, indicates the changes in societies that report higher Uncertainty
associated with the prosperous conditions Avoidance practices scores are more likely to
created in advanced industrial societies. This encourage traditional behavior emphasizing
dimension is revealed through the increase of hard work and self-denial, which will be
the knowledge and service sectors, reflected reflected in the survival scale. In contrast, the
in the concepts of postmodernization. The GLOBE societal Uncertainty Avoidance values
second dimension, termed traditional versus scale is negatively and significantly correlated
19-House.qxd 3/19/2004 5:34 PM Page 629
Table 19.12 GLOBE Uncertainty Avoidance and Inglehart’s Modernization and Postmodernization
Dimensions
scale scores was positive, indicating the higher the extent to which infrastructure development
the level of Uncertainty Avoidance practices and maintenance is planned and financed, and
in a society, the more its economic prosperity. the extent to which the values of hard work
In practical terms, a society that currently has and innovation support competitiveness. Across
formalization of social expectations and a high 40 cultures, the correlation is .44 ( p < .01) between
degree of orderliness enjoys more telephone societal support for economic competitiveness and
lines, cellular phones, personal computers, and GLOBE Uncertainty Avoidance practices. The
fax machines per 1,000 inhabitants. However, correlation with GLOBE Uncertainty Avoidance
equally interesting is the even greater negative values (r = − .24) is not statistically significant.
correlation with GLOBE Uncertainty Avoidance
values (r = −.80, p < .01), indicating that the World Competitiveness. Across 40 cultures
higher a society’s level of economic prosperity, common to both data sets, GLOBE Uncertainty
the less that society endorsed the value of Avoidance practices correlated .60 (p < .01) with
Uncertainty Avoidance. These results would the competitiveness index (for details see chapter
suggest that prosperous societies do not value 7 by Javidan & Hauser), signifying that more for-
the very factors that seem to have benefited malized and structured economies are more
them. On the other hand, in poor nations, char- competitive in the global environment. The corre-
acterized by low levels of rule orientation, sponding correlations of GLOBE Uncertainty
people recognize the value of rules. Avoidance values are −.49 (p < .01) with the
World Competitiveness Index. That the direction
Government Support for Economic Prosperity. of relationship with practices rankings is the oppo-
In Chapter 7, Javidan and Hauser proposed site of that with values is consistent with our earlier
three hypotheses around uncertainty avoidance. observed negative correlation between GLOBE
One of these, Hypothesis 8.3, stipulated that Uncertainty Avoidance practices and values.
countries that are high on uncertainty avoidance
tend to have governments that support economic
Uncertainty Avoidance
activities. Our results confirm this hypothesis.
and the Human Condition
We found that the extent to which a society
has high Uncertainty Avoidance practices is The correlations of GLOBE Uncertainty
highly correlated with governmental support for Avoidance with a number of variables relating to
economic development of the country (r = .74, societal well-being are reproduced in Table 19.14.
p < .01). This indicates that in uncertainty-
avoiding societies, the public sector is able to Societal Health. In uncertainty-avoiding societies,
function without much political interference and we may expect to find, in relative terms, fair
there is relatively less bribery and corruption. In administration of the justice system, personal
such societies, technological development is not security and protection of private property, higher
constrained by lack or insufficiency of financial quality of life, and relatively fewer problems
resources and is not restricted by the legal envi- posed by alcoholism and drug abuse in the work
ronment. In addition, the prevalence of competi- place. The correlation across 40 cultures was .76
tion laws discourages unfair business practices, (p < .01) with GLOBE Uncertainty Avoidance
economic literacy is high, and industrial rela- practices, and −.74 (p < .01) with GLOBE
tions are generally productive rather than hostile. Uncertainty Avoidance values. However, these
societies do not necessarily enjoy a high level
Societal Support for Economic Competitiveness. of physical well-being, as reflected by the low
The prevalence of rules and regulations, orderli- correlations with the variable “human health.”
ness, and structure in a society, as measured by
the GLOBE Uncertainty Avoidance practices Life Expectancy. Across 56 cultures common to
scales, is positively related to the extent to which both datasets, life expectancy was negatively
its political and social systems are adapted to associated with GLOBE Uncertainty Avoidance
long-term competitiveness in world markets, the values (r = −.44, p < .01), and positively corre-
extent to which its labor regulations are flexible, lated (r = .28, p < .05) with GLOBE Uncertainty
19-House.qxd 3/19/2004 5:34 PM Page 631
Correlation With
Uncertainty Uncertainty
Economic Variables N Avoidance Practices Avoidance Values
Avoidance practices. Predictably, then, we may up of the following items: Life expectancy,
expect to find a longer average life span in knowledge, and standard of living. In general,
countries where Uncertainty Avoidance prac- therefore, societies with greater uncertainty
tices are experienced and Uncertainty Avoid- avoiding mechanisms tended to enjoy higher
ance values are not endorsed. However, the qualities of life and human development.
driving force behind this correlation is probably
the availability of advanced medical technolo- Civil Liberties Index. This variable measures
gies and health services in the economically the extent to which members of a society enjoy
developed countries, which also rank lower on freedom in several areas, with higher values
Uncertainty Avoidance values. indicating less liberty for individuals. A moder-
ate negative correlation between GLOBE
General Satisfaction. In Chapter 7, Javidan and Uncertainty Avoidance practices and this index
Hauser proposed the hypothesis (Hypothesis (r = −.36, p < .01) supports the notion that
8.1) that societies that are high on uncertainty in societies characterized by a high degree of
avoidance tend to enjoy a greater general satis- orderliness, consistency, and structure, there
faction. This measure, which consists of items is greater assurance of individual liberty.
on people’s general feelings about themselves, However, the reason that one may not expect a
their health, their lives, and their countries, cor- stronger correlation is that democratic cultures
related .63 (p < .01) with GLOBE Uncertainty characterized by lower levels of rule orientation
Avoidance practices. These results suggest that may, nevertheless, have a high level of civil
people in presently rule-oriented cultures enjoy liberties. In societies that have a low level of
a better general satisfaction, confirming the civil liberties, the people are desirous of more
earlier hypothesis. structure and social order, as evidenced by a
strong positive correlation of civil liberties
Human Development Index. For the composite index with GLOBE Uncertainty Avoidance
index of human development, there is a mod- values (r = .64, p < .01).
est but significant positive correlation (r = .28, Similar results are found using the World
p < .05) with GLOBE Uncertainty Avoidance Values Survey findings (Inglehart et al., 1998).
practices, and a large negative correlation As reported in Table 19.15, rule-oriented
(r = −.59, p < .01) with GLOBE Uncertainty societies that scored higher on Uncertainty Avoid-
Avoidance values, across 56 cultures. Recall ance tend to have a higher regard for democra-
that the Human Development Index was made tic systems (r = −.51, p < .01), individual
19-House.qxd 3/19/2004 5:34 PM Page 632
Human
Uncertainty Societal Human Life General Development Civil
Avoidance Health Health Expectancy Satisfaction Index (HDI) Liberties
involvement in the political system (r = −.52, their society as formalized and orderly (i.e.,
p < .01), greater voice for people in the running scoring high in GLOBE Uncertainty Avoidance
of the country (p = −.53, p < .01), a greater practices) are those that also enjoy a high
admiration for democracy (r = −.49, p < .01), and quality of life and relative safety and security.
a less-active role for the government (r = −.62, Life expectancy in these societies tends to be
p < .01). The negative signs are due to the way somewhat higher. People in these societies are
these scales are measured (see Chapter 7, generally more satisfied and happy. The Human
Javidan & Hauser). In contrast, regarding Development Index, which includes measures
Uncertainty Avoidance values, respondents in of life expectancy, adult literacy, and standard
societies in the GLOBE sample reported to of living, yields correlations with GLOBE
desire more order have less of a regard for Uncertainty Avoidance that are consistent with
democracy (r = .40, p < .05), are more passive the other results in the table.
(r =.60, p < .01), have little voice in the running
of the country (r = .75, p < .01), dislike democ-
Uncertainty Avoidance
racy (r = .38, p < .05), and aspire to more rules
With Science and Technology
and structure, a more active government (r = .81,
p < .01), and more stability (r = .56, p < .01) in Perhaps in no other realm of human
their communities. endeavor would we expect uncertainty avoid-
In sum, as far as the human condition is con- ance, defined in terms of formalization and
cerned, societies whose respondents perceive structure, to be more influential than in the
19-House.qxd 3/19/2004 5:34 PM Page 633
conduct and progress of science and technology. returned the highest GLOBE Uncertainty
In Chapter 7, Javidan and Hauser hypothesized Avoidance practices scores, and Protestant cul-
(Hypothesis 8.2) that societies that are high on tures the lowest (Verweij, Ester, & Nauta, 1997).
uncertainty avoidance tend to enjoy stronger Because empirical studies of religious practices
scientific progress. Our findings confirm this are based primarily on church attendance, only
hypothesis. As Table 19.16 shows, societies Christian countries have been targeted for investi-
that are reported to have high Uncertainty gation. Across 16 countries, Verweij and col-
Avoidance practices also tend to be successful leagues (1997) found a high correlation (r = .77,
in basic science research. Across 40 cultures p < .05) between the percentage of Catholics in a
common to both datasets, the correlation country and Hofstede’s UAI. However, across the
between progress of science and technology is 12 countries common to the GLOBE and Verweij
.59 (p < .01) with reported Uncertainty and colleagues’ (1997) datasets, the percentage
Avoidance practices. In formalized and orderly of regular church attendance (defined as one or
cultures, therefore, one may also find more more times per week) correlated −.69 (p < .05)
company–university cooperation, more empha- with GLOBE Uncertainty Avoidance practices
sis on basic science and research, and more scores, and .34 (nonsignificant) with Uncertainty
interest among youth regarding science and Avoidance values scores—implying a lower rate
technology issues. There was a strong negative of church attendance in societies with high
relationship between GLOBE Uncertainty Uncertainty Avoidance practices and possibly a
Avoidance values and science and technology high rate of church attendance in cultures with
(r = −.58, p < .01). high Uncertainty Avoidance values. Verweij and
colleagues (1997) had also constructed a religious
orientation index derived from a factor score. This
Uncertainty Avoidance With Religion
index correlated −.73 (p < .05) with reported
The theoretical role of religion, viewed by GLOBE Uncertainty Avoidance practices scores
Hofstede (1980) as a coping mechanism for and .37 (nonsignificant) with the GLOBE Uncer-
uncertainty avoidance was touched on earlier in tainty Avoidance values scores, consistent with
this chapter. Several sources provide data for the relationships found with church attendance.
investigation of the empirical relation between The findings conveyed here show a negative rela-
these constructs. It is generally expected that tionship between reported GLOBE Uncertainty
people in cultures low in uncertainty avoidance Avoidance practices and degree of religiosity in a
would exhibit a lesser degree of religious society, which is contrary to the conventional
involvement to the extent that they do not need wisdom and Hofstede’s supposition.
to use religion to help them cope spiritually or Further supportive evidence is provided
practically with the uncertainties in life. by GLOBE’s measures of religiousness. In
The empirical evidence is not unequivocal. At Chapter 7, Javidan and Hauser developed a scale
the societal level, overall indices of religiosity of religious devotion and a scale of religious
show that predominantly Catholic countries dogma. Table 19.17 shows the correlations
19-House.qxd 3/19/2004 5:34 PM Page 634
between these scales and GLOBE measures of common fundamental tenets, differ in their
Uncertainty Avoidance. As can be seen from prescriptions for social order. Whereas all reli-
this table, GLOBE Uncertainty Avoidance prac- gions by definition promote the acceptance of
tices scores are negatively correlated with uncertainty by the very acknowledgment of a
degree of religious devotion (r = −.42, p < .01) higher power, different religions impose differ-
and degree of religious dogma (r = − .33, ent degrees of restriction or prescription for
p < .05). Societies that are reported to have high personal and social behavior.
Uncertainty Avoidance practices attach a lower Still, it would be difficult to make a state-
level of importance to religion, and less fre- ment about a society’s tolerance of uncertainty
quently attend religious services. On the other from religious faith per se. Practically every
hand, societies reported to value (aspire toward) major religion has different sects that have
more rules in their lives tend to attach a greater emerged through differences in the interpreta-
significance to religion, attend religious services tion of religious texts. Thus Hinduism, one of
more frequently, and have stronger beliefs in the oldest religions in existence, has seen differ-
heaven and hell, indicating a greater value of ent phases in its own following: In the reac-
religious devotion (r = .43, p < .01) and religious tionary birth of Buddhism when the practice of
dogma (r = .48, p < .01). Hinduism became too ritualistic, and in the birth
To summarize, it appears that the relation- of the Sikh faith as a militant response to the
ship between Uncertainty Avoidance and reli- perceived threat of Islam against the passive
gion is more complicated than that envisioned stance of Hinduism. Similar undercurrents may
by Hofstede and presumed in the literature. be observed in the Western world in the emer-
Those societies that are reported to have proce- gence of Christianity from the Jewish faith, in
dures and practices in place to reduce uncertain- the Catholic and Protestant persuasions, and
ties in peoples’ lives do not necessarily need in the scores of divisions that have occurred in
religion to create further structure and certainty. these faiths. There exists no extensive or inten-
Only respondents in societies reporting defi- sive study of the relation between different
cient processes and procedures, as well as those faiths and the extent to which they incorporate
desiring a higher level of structure and clarity in the tolerance or avoidance of uncertainty,
their lives, need religion to fill the gap. although we can speculate on their emergence
In addition to the findings presented in this as a historical response to uncertainties in the
chapter, it is very likely that many aspects of social milieu of the times.
cultures are culture-specific and, thus, are not
likely to be reflected in the findings reported in Uncertainty Avoidance
this chapter or others. An illustrative case per- and Financial Transactions
tains to the association between uncertainty
avoidance and religion, which has been posed as The relationship between GLOBE Uncertainty
a coping mechanism of societies (Hofstede, Avoidance scores and cash holding was exam-
1984, p. 111). The GLOBE data described here ined using data on 14 developed nations pub-
suggest that religiosity, or the extent of religious lished by Humphrey, Pulley, and Vesala (1996).
beliefs in a society, has a complex relationship It was hypothesized that societies exhibiting a
with the level of Uncertainty Avoidance in preference for high uncertainty avoidance
that society. Hofstede (2001) has discussed would have high cash holdings. The practice
(pp. 176–177) and summarized (p. 181) the of holding cash reserves is proposed to reduce
association between type of religion and his uncertainty in high uncertainty avoidance
UAI. The issues run deeper than these simplistic societies (Humphrey et al., 1996). Across the
explanations, however. 12 countries common to both datasets, the cor-
First, there are countries that are almost relation between Uncertainty Avoidance prac-
entirely of a single religious persuasion (e.g., tices scores and cash holdings per capita is .07
Pakistan, Iraq), while at the other extreme there (p > .10, two-tailed). The corresponding
are highly secular states (e.g., India, the United correlation with Uncertainty Avoidance values
States). Religions, although possessing many scores is −.27 (p > .10, two-tailed). Of the
19-House.qxd 3/19/2004 5:34 PM Page 635
12 countries analyzed, all but 2 countries are into 10 geographical regions (details of these
among the 16 developed countries that scored clusters are discussed in Chapter 10 by Gupta &
lowest on Uncertainty Avoidance values. If Hanges). Analysis of variance tests (Table
these two countries (Italy and Japan) are 19.18) were conducted to determine if cross-
excluded from the analysis, the correlations spi- regional differences exist in the practices and
ral to .55 (p < .10, two-tailed) between cash values of societal Uncertainty Avoidance
holdings and Uncertainty Avoidance practices as measured by the GLOBE scales. There
scores, and −.80 (p < .01, two-tailed) between exist significant cross-regional differences in
cash holdings and Uncertainty Avoidance val- Uncertainty Avoidance societal practices
ues scores. (F9, 51 = 11.75, p < .01) as well as in societal
Correlations with the number of noncash values (F9, 51 = 17.36, p < .01). Further analysis
transactions per capita are consistent with the showed that regional differences account
results for cash holdings. As cash holdings for 68% of the variation in GLOBE societal
decrease, we may expect the number of noncash Uncertainty Avoidance practices scores,
transactions to go up. As expected on the basis of whereas 32% represented idiosyncratic societal
cash-holding correlations, the number of non- differences. With respect to GLOBE societal
cash transactions varies as an inverse function of Uncertainty Avoidance values, 75% of the vari-
the GLOBE Uncertainty Avoidance practices ation was explained by regional differences,
scores (r = −.78, p < .01) and as a direct function whereas 25% represented unique societal differ-
of GLOBE Uncertainty Avoidance values scores ences. In other words, there is substantial within-
(r = .43, p > .10, two-tailed). Among the devel- region commonality among societies within
oped nations (with the exception of Japan and clusters with respect to Uncertainty Avoidance
Italy), a higher GLOBE Uncertainty Avoidance as well as smaller nontrivial differences.
values orientation is associated with lower cash Table 19.18 shows the mean and standard
holdings and higher noncash transactions. deviation of societal GLOBE Uncertainty
Avoidance practices and values scores in each
of the regions. It can be seen that Southern Asia,
UNCERTAINTY Middle East, Sub-Saharan Africa, Eastern
AVOIDANCE REGIONAL CLUSTERS Europe, and Latin America scored relatively
high on GLOBE Uncertainty Avoidance values
As discussed in Chapter 8 by Hanges and measures, whereas Northern European and
Dickson, there are significant differences in Anglo societies scored relatively low. Nordic
values and practices for every cultural dimension Europe and Germanic Europe scored highest
among the 61 societies (i.e., recall that we use in Uncertainty Avoidance societal practices.
the term societal cultures rather than nations). Eastern Europe, Latin America, and Middle
Societies in the GLOBE sample were clustered East cultures, on the other hand, scored less than
19-House.qxd 3/19/2004 5:34 PM Page 636
the midpoint of 4.0 on GLOBE Uncertainty Table 19.19 provides grand mean scores for
Avoidance practices. GLOBE societal Uncertainty Avoidance prac-
tices and values across each of the climatic
Uncertainty Avoidance categories. The GLOBE Uncertainty Avoidance
and Climatic Differences practices scores for tropical wet and dry
(savanna), subtropical humid, and subtropical
The GLOBE sample of societies is grouped
wet and dry (Mediterranean) climate regions
into seven categories of climates (see Chapter 10
by Gupta & Hanges). Analysis of Variance average to less than the midpoint of the GLOBE
tests suggest significant climatic differences for scale (i.e., 4.0). The reported practices are
Uncertainty Avoidance societal practices (F6, distinctively strong in the maritime climate
54 = 4.41, p < .01) as well as for societal values region (marine west coast). In contrast, reported
(F6, 54 = 10.72, p < .01). Further analysis Uncertainty Avoidance values are distinctively
(not shown) showed that climatic differences strong in tropical wet and dry climate; the
account for 33% of the variation in societal average of GLOBE Uncertainty Avoidance
Uncertainty Avoidance practices scores, whereas values scores is less than the midpoint of 4.0
67% represented idiosyncratic societal differ- for the temperate marine west coast climate
ences. With respect to GLOBE societal Uncer- region. To summarize, temperate climates
tainty Avoidance values scores, 54% of the (especially maritime, but also continental) tend
variation was explained by climatic differences, to have strong Uncertainty Avoidance practices,
whereas 46% represented unique societal differ- but do not value Uncertainty Avoidance. The
ences. In other words, climatic differences have a opposite is true for the nontemperate wet and
considerably stronger effect on reported societal dry climates, especially in the tropics, but also
Uncertainty Avoidance values than on practices. extending even to the subtropics.
Standard Standard
Geographic Region N Mean Deviation Mean Deviation
Standard Standard
Climate N Mean Deviation Mean Deviation
colleagues’ data sample, “requesting written consistent with our hypothesized relationships,
documentation” had the strongest correlation although further research is needed to verify this
with GLOBE societal Uncertainty Avoidance as association better.
reflected by current practices in the GLOBE
ranking of countries (r = .54, p < .05, rho = .44,
p > .10). Curiously, checking for criminal records UNCERTAINTY AVOIDANCE
yielded a moderate negative correlation with AND FORMALIZATION
Uncertainty Avoidance current practices
(r = −.34, p > .10, rho = −.50, p < .10). It may be In his work on the influence of national cultures
that criminal backgrounds are more rare and, on corporate cultures, Trompenaars (1994) asked
therefore, less of an issue in highly uncertainty- respondents to choose between four possible
avoiding cultures. descriptions of their company. The descriptions
were keyed, respectively, to reflect the power pri-
ority of the “family” type organizational culture,
Uncertainty Avoidance
the role dominance of the “Eiffel Tower” culture,
and Championing Strategies the task orientation of the “guided missile” cul-
Shane (1994) hypothesized and observed ture, and the person orientation of the “incubator”
a negative relationship between Hofstede’s culture. He obtained results from 12 countries,
(1980) UAI and a preference for autonomy mapped along two dimensions: degree of central-
(Spearman’s rho = −.32, p < .05, n = 24). The ization and degree of formalization. As the time of
GLOBE values rankings (of countries common data collection in this study coincides with
to both studies) supported this relationship GLOBE’s, the GLOBE Uncertainty Avoidance
(Spearman’s rho = − .31, n = 18, p < .10) to practices scores can be rank-correlated with the
much the same extent. Predictably, the GLOBE ranking of countries on the two organizational
current practices rankings correlated positively culture dimensions. Across the 11 countries com-
(r = .24, p > .10) with preference for autonomy. mon to both datasets, GLOBE Uncertainty
Although the comparison of Shane and GLOBE Avoidance practices scores rank-correlated .40
scores was not statistically significant, the rela- (p > .10) with the degree of formalization and −.26
tionships are worth noting. (p > .10) with the degree of centralization.22 Cyert
Another innovation championing strategy and March (1963) posited that organizations adopt
that we hypothesized as relevant to Uncertainty procedures that minimize the need for predicting
Avoidance is monitoring. It has been proposed uncertain events in the future. We theorize that
that some degree of control exercised over the in countries with high levels of uncertainty
innovation process will help a team achieve its avoidance, we may expect to find a high degree of
goals (e.g., Sathe, 1988). Among the monitoring formalization and decentralization.
processes advocated is the approval of funds on
a periodical basis in installments rather than in The Relationship Between
entirety at one time. Monitoring can be seen as
Societal and Organizational Practices
an attempt to reduce uncertainty of outcomes at
least in part by minimizing the losses on bad
and Societal and Organizational Values
investments. Shane (1994) did not hypothesize a As indicated in Chapter 2 (House & Javidan,
relationship between UAI scores and monitoring.20 this volume), the GLOBE theoretical model
Had he done so, he would have found a rather postulates that societal practices and values
large rank correlation of .47 (p < .05) with affect organizational practices and values. Two
Hofstede’s UAI,21 consistent with the notion hierarchical linear models (HLMs) were con-
that people in high uncertainty avoidance ducted to test these hypotheses for organiza-
cultures prefer monitoring as an innovation tional Uncertainty Avoidance practices and
championing strategy. Shane’s (1994) rankings values. We tested the GLOBE hypothesis
on monitoring correlate .38 (p > .10) with the regarding the effect of societal culture on orga-
GLOBE values rankings and −.43 (p < .10) nizational culture by conducting HLM analyses
with current practices rankings. These are in which societal Uncertainty Avoidance
19-House.qxd 3/19/2004 5:34 PM Page 641
Table 19.20 HLM Regression of GLOBE Societal Culture Dimensions on Organizational Uncertainty
Avoidance Practices
Change in
Organizational Organizational Predictor
Uncertainty Uncertainty Significant
Predictors Avoidance R2 Avoidance R2 .05 level
Table 19.21 HLM Regression of GLOBE Societal Culture Dimensions on Organizational Uncertainty
Avoidance Values
Change in
Organizational Organizational Predictor
Uncertainty Uncertainty Significant
Predictors Avoidance R2 Avoidance R2 .05 level
predicted organizational Uncertainty Avoidance. and practices affect organizational cultural values
As shown in Table 19.20 and Table 19.21, soci- and practices.23
etal Uncertainty Avoidance practices did have a
significant and strong positive relationship with HLM Analysis: Uncertainty
organizational Uncertainty Avoidance practices
Avoidance and Leadership CLTs
(p < .05). We also found a strong relationship
between societal Uncertainty Avoidance values This chapter presents relationships between
and organizational Uncertainty Avoidance val- one culture dimension (Uncertainty Avoidance)
ues (p < .01). Therefore, this analysis supports a and the six culturally endorsed leadership
principal proposition in the GLOBE theoretical dimensions. Competitive tests of all culture
model (i.e., Proposition 3, Figure 2.1, Chapter 2, dimensions and GLOBE dimensions are
by House and Javidan): societal cultural values presented in Chapter 21 by Dorfman and
19-House.qxd 3/19/2004 5:34 PM Page 642
Figure 19.1 Uncertainty Avoidance Cultural Values as Drivers for CLT Leadership Dimensions
* Only statistically significant relationships are shown (see Table 21.10, Chapter 21 by Dorfman et al.). Relationship
significant at p < .05. The most important leadership CLT relationships are bolded (i.e., relationship is significant at both
society and organizational levels of analyses or highest HLM coefficient within each level of analysis).
O = Organizational level
S = Societal level
The societal Uncertainty Avoidance cultural majority of the explained variance (94.5%)
values scale was negatively related (p < .05) to was associated with forces operating at the
this CLT dimension. Charismatic/Value-Based societal level of analysis.
leadership is less likely to be a part of the shared
leadership belief system in societies reported to The organizational Uncertainty Avoidance
espouse Uncertainty Avoidance values. cultural practices scale was negatively related
(p < .05) to this CLT dimension, as was the
• Team-Oriented leadership. Uncertainty organizational Uncertainty Avoidance cultural
Avoidance practices and values were signifi- values scale (p < .01). Participative leadership is
cantly related to this CLT dimension and less likely to be a part of the shared leadership
explained a total of 25.2% of the organizational belief system in organizations reported to have
and societal variance for this dimension. Uncertainty Avoidance practices and it is less
Approximately 15.6% of this explained vari- likely to be a part of the shared leadership belief
ance was associated with forces operating at the system in organizations reported to espouse
organizational level of analysis. The remaining Uncertainty Avoidance values.
portion of the explained variance (84.4%) was The societal Uncertainty Avoidance cultural
associated with forces operating at the societal values scale was negatively related (p < .01) to
level of analysis. this CLT dimension. Participative leadership is
less likely to be a part of the shared leadership
The organizational Uncertainty Avoidance belief system in societies reported to espouse
cultural practices scale was negatively related Uncertainty Avoidance values.
(p < .05) to this CLT dimension, whereas the • Humane-Oriented leadership. Uncertainty
organizational Uncertainty Avoidance cultural Avoidance cultural practices and values were
values scale was positively related (p < .01) to significantly related to this CLT dimension and
this CLT dimension. Team-Oriented leadership explained a total of 20.9% of the organizational
is less likely to be a part of the shared leadership and societal variance for this dimension.
belief system in organizations reported to have Approximately 32.2% of this explained vari-
Uncertainty Avoidance practices, but it is more ance was associated with forces operating at the
likely to be a part of the shared leadership belief organizational level of analysis. The remaining
system in organizations reported to espouse portion of the explained variance (67.8%) was
Uncertainty Avoidance values. associated with forces operating at the societal
The societal Uncertainty Avoidance cultural level of analysis.
practices scale was negatively related (p < .05)
to this CLT dimension, whereas the societal The organizational Uncertainty Avoidance
Uncertainty Avoidance cultural values scale cultural practices scale was negatively related
was positively related (p < .05) to this CLT (p < .05) to this CLT dimension, whereas the
dimension. Team-Oriented leadership is less organizational Uncertainty Avoidance cultural
likely to be a part of the shared leadership belief values scale was positively related (p < .01) to
system in societies reported to have Uncertainty this CLT dimension. Humane-Oriented leader-
Avoidance practices, but it is more likely in ship is less likely to be a part of the shared lead-
societies reported to espouse Uncertainty ership belief system in organizations reported to
Avoidance values. have Uncertainty Avoidance practices, but it is
more likely to be a part of the shared leadership
• Participative leadership. Uncertainty belief system in organizations reported to
Avoidance cultural practices and values were espouse Uncertainty Avoidance values.
significantly related to this CLT dimension The societal Uncertainty Avoidance cultural
and explained a total of 33.8% of the organi- values scale was positively related (p < .01) to
zational and societal variance for this dimen- this CLT dimension. Humane-Oriented leader-
sion. Approximately 5.5% of this explained ship is more likely to be a part of the shared
variance was associated with forces operating leadership belief system in societies reported to
at the organizational level of analysis. The espouse Uncertainty Avoidance values.
19-House.qxd 3/19/2004 5:34 PM Page 644
Uncertainty Avoidance. In this chapter we have Finally, uncertainty avoidance has implications
demonstrated that organizations reflect the cul- for which leadership styles are seen as most
ture (values, but not practices) in the society in effective in society. A series of hierarchical lin-
which they are embedded. The notion that orga- ear modeling procedures performed to examine
nizational values are reflected in the society is the influence of Uncertainty Avoidance on
consistent with Hofstede’s (2001) thesis regard- implicit leadership theories at the individual,
ing uncertainty avoidance as a value. A negative organizational, and societal levels of analysis
association between GLOBE Uncertainty showed that cultural dimensions differently
Avoidance practices and values at the societal influence each of six leadership styles: Higher
level suggests that the desire for order and struc- Uncertainty Avoidance values were associated
ture may be strongly influenced by the extent of with higher Team-Oriented, Humane-Oriented,
order and structure present in society. Notably, and Self-Protective leadership CLT dimen-
whereas the industrialized nations ranked high sions; and higher Uncertainty Avoidance val-
in GLOBE Uncertainty Avoidance practices ues were associated with lower Participative
and low in values, the reverse is the case with and Charismatic/Value- Based leadership CLT
the developing nations. GLOBE Uncertainty dimensions.
Avoidance scores are highly correlated among In conclusion, the GLOBE study offers
the three industries from which data were further insights into the construct of Uncertainty
obtained: food processing, finance, and tele- Avoidance and its relationship to societal, orga-
communications industries. This indicates a nizational, and leadership variables. The distinc-
high level of consistency among industries tion between Uncertainty Avoidance values
within a societal culture. and practices scores is brought into sharp focus
A review of the prior research on this through a negative correlation between the
cultural dimension in relation to GLOBE data scores. Uncertainty Avoidance, as measured
described in this chapter suggests that uncer- in the GLOBE study, also has conceptual and
tainty avoidance may be related to several soci- empirical links to Power Distance, Individualism,
etal, economic, and organizational variables. and other dimensions of national culture.
Specifically, Uncertainty Avoidance practices
and values, as measured by the GLOBE scales,
were found to be related to innovation, percep- ENDNOTES
tions of risk, per capita cash holdings, and
growth, among other variables. The data suggest 1. The three personality dimensions for the EPQ
that societies with high scores for GLOBE Uncer- are neuroticism versus emotional stability, extraver-
tainty Avoidance practices tend to have a higher sion versus introversion, and psychoticism versus ego
level of economic prosperity and enjoy more control.
civil liberties, a higher level of competitiveness 2. Hofstede (1984) also distinguished the intoler-
in the global market, greater security, higher life ance of ambiguity as an attitudinal component of
expectancy, and greater general satisfaction. the authoritarian personality from the intolerance of
Higher Uncertainty Avoidance practices were ambiguity as a cultural norm. The first is presumably
also related to more emphasis on basic science an aspect of power distance, the second of uncertainty
and research. avoidance.
At the organizational level, uncertainty avoid- 3. The Hofstede (1980) model does not address
ance has implications for organizational variables the individual level of analysis. However, Hofstede,
such as selection practices. The above review of Bond, and Luk (1993) suggest that some aspects of
the literature suggests that uncertainty-avoiding cultural values could be effectively studied at differ-
societies may have a relatively high degree of ing levels of analysis, although these manifestations
formalization and decentralization in organiza- will not be the same at the various levels of analysis.
tional structure. It is plausible that appropriate 4. On the basis of research conducted into the
uncertainty avoidance practices make it possible 1980s and 1990s using value survey modules (VSM)
for decentralization of operations rather than developed by Hofstede, a new set of measures was
management by personal control. developed (VSM 94). The uncertainty avoidance
19-House.qxd 3/19/2004 5:34 PM Page 646
measure was amended such that the employment and persuasion based on financial measures could not
longevity question was removed, replaced by two addi- be tested.
tional questions: “One can be a good manager without 16. The four companies were a life and property
having precise answers to most questions that subordi- insurance company, a financial services firm, a petro-
nates may raise about their work” and “Competition chemical firm, and a consumer electronics firm.
between employees usually does more harm than 17. Our response bias correction procedure iden-
good.” Unfortunately, no psychometric properties were tified response bias in some countries for this scale.
made available justifying the selection and inclusion of We recomputed the predicted response bias corrected
these items as measures of uncertainty avoidance. scale score for each country:
5. Using the Hofstede dimensions as a paradigm Practices: Indonesia, 3.92 (moves from band B to
is practiced whenever the dimensions, or the Hofstede band C); Morocco, 3.95 (no change in band); Qatar,
country rankings on which the dimensions are based, 4.26 (moves from band C to band B); and Taiwan,
are used as a theoretical framework external to their 4.04 (moves from band B to band C).
original setting. For a complete discussion of this procedure and
6. Twenty-eight replications examined matched all response bias corrected scores, see Appendix B.
populations with paired sets of a minimum of two 18. Our response bias correction procedure iden-
separate populations; 16 studies examined single tified response bias in some countries for this scale.
populations; 13 studies made use of research designs We recomputed the predicted response bias corrected
and measures different from Hofstede (1980). scale score for each country:
7. The Hoppe study had 18 countries in com- Values: France, 4.65 (moves from band C to
mon with the Hofstede sample. band B) and Morocco, 5.77 (no change in band).
8. The average Salzburg alumni score was For a complete discussion of this procedure and
32.9 points lower on UAI than the Hofstede sample all response bias corrected scores, see Appendix B.
of IBM employees. 19. Controlling for age, Hofstede reported Japan
9. To measure uncertainty, the authors modified to be highest on UAI.
measures developed by Gudykunst and Nishida (1986), 20. In a later study involving 30 countries, Shane
to measure anxiety they used measures developed by et al. (1995) again missed an opportunity to investi-
Stephan and Stephan (1985), and to assess perceived gate this influential variable in connection with
effectiveness of communication the authors developed uncertainty avoidance.
original measures (see Gudykunst, & Chua, Nishida, 21. This correlation was obtained by ranking the
1986). country mean scores on monitoring.
10. There was also a significant multivariate 22. With the sample size of 11 nations, however,
main effect for culture, and significant univariate none of the correlations are statistically significant,
effects for anxiety, uncertainty, and effectiveness for but the relationship is worth noting.
relationship and culture. 23. As reported in Chapter 20 by Brodbeck,
11. The authors do not provide correlation values; Hanges, Dickson, Gupta, and Dorfman, we found that
only the p values associated with each relationship all the cultural dimensions of organizational cultural
and the model fit estimates. values and practices significantly differed across
12. These years were carefully selected to closely societies. Although important, this prior analysis
match the time period of the data collection of did not identify the particular aspect of societal
Hofstede (1980). differences that was related to organizational culture.
13. This control was achieved by taking national In the present chapter, we found that societal and
rates of innovation as per capita numbers of trade- organizational Uncertainty Avoidance practices were
marks granted in the United States and world markets not significantly related (p > .05). In contrast, we
to nationals from the 33 countries under study. found strong results for societal and organizational
14. Shane did not provide correlations of the Uncertainty Avoidance values (R2 Total = 23.1%, R2
independent and dependent variable, thus a more Societal = 84.5%, p < .01). As discussed in Chapter 11
detailed discussion is precluded. by Hanges, Dickson, and Sipe, the R2 Total considers
15. Reliable measures for the championing strat- all levels of analysis (i.e., individual, organizational,
egy of persuasion could not be created for this study, and societal), whereas the R2 Societal isolates the
thus the relationship between uncertainty avoidance societal-level portion of the dependent variable and
19-House.qxd 3/19/2004 5:34 PM Page 647
indicates the percentage of variance accounted for by information environment. Journal of Occupational
the predictor at only this level. Whereas we have pri- Psychology, 58, 67–97.
marily taken the conservative approach and reported Ashkanini, M. G. A. (1984). A cross-cultural per-
the R2 Total in GLOBE, several scholars suggest that spective on work related values. Unpublished
R2 Societal provides a more accurate description of doctoral dissertation, International University,
aggregated relationships. For further discussion, see San Diego.
the paper by Lance and James (1999). Aspinwall, L., & Staudinger, U. (2003). A psychology
24. Results between the single HLM and multiple of human strengths: Fundamental questions
HLM tests will likely differ somewhat. The differ- and future directions for a positive psychology.
ences between the results of the multiple HLMs and Washington, DC: American Psychological
single HLMs are conceptually similar to the differ- Association.
ences between a multiple regression analysis and a Atuahene-Gima, K., & Li, H. (2002). When does
correlation coefficient. Table 21.10 in Chapter 21 by trust matter? Antecedents and contingent effects
Dorfman et al. presents both single and multiple of supervisee trust on performance in selling
HLM coefficients. In addition, the relationships for new products in China and the United States.
all culture dimension values are summarized in Journal of Marketing, 66(3), 61–81.
Chapter 3 (Javidan & House). Babrow, A. S. (1992). Communication and problem-
atic integration: Understanding diverging proba-
bility and value, ambiguity, ambivalence, and
REFERENCES impossibility. Communication Theory, 2, 95–130.
Babrow, A. S., Hines, S. C., & Kasch, C. R. (2000).
Ackelsberg, R., & Arlow, P. (1985). Small businesses Managing uncertainty in illness explanation: An
do plan and it pays off. Long Range Planning, application of problematic integration theory. In
18(5), 61–67. B. Whaley (Ed.), Explaining illness: Research,
Adler, N. J. (1991). International dimensions of orga- theory, and strategies (pp. 41–67). Hillsdale, NJ:
nizational behavior. Boston: Plus Kent. Lawrence Erlbaum.
Adorno, T. W., Frenkel-Brunswick, E., Levinson, Bailey, J. B., Chen, C. C., & Dou, S.-G. (1997).
D. J., & Sanford, R. N. (1950). The authoritar- Conceptions of self and performance-related
ian personality. New York: Harper. feedback in the U.S., Japan, and China. Journal
Arrindell, W. A., Hatzichristou, C., Wensink, J., of International Business Studies, 28, 605–625.
Rosenberg, E., van Twillert, B., Stedema, J., & Basabe, N., Paez, D., Valencia, J., Gonzalez, J. L.,
Meiijer, D. (1997). Dimensions of national cul- Rime, B., & Diener, E. (2002). Cultural dimen-
ture as predictors of cross-national differences sions, socioeconomic development, climate, and
in subjective well-being. Personality and emotional hedonic level. Cognition and
Individual Differences, 23(1), 37–53. Emotion, 16(1), 103–125.
Ashford, S. J. (1986). Feedback-seeking in individual Baskerville, R. F. (2003). Hofstede never studied
adaptation: A resource perspective. Academy of culture. Accounting, Organizations and Society,
Management Journal, 29, 465–487. 28, 1–14.
Ashford, S. J. (1988). Individual strategies for coping Beatty, C. A., & Gordon, J. R. M. (1988). Barriers to
with stress during organizational transitions. the implementation of CAD/CAM systems.
Journal of Applied Behavioral Science, 24, 19–36. Sloan Management Review, 29(4), 25–33.
Ashford, S. J., Blatt, R., & VandeWalle, D. (in press). Bennett, N., Herold, D. M., & Ashford, S. J. (1990).
Reflections on the looking glass: A review of The effects of tolerance for ambiguity on feed-
research on feedback-seeking behavior in orga- back-seeking behavior. Journal of Occupational
nizations. Journal of Management. Psychology, 63, 343–348
Ashford, S. J., & Cummings, L. L. (1983). Feedback Berger, C. R., & Bradac, J. J. (1982). Language and
as an individual resource: Personal strategies of social knowledge: Uncertainty in interpersonal
creating information. Organizational Behavior relations. London: Edward Arnold.
and Human Performance, 32, 370–398. Berger, C. R., & Calabrese, R. J. (1975). Some
Ashford, S. J., & Cummings, L. L. (1985). Proactive explorations in initial interaction and beyond:
feedback seeking: The instrumental use of the Toward a developmental theory of interpersonal
19-House.qxd 3/19/2004 5:34 PM Page 648
communication. Human Communication Clark, R. W. (1984). Einstein: The life and times.
Research, 1, 99–112. New York: Avon Books.
Bochner, S. (1965). Defining intolerance of ambigu- Cyert, R. M., & March, J. G. (1963). A behavioral
ity. Psychological Record, 15, 393–400. theory of the firm. Englewood Cliffs, NJ:
Bontempo, R. N., Bottom, W. P., & Weber, E. U. Prentice Hall.
(1997). Cross-cultural differences in risk percep- Dawson, J. W. (1999, June). Godel and the limits of
tion: A model-based approach. Risk Analysis, logic. Scientific American, pp. 76–81.
17(4), 479–488. de Bettignies, H. C., & Evans, P. L. (1977). The
Bracket, J. S., Keats, B. W., & Pearson, J. N. (1988). cultural dimension of top executives’ careers:
Planning and financial performance among A comparative analysis. In T. D. Weinshall (Ed.),
small firms in a growth industry. Strategic Culture and management. Harmondsworth, UK:
Management Journal, 9, 591–603. Penguin.
Bradac, J. J. (2001). Theory comparison: Uncertainty Demerath, L. (1993). Knowledge-based affect. Social
reduction, problematic integration, uncertainty Psychology Quarterly, 56, 136–147.
management, and other curious constructs. Dessler, G. (1993). Value-based hiring builds com-
Journal of Communication, 51(3), 456–476. mitment. Workforce, 72(11), 98–101.
Brett, J. M., Feldman, D. C., & Weingart, L. R. Diener, E. (2000). Subjective well-being: The science
(1990). Feedback-seeking behavior of new hires of happiness and a proposal for a national index.
and job changers. Journal of Management, 16, American Psychologist, 55(1), 34–43.
737–149. Diener, E., Diener, M., & Diener, C. (1995). Factors
Brown, S. P., Ganesan, S., & Challagalla, G. (2001). predicting the subjective well-being of nations.
Self-efficacy as a moderator of information- Journal of Personality and Social Psychology,
seeking effectiveness. Journal of Applied 69(5), 851–864.
Psychology, 86, 1043–1051. Dimaggio, P. J., & Powell, W. (1983). The iron cage
Budner, S. (1962). Intolerance of ambiguity as a revisited: Institutional isomorphism and collec-
personality variable. Journal of Personality, 30, tive rationality in organizational fields. American
29–59. Sociological Review, 48, 147–160.
Burgelman, R. (1983). A process model of internal d’Iribarne, P. (1997). The usefulness of an ethno-
corporate venturing in the major diversified graphic approach to the international compari-
firm. Administrative Science Quarterly, son of organizations. International Studies of
28, 223–244. Management and Organization, 26(4), 30–47.
Callister, R. R., Kramer, M. W., & Turban, D. B. Doney, P. M., Cannon, J. P., & Mullen, M. R. (1998).
(1999). Feedback seeking following career tran- Understanding the influence of national culture
sitions. Academy of Management Journal, 42, on the development of trust. Academy of
429–438. Management Review, 23(3), 601–620.
Chakrabarti, A. (1974). The role of champion in Dunphy, D., & Shi, J. (1986). A comparison of enter-
product innovation. California Management prise management in Japan and the People’s
Review, 17(2), 58–62. Republic of China (Working Paper 86/016).
Chew, I. K. H., & Putti, J. (1993). A comparative Australian Graduate School of Management,
study of work-related values of Singaporean The University of Sydney and The University of
and Japanese managers in Singapore (Working New South Wales.
Paper). Nanyang Technological University, Earley, P. C. (1997). Face, harmony and social
Singapore. structure. New York: Oxford University
Chow, C. W., Shields, M. D., & Chan, Y. K. Press.
(1991). The effects of management controls Eysenck, H. J., & Eysenck, M. (1975). Manual of the
and national culture on manufacturing perfor- Eysenck Personality Questionnaire. London:
mance. Accounting, Organizations and Society, Hodder and Stoughton.
209–226. Fernandez, D. R., Carlson, D. S., Stepina, L. P., &
Clampitt, P. G., & de Koch, R. J. (2002). The essence Nicholson, J. D. (1997). Hofstede’s country
of leadership (2nd ed.). New York: M. E. classification 25 years later. Journal of Social
Sharpe. Psychology, 137(1), 43–54.
19-House.qxd 3/19/2004 5:34 PM Page 649
organizations across nations. Thousand Oaks, Jeanquart-Barone, S., & Peluchette, J. L. V. (1999).
CA: Sage. Examining the impact of the cultural dimension
Hofstede, G., Bond, M. H., and Luk, C.-L. (1993). of uncertainty avoidance on staffing decisions:
Individual perceptions of organizational cul- A look at U.S. and German firms. Cross
tures: A methodological treatise on levels of Cultural Management, 6(3), 3–12.
analysis. Organization Studies, 14(4), 483–503. Judge, T., & Bono, J. E. (2001). Relationship of core
Hoppe, M. H. (1990). A comparative study of country self-evaluation traits—self-esteem, generalized
elites: International differences in work-related self-efficacy, locus of control, and emotional
values and learning and their implications stability—with job satisfaction and job perfor-
for management training and development. mance: A meta-analysis. Journal of Applied
Unpublished doctoral thesis, University of Psychology, 86, 80–92.
North Carolina at Chapel Hill. Kahn, R. L., Wolfe, D. M., Quinn, R. P., Snoek, J. D., &
Howell, J., & Higgins, C. (1990). Champions of tech- Rosenthal, R. A. (1964). Organizational stress:
nological innovation. Administrative Science Studies in role conflict and role ambiguity.
Quarterly, 35, 317–341. New York: John Wiley.
Howell, J., & Higgins, C. (1991). Champions of Kale, S. H., & McIntyre, R. P. (1991). Distribution
change: Identifying, understanding, and sup- channel relationships in diverse cultures.
porting champions’ technological innovations. International Marketing Review , 8(3), 31–45.
Organizational Dynamics , 10(1), 40–55. Kanter, R. (1982). The change masters. New York:
Humphrey, D. B., Pulley, L. B., & Vesala, J. M. Simon and Schuster.
(1996). Cash, paper, and electronic payments: A Kanter, R. (1988). When a thousand flowers bloom:
cross-country analysis. Journal of Money, Structural, collective, and social conditions for
Credit, and Banking, 28(4), 914–939. innovation in organizations. In L. Cummings & B.
Huo, P. Y., & Randall, D. M. (1991). Exploring sub- Staw (Eds.), Research in organizational behavior
cultural differences in Hofstede’s value survey: (Vol. 10, pp. 169–211). Greenwich, CT: JAI.
The case of the Chinese. Asia Pacific Journal of Kelly, J. R., & Barsade, S. G. (2001). Mood and emo-
Management, 8(2), 159–173. tions in small groups and work teams.
Ilgen, D. R., Fisher, C. D., & Taylor, M. S. (1979). Organizational Behavior & Human Decision
Consequences of individual feedback on behav- Processes, 86, 99–130.
ior in organizations. Journal of Applied Kim, U., Triandis, H. C., Kagitçibasi, C., Choi, S.-C., &
Psychology, 64, 349–371. Yoon, G. (1994). Introduction. In U. Kim, H. C.
Inglehart, R. (1997). Modernization and postmod- Triandis, C. Kagitçibasi, S.-C. Choi, & G. Yoon
ernization: Cultural, economic and political (Eds.), Individualism and collectivism: Theory,
change in 43 societies. Princeton, NJ: Princeton method, and applications (pp. 1-16). Thousand
University Press. Oaks, CA: Sage.
Inglehart, R., & Baker, W. E. (2000). Modernization, Kluger, A. N., & DeNisi, A. (1996). The effects of
cultural change, and the persistence of tradi- feedback interventions on performance: A histor-
tional values, American Sociological Review, ical review, a meta-analysis, and a preliminary
65(1), 19–51. feedback intervention theory. Psychological
Inglehart, R., Basanez, M., & Moreno, A. (1998). Bulletin, 119, 254–284.
Human values and beliefs: A cross-cultural Kohn, M. L. (1971). Bureaucratic man: A portrait and
sourcebook. Ann Arbor: University of Michigan an interpretation. American Sociological
Press. Review, 36, 461–474.
Inkson, J. H. K., Schwitter, J. P., Pheysey, D. C., & Kreitler, S., Maguen, T., & Kreitler, H. (1975). The
Hickson, D. J. (1970). A comparison of organi- three faces of intolerance of ambiguity. Archiv
zation structure and managerial roles: Ohio, Für Psychologie, 127, 238–250.
U.S.A., and the Midlands, UK. Journal of Lance, C. E., & James, L. R. (1999). A proportional
Management Studies, 7, 347–363. variance accounted for index for some
Inlis, I. R. (2000). The central role of uncertainty cross-level and person-situation research
reduction in determining behaviour. Behaviour, designs. Organizational Research Methods, 2,
137, 1567–1599. 395–418.
19-House.qxd 3/19/2004 5:34 PM Page 651
Lebas, M., & Weigenstein, J. (1986). Management McNally, G. M. (1980). Responsibility accounting
control: The roles of rules, markets, and culture. and organizational control: Some perspectives
Journal of Management Studies, 23, 259–273. and prospects. Journal of Business Finance and
Leung, T., & Yeung, L. L. (1995). Negotiation in the Accounting , 7(2), 165–181.
People’s Republic of China: Results of a survey Millendorfer, J. (1976). Mechanisms of socio-
of small businesses in Hong Kong. Journal of psychological development. Vienna: Studien-
Small Business Management, 33(1), 70–77. gruppe für Internationale Analysen.
Le Vine, R. A. (1973). Culture, behavior and person- Miller, V. D., & Jablin, F. M. (1991). Information
ality: An introduction to the comparative study seeking during organizational entry: Influences,
of psychosocial adaptation. Chicago: AVC. tactics, and a model of the process. Academy of
Lowe, S. (1996). Hermes revisited: A replication of Management Review, 16, 92–120.
Hofstede’s study in Hong Kong and the UK. Morrison, E. W. (1993). Newcomer information
Asia Pacific Business Review , 2(3), 101–119. seeking: Exploring types, modes, sources, and
Lumpkin, G. T., Shrader, R. C., & Hills, G. E. (1998). outcomes. Academy of Management Journal,
Does formal business planning enhance the 36, 557–589.
performance of new ventures? In P. S. Reynolds, Morrison, E. W. (2002). Information seeking with
W. D. Bygrave, N. M. Carter, S. Manigart, C. M. organizations. Human Communication Research,
Mason, G. D. Meyer, & K. G. Shaver (Eds.), 28, 229–242.
Frontiers of entrepreneurship research (pp. 180- Nakane, C. (1974). The social system reflected in
189). Wellesley, MA: Babson College Press. interpersonal communication. In J. Condon &
Lyle, M. A., Baird, L. S., Orns, B., & Kuratko, D. E. M. Saito (Eds.), Intercultural encounters with
(1995). Formalized planning in small business: Japan (124–131). Tokyo: Simul Press.
Increasing strategic choice. Journal of Small Nakata, C., & Sivakumar, K. (1996). National culture
Business Management, 33(1), 38–50. and new product development: An integrative
Lynn, R., & Hampson, S. L. (1975). National differ- review. Journal of Marketing, 60, 61–72.
ence in extraversion and neuroticism. British Norton, R. (1975). Measurement of ambiguity
Journal of Social and Clinical Psychology, 14, tolerance. Journal of Personality Assessment,
223–240. 39, 607–619.
Lynn, R., & Hampson, S. L. (1977). Fluctuations in Pascale, R. T., & Athos, A. G. (1983). The art of Japanese
national levels of neuroticism and extraversion, management. New York: Penguin.
1935-1970. British Journal of Social and Pelto, P. J. (1968, April). The difference between “tight”
Clinical Psychology, 16, 131–137. and “loose” societies. Trans-Action, 37–40.
Lynn, R., & Martin, T. (1995). National differences Perrow, C. (1972). Complex organizations: A critical
for 39 nations in extraversion, neuroticism, psy- essay. Glenview, IL: Scott, Foresman.
choticism and economic, demographic and other Peterson, M. F., Smith, P. B. Akande, A., Ayestaran,
correlates. Personality and Individual Differences, S., Bochner, S., Callan, V., Cho, N., et al.
19, 403–406. (1995). Role conflict, ambiguity, and overload:
Lytle, A. L., Brett, J. M., Barsness, Z. I., Tinsley, C. H., A 21-nation study. Academy of Management
& Janssens, M. (1995). A paradigm for con- Journal, 38(2), 429–452.
firmatory cross-cultural research in organization Pooyam, A. (1984). Acculturation, acculturative
behavior. In Research in organizational behav- stress and their relationships to both work and
ior (Vol. 17, pp. 167–214). Greenwich, CT: JAI. non-work outcomes. Unpublished doctoral dis-
Maldonado, E. M. (1983). Mexican-American partici- sertation, Management Science, University of
pation in law enforcement in California: Law Texas at Dallas.
enforcement: the impact of cultural related Punnett, B. J., & Withaney, S. (1988). Testing the
value differences. Unpublished doctoral thesis, validity of Hofstede’s value indices. Unpublished
Claremont Graduate School, Claremont, CA. paper, University of Windsor, Ontario, Canada.
McClelland, D. C. (1961). The achieving society. Randall, D. M. (1993). Cross-cultural research
Princeton, NJ: Van Nostrand Reinhold. on organizational commitment: A review and
McGregor, D. (1960). The human side of enterprise. application of Hofstede’s Values Survey Module.
New York: McGraw-Hill. Journal of Business Research, 26(1), 91–110.
19-House.qxd 3/19/2004 5:34 PM Page 652
Reader’s Digest. (1970). A survey of Europe today. model. Journal of Cross-Cultural Psychology,
London: Author. 21(1), 109–118.
Reeder, J. A. (1987). When west meets east: Cultural Shane, S. (1993). Cultural influences on national rates
aspects of doing business in Asia. Business of innovation. Journal of Business Venturing, 8,
Horizons, 30, 69–74. 59–73.
Renn, R. W., & Fedor, D. B. (2001). Development and Shane, S. (1994). Cultural values and the championing
field test of a feedback seeking, self-efficacy, process. Entrepreneurship Theory and Practice,
and goal setting model of work performance. 18(4), 25–41.
Journal of Management, 27, 563–583. Shane, S. (1995). Uncertainty avoidance and the pref-
Robinson, W. S. (1950). Ecological correlations and erence for innovation championing roles. Journal
the behavior of individuals. American Sociological of International Business Studies, 26, 47–68.
Review, 15, 351–357. Shane, S., Venkataraman, S., & MacMillan, I. (1995).
Rossini, F. D. (1950). Chemical thermodynamics. Cultural differences in innovation champion-
New York: John Wiley. ing strategies. Journal of Management, 21(5),
Ryan, A. M., McFarland, L., Baron, H., & Page, R. 931–952.
(1999). An international look at section practices: Shapero, A., & Sokol, L. (1982). The social dimen-
Nation and culture as explanations for variability sions of entrepreneurship. In C. Kent, D. Sexton,
in practice. Personnel Psychology, 52, 359–391. & K. Vesper (Eds.), Encyclopedia of entrepre-
Ryans, C. C. (1997). Resources: Writing a business neurship (pp. 72–90). Englewood Cliffs, NJ:
plan. Journal of Small Business Management, Prentice Hall.
35(2), 95–98. Singh, J. P. (1990). Managerial culture and work-
Rydell, S., & Rosen, E. (1966). Measurement and related values in India. Organization Studies,
some correlates of need cognition. Psychological 11(1), 75–101.
Reports, 19, 139–165. Smith, P. B., & Schwartz, S. H. (1997). Values. In
Salter, S. B., & Niswander, F. (1995, Second Quarter). J. W. Berry, M. H. Segall, & K. Kagitçibasi
Cultural influence on the development of (Eds.), Handbook of cross-cultural psychology
accounting systems internationally: A test of (Vol. 3, 77–118). Needham, MA: Allyn &
Gray’s [1988] theory. Journal of International Bacon.
Business Studies, 379–397. Snyder, C. R., & Lopez, S. J. (2002). Handbook of
Sathe, V. (1988). Fostering entrepreneurship in the positive psychology. New York: Oxford
large diversified firm. Organizational Dynamics, University Press.
7(2), 20–32. Soeters, J. L., & Schreuder, H. (1988). The interaction
Schon, D. (1963). Champions for radical new inven- between national and organizational cultures in
tions. Harvard Business Review, 41, 77–86. accounting firms. Accounting, Organizations
Schramm-Nielsen, J. (2000). How to interpret uncer- and Society, 13, 75–85.
tainty avoidance scores: A comparison study of Søndergaard, M. (1994). Hofstede’s consequences: A
Danish and French firms. Cross Cultural Manage- study of reviews, citations and replications.
ment, 7(4), 3–11. Organization Studies, 15, 447–456.
Schwartz, S. H. (1999). A theory of cultural values Steel, P., & Ones, D. S. (2002). Personality and hap-
and some implications for work. Applied piness: A national-level analysis. Journal of
Psychology: An International Review, 48, Personality and Social Psychology, 83, 767–781.
23–47. Stephan, W., & Stephan, C. (1985). Intergroup anxi-
Schyns, P. (1998). Crossnational differences in hap- ety. Journal of Social Issues, 41, 157–166.
piness: Economic and cultural factors explored. Stohl, C. (1993). European managers’ interpreta-
Social Indicators Research, 43, 3–26. tions of participation: A semantic network
Seligman, M., & Csikszentmihalyi, M. (2000). Positive analysis. Human Communications Research,
psychology. American Psychologist, 55, 5–14. 20(1), 97–117.
Seligman, M. E. P. (1999). The president’s address. Sully de Luque, M. F., & Sommer, S. M. (2000). The
American Psychologist, 54, 559–562. impact of culture on feedback seeking behavior:
Shackleton, V. J., & Ali, A. H. (1990). Work-related An integrated model and propositions. Academy
values of managers: A test of the Hofstede of Management Review, 25, 829–849.
19-House.qxd 3/19/2004 5:34 PM Page 653
Tannenbaum, A. S., Kavcic, B., Rosner, M., Wanberg, C. R., & Kammeyer-Mueller, J. D. (2000).
Vianello, M., & Wieser, G. (1974). Hierarchy in Predictors and outcomes of proactivity in
organizations. San Francisco: Jossey-Bass. the socialization process. Journal of Applied
Tayeb, M. (1994). Organizations and national culture: Psychology, 85, 373–385.
Methodology considered. Organization Studies , Weber, E. U., & Hsee, C. (1998). Cross-cultural
15(3), 429–446. differences in risk perception, but cross-cultural
Thorndike, E. L. (1939). On the fallacy of imputing similarities in attitudes toward perceived risk.
the correlations found for groups to the indi- Management Science, 44, 1205–1218.
viduals or smaller groups composing them. Weber, E. U., & Milliman, R. A. (1997). Perceived
American Journal of Psychology, 3, 122–124. risk attitudes: Relating risk perception to risky
Triandis, H. C. (1989). The self and social behavior choice. Management Science, 43, 123–145.
in differing cultural contexts. Psychological Westwood, R. G., & Everett, J. E. (1987). Culture’s
Review, 96, 506–520. consequences: A methodology for comparative
Trompenaars, F. (1994). Riding the waves of culture. management studies in Southeast Asia. Asia
New York: Irwin. Pacific Journal of Management , 4(3), 187–202.
Turner, J. H. (1988). A theory of social interaction. Williamson, O. (1975). Markets and hierarchies.
Palo Alto, CA: Stanford University Press. New York: Free Press.
Ueno, S., & Sekaran, U. (1992). The influence of culture Witkin, H. A., & Berry, J. W. (1975). Psychological
on budget control practices in the USA and Japan: differentiation in cross-cultural perspective.
An empirical study. Journal of International Journal of Cross-Cultural Psychology, 6, 4–87.
Business Studies, 23(4), 659–674. Yates, J. F., Zhu, Y., Ronis, D. L., Wang, D. F.,
Van de Ven, A. (1986). Central problems in the man- Shinotsuka, H., & Masanao, T. (1989). Probability
agement of innovation. Management Science, judgment accuracy: China, Japan, and the
32(5), 590–607. United States. Organizational Behavior and
van Gunsteren, H. R. (1976). The quest for control: A Human Decision Processes, 43, 145–171.
critique of the rational-central-rule approach in Yeh, R. (1988, August). Values of American,
public affairs. London: Wiley. Japanese and Taiwanese managers in Taiwan: A
Venkataraman, S., Shane, S., McGrath, R., & test of Hofstede’ framework. In Best Paper
MacMillan, I. (1993). Some central tensions Proceedings at the annual meeting of the
in the management of corporate venturing. In Academy of Management Annual Meetings,
S. Birley & I. MacMillan (Eds.), Entrepreneurship Anaheim, CA.
research: Global perspectives (pp. 177–199). Zarzeski, M. T. (1996). Spontaneous harmonization
Amsterdam: Elsevier Science. effects of culture and market forces on account-
Verweij, J., Ester, P., & Nauta, R. (1997). ing disclosure practices. Accounting Horizons,
Secularization as an economic and cultural 10(1), 18–37.
phenomenon: A cross-national analysis. Journal Zhao, J. J. (2000). The Chinese approach to interna-
for the Scientific Study of Religion, 36(2), tional business negotiation. Journal of Business
309–324. Communication, 37, 209–237.
Walsh, T. (1995). Biosociology: An emerging para-
digm. Westport, CT: Praeger.
20-house.qxd 3/19/2004 5:06 PM Page 654
20
SOCIETAL CULTURE AND
INDUSTRIAL SECTOR INFLUENCES
ON ORGANIZATIONAL CULTURE
FELIX C. BRODBECK
PAUL J. HANGES
MARCUS W. DICKSON
VIPIN GUPTA
PETER W. DORFMAN
major premise of the GLOBE study is The purpose of the present chapter is to
Form Beta asked about societal cultural practices patterns in times of economic downturn, overall
and values1. Multiple organizations from one or demand appears to be stable in this industry.
more of three different industries (finance, food In fact, the processed food sector is frequently
processing, telecommunications) were sampled considered a safe haven for global investors in
from the 62 different cultures. times of slower economic growth.
In the next section, we will discuss the Finally, with regard to the financial services
rationale for choosing the three specific indus- industry, the dismantling of regulatory barriers
tries sampled in GLOBE. In particular, we will separating banking, insurance, and securities
discuss the range of environmental pressures segments has facilitated development of a
experienced by organizations in these different global capital market. Over the years, commer-
industries to indicate why we expect sufficient cial paper and corporate bonds have substituted
variation in these types of industries to expect bank loans, while mutual funds and securities
an effect on organizational cultural practices. have replaced some bank deposits. There has
also been competition from outside each
industry. Although automotive and aerospace
INDUSTRY SECTORS AND firms have had their own financial service sub-
ORGANIZATIONAL CULTURAL PRACTICES sidiaries for a long time, as have retailers and
several consumer electronic firms offering
The GLOBE data were sampled from the credit facilities, telecommunication and utility
telecommunications, food processing, and firms have also begun offering payment and
financial services industrial sectors. We selected other services through their distribution net-
these three industries because we believed that works and customer relationships. The result
they were present in most, if not all, countries in is growth of financial service companies that
the world, and because we believed that these offer a whole range of services, combining
industries systematically differed from one financial services production with distribution
another. In this section, we provide brief networks and brand names (Clasessens,
descriptions of these industrial sectors. We Glaessner, & Klingebiel, 2000).
obtained this information by reviewing elec- The financial services firms are intermedi-
tronic databases such as the Standard & Poor’s aries who seek to profit from exposing them-
industry surveys or the Economist Intelligence selves to risks and not matching assets and
Unit’s industry reports. liabilities on a one-to-one basis. The financial
The telecommunication industrial sector is services industry therefore is quite sensitive
relatively new in comparison to the financial to economic fluctuations. Demand for these
and food services industrial sectors. Although services tend to peak during the early phases of
the telecommunication industrial sector initially economic upturn, when interest rates are low.
embraced bureaucratic cultural practices and As economic growth proceeds, organizations
values, major transformations of this industrial that fall outside of the financial services sector
sector started in the 1980s. For example, the tend to provide sources of financing to con-
British Telecommunication Company was pri- sumers, thereby decreasing demand for the
vatized during this decade and the AT&T services of traditional financial services indus-
monopoly was broken up in the United States. try organizations. Consequently, there is much
The telecommunication industrial sector was emphasis on international diversification of
among the fastest growing sectors in the 1990s. portfolios, which makes the industry more
The food-processing industry, in contrast, is sensitive to economic crises in any part of the
one of the oldest industrial sectors. This sector world, and less sensitive to the societal factors.
is currently facing new market challenges, such In summary, there are differences in the envi-
as customer preferences for fresher, organic, ronments experienced by these three industries.
ready-to-eat, and ecologically friendly food. The The food services industry experiences a more
processed food industry is among the least sen- stable environment, the telecommunications
sitive to the general economic turbulence. Though industry has experienced a more turbulent envi-
people often tend to shift their consumption ronment since the 1980s, and the financial services
20-house.qxd 3/19/2004 5:06 PM Page 656
industry is quite sensitive to environmental founders and other organizational leaders are
pressures. Indeed, consistent with this perspec- thus differentially influenced by the larger
tive, Harvey (personal communication with society’s values, and these founders and leaders
Vipin Gupta, July 27th, 2001) has examined the have differential effects on the organizations
volatility of these three industries worldwide. they create and lead. Thus, an effect of a value
He found that the global financial services indus- at the societal level on the analogous value
trial sector is the most volatile of our three indus- at the organizational level is mediated by the
tries. The food services industrial sector has been values of founders. Resource dependency
relatively stable for a long time. Finally, Harvey theory and institutional theory also suggest that
has documented that the telecommunications coercive isomorphism (based on political and
industry experienced a dramatic change world- legal pressures), mimetic isomorphism (based
wide after the 1980s. As documented by Harvey’s on modeling other successful organizations),
research, these three industries have experienced and normative isomorphism (based on rules and
different environments over their histories and norms) all operate to influence organizations
the organizational cultural practices seen in these through the society in which they are embedded.
industries probably differ as a result. Further, the international business literature
suggests that organizations compete in a global-
izing world on the basis of their home-base
HYPOTHESES advantages. On the basis of this discussion, we
suggest the following hypothesis on the signifi-
We propose and test three potential relationships cance of home-based societal factors in the
between industrial sectors and societal cultures. practices of the organizations:
The first hypothesis is concerned with the effects
of societal systems on organizational cultures. Hypothesis 1: The societal system
The simple fact that most people live the major- has a significant effect on
ity of their lives within a single culture makes it organizational cultural practices
highly likely that the normative prescriptions
and behavioral expectations of that culture would The second hypothesis is concerned with a
be reflected in the organizations in which they common industry sector effect on organizational
function. Given that there is a long tradition in cultural practices. There is evidence that organi-
the management literature of viewing organiza- zational practices respond to the industry-related
tions as “open systems influenced by the envi- contextual contingencies. For example, Kerr,
ronment” (Katz & Kahn, 1966; Lee & Barnett, Dunlop, Harbison, and Charles (1960) and
1997, p. 398), societal culture should be a major Hickson, Hinings, McMillan, and Schwitter
source of influence on organizational systems. (1974) indicate that technology and machines
Further, the founder of an organization, along directly influence how people perform their
with other influential organizational leaders, work, regardless of nationality, and constrain
probably has a major influence on the practices, organization design choices. In addition, several
policies, and culture of an organization (Schneider, theoretical mechanisms suggest that industry
1987; Schneider, Goldstein, & Smith, 1995). It is might have a direct effect on organizational cul-
likely that these leaders try to develop organiza- ture. Resource dependency theory, for example,
tional practices and policies that appear to be sen- suggests that organizations attempt to control
sible and strategic for their organization. the resources that are necessary to them (Pfeffer,
Although there is variation in the extent to 1981; Pfeffer & Salancik, 1978). The common-
which any given individual shares a society’s ality of the perception of necessary resources
cultural values and beliefs, cultural immersion by organizational leaders in the same industry
theory as well as cognitive researchers (e.g., probably reduces the range of practices in these
Hanges, Lord, & Dickson, 2000) have argued organizations as these leaders attempt to maintain
that societal culture influences the beliefs of their organization’s technological and operational
individuals regarding what is sensible or strate- efficiency as well as achieve social legitimacy for
gic within a particular society. Values of their organization (Abernathy & Chua, 1996).
20-house.qxd 3/19/2004 5:06 PM Page 657
Further, institutional theory takes into consideration based on scientific management principles.
the larger institutional environment, which The German workers, who were vocationally
consists of the elaboration of rules, practices, qualified and experienced, underwent extensive
symbols, beliefs, and normative requirements on-the-job training and were given responsibil-
to which individual organizations must conform ity for multiple baking operations within semi-
to receive support and legitimacy (DiMaggio & autonomous, three-person teams. In contrast,
Powell, 1983; Meyer & Rowan, 1991). In an the U.S. firms used standardized production
attempt to achieve legitimacy and support, insti- procedures that were supported by a greater use
tutional theory suggests that organizations mimic of automated or dedicated production equip-
other organizations in their industry (DiMaggio & ment. The contrast between both systems is
Powell, 1983). Finally, so called “value added clear: The American experts relied on work
networks” within industries link organizations principles consistent with scientific manage-
into collectives. Such collectives tend to increase ment, and the Germans on work principles
the similarity of member organizations’ strategic consistent with sociotechnical systems.
profiles (Abrahamson & Fombrun, 1994) and In addition, a simple imitation of worldwide
thus foster the development of industry-specific industry practices is not competitively advanta-
interorganizational macrocultures, such as beliefs geous for any organization existing within a
that are shared by managers across organizations. particular society. The administrative heritage
An industrial culture, therefore, emerges because of the society plays an important role in the
of the similarity in practices by organizations adaptation of industrywide practices by any orga-
within a given industry. More generally, in global nization (Bartlett & Ghoshal, 1990). Such
industries, one would expect a shared industry- society-specific adaptation of the industrywide
wide practice to emerge as a result of bench- practices can help firms gain robust home-based
marking, and of common and converging advantage in specific industry sectors and be
technical imperatives, needs of customers, and differentiated from the competing organizations
regulatory standards. Given this review, we in other societies. Further, given the significance
advance the following hypothesis on the existence of the societal systems, one might observe the
of some global element in the industries sampled: emergence of local industry clusters, as opposed
to global industry clusters. In the local industry
Hypothesis 2: The industrial clusters, the organizations within a specific
sector has a significant main effect industry may learn from one another and develop
on organizational cultural practices similar organizational culture practices. Thus, we
have the following third hypothesis that postu-
The third, and final, hypothesis is concerned lates the existence of such local industry clusters:
with a society by industry interaction. Sufficient
evidence indicates that industrial sector effects Hypothesis 3: There is a significant
are often mediated and shaped by societal cul- industry sector-by-societal system
tures. For instance, the regulatory environment interaction effect on organizational
tends to differ across societies, as is the case cultural practices
with financial services as well as telecommu-
nications industry. Empirical evidence of this What factor drives the development of a
interaction was provided in a study by Mason convergent organizational practice within a
and Finegold (1997). These authors compared local industry cluster? We suggest that societal
the organizational practices of eight biscuit culture values play an important role in this
manufacturing factories in Germany and the process. For instance, if the value of uncertainty
United States. After controlling for size of plant, avoidance is prominent in a society, then the
the productivity of these factories was equiva- organizations may converge in adopting some
lent even though the organizations in Germany practices that respond to this common societal
appeared to be characterized as adopting value. The degree to which the organizations
sociotechnical systems principles whereas in the converge in responding to the common societal
United States, the factories were using practices values may, however, vary by industry. The
20-house.qxd 3/19/2004 5:06 PM Page 658
Argentina 4 3 7
Australia 4 3 7
Brazil 5 4 9
Canada (English) 3 3 6
China 3 3 6
Egypt 3 5 8
England 4 2 6
Finland 3 2 2 7
Georgia 4 3 7
Germany (West) 4 4 2 10
Greece 4 3 7
Hungary 2 3 5
India 4 5 9
Italy 4 3 7
Netherlands 3 3 6
Philliplines 3 3 6
Poland 2 2 4
Singapore 5 2 7
Slovenia 3 2 3 8
South Africa
(Black sample) 5 3 8
South Korea 3 3 3 9
Sweden 4 6 4 14
Switzerland 3 3 6
Taiwan 3 3 6
Thailand 5 5 5 15
Turkey 7 3 10
USA 3 3 2 8
Total 92 73 43 208
NOTES:
N = number of organizations per industry.
Data sets with less than two organizations per industry per country were excluded.
Data sets from countries with only one industry sampled were excluded.
Average number of respondents per organization = 18.6 (minimum 7; maximum 148).
scales (Hypothesis 4). A detailed description of on all nine organizational cultural practice
HLM analyses of GLOBE data is presented in dimensions. If considering variance for each
Chapter 11, by Hanges, Dickson, and Sipe. dependent variable (e.g., GLOBE Power
Distance organizational practices scale
RESULTS η(org) = 37%) that occurs only at the organi-
zational level (η2(org) in Table 20.2), societal
systems accounted for 21% to 47% of this
Societal Effects on Organizational Culture
variance. Consistent with Hypothesis 1 and
Our first hypothesis concerned whether the original GLOBE conceptual model,
societal systems affect organizational cultural societal-level differences have a substantial
practices. As shown in Table 20.2, we found impact on the cultural practices of organiza-
that societal systems had a significant effect tions. The relationship between societal and
20-house.qxd
660
Table 20.2 Results for Analyses of Variance With the GLOBE Organizational Culture Practices (“As Is”) Scales as the Dependent Variable
3/19/2004
Organizational
Cultural Practices Society Industry Society * Industry
5:06 PM
Institutional
Page 660
In-Group
Collectivism Yes .21 No .00 No .07
Performance
Orientation Yes .26 No .00 No .11
NOTE: η2 is an estimator of the amount of variance in the dependent variable accounted for by a particular factor. The significance level used for these analyses was the traditional p < .05
level.
20-house.qxd 3/19/2004 5:06 PM Page 661
7
Organizational Uncertainty Avoidance Cultural Practices
Financial
4 Food
Telecomm
1
Low Medium High
Figure 20.1 Society and Industry Interaction on GLOBE Organizational Uncertainty Avoidence Cultural
Practices
7
OrganizationalAssertiveness Cultural Practices
Financial
4 Food
Telecomm
1
Low Medium High
Figure 20.2 Society and Industry Interaction on GLOBE Organizational Assertiveness Cultural Practices
20-house.qxd 3/19/2004 5:06 PM Page 663
7
Organizational Gender Egalitarianism Cultural Practices
Financial
4 Food
Telecomm
1
Low Medium High
Figure 20.3 Society and Industry Interaction on GLOBE Organizational Gender Egalitarianism Cultural
Practices
Figure 20.3 shows the interaction between use of technology, with greater information,
societal Gender Egalitarianism cultural values communication, transportation, safety, and
and industry on organizational Gender Egalitari- health technology indicating investments in
anism cultural practices. Once again, industry- security, comfort, and resolution of uncertainty,
specific convergence in organizational Gender all of which are likely to have a strong influence
Egalitarianism cultural practices is not evident on the power in organizations.
in the financial services industry but exists for Indeed, our results indicated that societal-
the telecommunications and food industries. level Uncertainty Avoidance cultural values
Though the societal Power Distance values captured the industry by society interaction on
and industry interaction effect on organizational organizational Power Distance culture prac-
Power Distance cultural practices was not sig- tices. Figure 20.4 shows the Uncertainty
nificant, we decided to investigate if any non- Avoidance societal cultural values by industry
isomorphic societal culture value dimension interaction for the organizational Power
might explain the industry–society interaction Distance cultural practices. As can be seen in
effect on organizational Power Distance culture this figure, industry plays an important role in
practices. The most appropriate nonisomor- the positive association of organizational Power
phic societal culture value dimension influenc- Distances practices and societal Uncertainty
ing organizational Power Distance practices is Avoidance values. The association is particu-
probably Uncertainty Avoidance. Hofstede larly strong with respect to the telecommuni-
(2001) holds that organizational cultures are cation and food industries, but not with the
influenced strongly by the interaction between finance industries.
power distance and uncertainty avoidance. In summary, we found support for our cul-
Also, in Chapter 9, Gupta, de Luque, and House tural value interpretation that lies behind the
suggest that the GLOBE Uncertainty Avoidance interaction between industry and societal effects
cultural practice scale is strongly related to the as it affects organizational practices. Taken
20-house.qxd 3/19/2004 5:06 PM Page 664
7
Organizational Power Distance Cultural Practices
Financial
4 Food
Telecomm
1
Low Medium High
Figure 20.4 Society and Industry Interaction on GLOBE Organizational Power Distance Cultural
Practices
together, the organizations in the financial The financial services sector is very closely
industrial sector do not appear to have industry- involved with all the other industries, and so its
specific local organizational practices, whereas practice innovations could rapidly gain cross-
the organizations in the food and, particularly, industry acceptance. Such a process will not
the telecommunications industrial sectors are show up as an interaction-effect of society and
more likely to have industry-specific organiza- industry, but instead will show up as shared
tional practices for four of the nine GLOBE society effects on organizational practices of all
dimensions. Further, the industry-specific the industries in a society.
convergent local organizational practices are
associated with isomorphic societal culture
values for three of the practice dimensions DISCUSSION
(Uncertainty Avoidance, Gender Egalitarianism,
and Assertiveness) and with Uncertainty Our findings can be summarized succinctly.
Avoidance societal culture value for the fourth First, organizational cultural practices differ
practice dimension (Power Distance). It is pos- among the societal systems investigated regard-
sible that a greater emphasis on the global mar- less of industrial sector. This means that the
ket norms in the financial industrial sector is societal system is an important influence on
responsible for these results. These norms, in organizational culture practices irrespective of
turn, probably make the organizations in the industry. Second, we found almost no industry-
financial industrial sector less sensitive to sector effects on organizational practices across
developing a local industry-specific practice societies. Although this may appear surprising
culture. It is also possible that the local indus- given the literature suggesting likely industry
try-specific practice innovations of the finan- effects, it is explained by the third major finding
cial services organizations more rapidly diffuse regarding the interaction of societies and indus-
to the local organizations in other industries. tries. For our sample of industries, we did find
20-house.qxd 3/19/2004 5:06 PM Page 665
society-specific differences in the organizational explain the strong and significant influences of
practices of the three industries. the societal system on all organizational culture
The few and weak industry effects on orga- practices.
nizational culture across societies strongly sug- Finally, we would like to highlight some lim-
gest that the major force shaping organizational itations of the current study, and directions for
practices is rooted in societal-level systems further research. First, there may be considerable
rather than industry-specific work systems. We variability among common characteristics for
did, however, find societal system by industry distinguishing industries within societies, such as
interactions (moderate to high in effect size) for technology, growth rate, governmental regulations,
four of the nine organizational cultural practice national economic systems, traditional strength
dimensions investigated. This interaction supports of labor unions, or status of an organization
the paramount role of societal cultural variables as a national monopoly. The effects of such
as influencing organizational practices, either differences on industry characteristics and on
directly or indirectly. organizational culture have been demonstrated
We also found evidence that indirect effects in empirical studies that were conducted in
of societal system on organizational culture singular societies (e.g., Chatman & Jehn, 1994;
practices are influenced by cultural values. Dansereau & Alutto, 1990; Phillips, 1994).
Three of the four society-by-industry interaction There is a need to examine the effects of these
effects on organizational cultural practice differences in a larger cross-cultural sample,
dimensions (on Uncertainty Avoidance, on like the one used here.
Gender Egalitarianism, and on Assertiveness) Second, the three industries’ profiles may
were explained by isomorphic societal cultural also vary considerably across societies depend-
values. The fourth interaction effect (on Power ing on the degree of modernization, economic
Distance) was explained by Uncertainty growth, gross national product, or more gener-
Avoidance societal cultural value. In other ally on the political and economic systems. For
words, issues of how to deal with uncertainty an example, differences in product market con-
played a key role in determining organizational centration among food, finance, and telecom-
power distance differences among industries munication may be a consequence of differences
within societies. In societies with high scores on among societies. At the time of data gathering
the GLOBE Uncertainty Avoidance values for GLOBE, in some societies (e.g., the United
scale, respondents in both food and telecommu- States), finance, food, and telecommunication
nications organizations tended to report higher industries all consisted of privately owned com-
levels of organizational power distance cultural panies with a high number and variability of
practices than respondents from these industries competitors. In other societies (e.g., Germany,
in societies with lower scores on the GLOBE Austria), one state monopolist dominated the
Uncertainty Avoidance scale. Furthermore, the telecommunication industry or private compa-
industry-specific power distance practices of the nies were entering the very recently opened
telecommunications industry organizations market, whereas finance and food were always
were most sensitive to the uncertainty avoid- fully private. Thus, aspects in organizational
ance values of the broader society, whereas culture that relate to product market concentra-
those of the financial industry organizations tion should be more different between telecom-
were least sensitive to the uncertainty avoidance munication and the other two industries within
norms of the broader society. Similar results Germany than within the United States. Similar
were obtained for the effects of isomorphic interaction effects are to be expected for other
societal value dimensions. industry characteristics, for example, differen-
These findings may suggest that the more tial customer demands between industries, as
pervasive an industry is (finance being highly documented in the change from reliability to
pervasive), the more likely should this industry’s novelty for the AT&T monopolist in telecom-
organizations’ cultures be shared with other munication in the United States. Similarly, dif-
industries in the society. Indeed, interlinkages ferent societal expectations—for example, the
among various industries within a society could preeminence of property rights versus human
20-house.qxd 3/19/2004 5:06 PM Page 666
rights like safety and health issues—should most important and which least important?
have profound effects on the food industry, but Thus, across a multination, multi-industry
not so much on the finance and telecommuni- sample of organizations such as the GLOBE
cations industries. These and similar social sample, we expect society main effects on orga-
changes that have taken place in most Western nizational culture across all industries to be
modernized societies are currently taking place stronger than industry main effects across all
in others and have not yet begun in many other societies. Furthermore, as adaptation to indus-
societies. In this respect the GLOBE database trial conditions tends to be contingent on society
warrants further research: for example, for factors, as was demonstrated in earlier discus-
particular sets of countries and industries that sion of the study of biscuit manufacturing facto-
differ meaningfully in the above-mentioned ries in Germany and the U.S. (Mason &
characteristics, particular hypotheses about Finegold, 1997), one might also expect stronger
societal culture by industry interactions can be society-by-industry interaction effects than
investigated. industry effects alone across societies.
Third, an important explanation for the lack The interaction effect, however, is not likely
of industrial sector main effects in the GLOBE to overpower the pervasive effect of societal
study might be due to the organizational cul- culture. Unfortunately, significance tests of the
tural practice dimensions studied in Project magnitude of the explained variance for variables
GLOBE. That is, we deliberately developed at these different levels of analysis are not
organizational cultural practice scales that were available. However, we can examine the average
isomorphic to our societal cultural practice amount of variance accounted for by each
dimensions. This decision probably influenced main effect, and the interaction effects, to assess
the size of the relationships we found between whether these magnitudes were consistent with
organizational and societal practices. Clearly, it our hypotheses.
is reasonable to expect that organizations Thus, we predicted that the strongest influ-
reflect the societies in which they are embed- ence on organizational practices would be soci-
ded. We did not create or measure organiza- etal differences, followed by the industry by
tional variables that are likely to be particularly society interactions, and then the main effect
salient to organizational members from one due to industry. Although we could not formally
industrial sector over the others. For instance, test this assertion, results reported in Table 20.2
the food services industrial sector might be less provide indirect support. Specifically, the level
concerned with innovation of services in com- of organizational-level variance (for all nine
parison to organizations in the financial or scales together) accounted for by society was
telecommunication sectors. Stronger support 49% (9% of explained total variance within and
for an industrial sector main effect might have between organizations) followed by 29% orga-
been obtained if the organizational cultural nizational level variance for the industry by
practices scales focused on the practices that society interaction (4% of explained total vari-
were important for distinguishing one industrial ance within and between organizations), and 4%
sector from another. organizational-level variance for the industry
main effect (1% of explained total variance
within and between organizations). Therefore,
COMPARING EFFECT SIZES: the results are consistent with our predictions
SOCIETY, INDUSTRY, AND regarding relative effect sizes.
SOCIETY-BY-INDUSTRY INTERACTIONS
Kerr, C., Dunlop, J. T., Harbison, F. H., & Charles, ceremony. In W. W. Powell & P. J. DiMaggio
A. M. (1960). Industrialism and industrial man. (Eds.), The new institutionalism in organizational
Cambridge, MA: Harvard University Press. analysis (pp. 41–62). Chicago: University of
Kirk, R. E. (1995). Experimental design: Procedures Chicago Press.
for the behavioral sciences. San Francisco, CA: Pfeffer, J. (1981). Power in organizations.
Brooks/Cole. Marshfield, MA: Pitman.
Lee, M., & Barnett, G. A. (1997). A symbols-and- Pfeffer, J., & Salancik, G. R. (1978). The external
meaning approach to the organizational cultures control of organizations: A resource dependency
of banks in the United States, Japan, and perspective. New York: Harper & Row.
Taiwan. Communication Research, 24(4), Phillips, M. E. (1994). Industry-mindsets: Exploring
394–412. the cultures of two macro-organizational settings.
Mason, G., & Finegold, D. (1997, October). Organization Science, 5(3), 384–402.
Productivity, machinery and skills in the United Schneider, B. (1987). The people make the place.
States and Western Europe. National Institute of Personnel Psychology, 14, 437–453.
Economic Review, 162, 85–98. Schneider, B., Goldstein, H. W., & Smith, D. B. (1995).
Meyer, J. W., & Rowan, B. (1991). Institutional The ASA framework: An update. Personnel
organizations: Formal structure as myth and Psychology, 48, 747–773.
21-House.qxd 3/19/2004 5:08 PM Page 669
21
LEADERSHIP AND
CULTURAL VARIATION
The Identification of Culturally
Endorsed Leadership Profiles
PETER W. DORFMAN
PAUL J. HANGES
FELIX C. BRODBECK
n this chapter we discuss implicit leadership We extend ILT to the cultural level of analysis
AUTHORS’ NOTE: The first author would like to acknowledge the help provided by Jon Howell and Renée
Brown. Jon’s perceptive feedback and encouragement along with Renée’s excellent copyediting improved this
chapter immeasurably.
669
21-House.qxd 3/19/2004 5:08 PM Page 670
leader perceptions are a function of the overlap agreement concerning beliefs about task
between an observer’s leadership belief sys- requirements that in turn are related to team
tems and the attributes of the person being effectiveness. Since effective leadership is
rated, it is not known whether all attributes in generally critical for team success, a shared
a person’s leadership belief system are equally team mental model of effective leaders is also
important. Is it more important for a leader likely to emerge among team members.
to exhibit behavior consistent with culture- On the basis of the organizational culture
specific expectations, or for a leader to exhibit literature (e.g., Schein, 1992), it can be argued
behavior consistent with universally held that organizational members, like team mem-
leadership expectations? In other words, we bers, share mental models and implicit theories
know little about the effect of leadership that about the functioning and leadership of orga-
violates culturally endorsed norms. The pre- nizations. This is exactly what Bass (1990)
sent results from GLOBE are useful for spec- argued when he indicated that most people of
ifying the content of CLT profiles for a variety the same culture hold a common set of beliefs
of cultures. Later phases of GLOBE will explore about the attributes of a typical leader. The com-
the consequences of violating culturally monality in the leadership belief systems of
endorsed norms. organizational members probably results from
repeated exposure to common organizational
policies, practices, and procedures (e.g., organi-
Conceptual Extension of zational reward systems, common task requi-
Implicit Leadership Theory to Culturally rements) that were in place partially because of
the founder’s beliefs and values (Kraiger &
Endorsed Implicit Leadership Theory
Wenzel, 1997). In addition to the beliefs about
Even though most of the social cognitive leaders shared by team or organizational mem-
literature focuses on the information processing bers, researchers have argued that enduing, per-
of individuals, there has been some recent vasive cultural values can also influence shared
discussion of the tendency for people within conceptions of members of nations concerning
teams, groups, or cultures to develop common- effective leadership (e.g., Hanges et al., 2000).
alities in the content of their leadership beliefs. In this chapter, the hypothesis that organiza-
For example, in the team literature the concept of tional and societal cultures are associated with
“team mental models” has more recently been culturally endorsed leadership belief systems is
identified as an important construct. Klimoski tested and confirmed. To the extent that cul-
and Mohammed (1994) defined team mental tural influences are important and enduring,
models as the team members’ organized under- there should be important differences in the
standing and mental representation of knowledge CLT profiles endorsed in various cultures. For
or beliefs about key elements of the team’s envi- instance, it is likely that individualistic values
ronment. Not surprisingly, team effectiveness is espoused in societies such as the United States
believed to be partly a function of the extent to would result in a strongly shared belief
which team members share their conceptualiza- that rugged individualists are effective leaders,
tions of key task and situational requirements whereas more group-oriented leaders should
(Duncan et al., 1996). It is believed that team per- be viewed as effective in collectivistic societies.
formance is enhanced if there is greater com- Thus, what is different in our approach to
monality among the mental models of team implicit leadership theory is that we not only
members because team members can process invoke the notion of shared beliefs about
information faster with fewer errors of interpreta- leaders as theoretical and practical perspectives
tion, resulting in increased coordination among ascribing how individuals within a culture view
team members (Mohammed & Dumville, 2001; leaders and leadership, we test the differences
Mohammed, Klimoski, & Rentsch, 2000). As in the content of these CLT profiles among
predicted, Rentsch and Klimoski (2001) found cultures. The empirical formulation and testing
that several antecedents such as similar demogra- of culturally relevant CLT profiles has never
phy and team experience led to team member been done before.1
21-House.qxd 3/19/2004 5:08 PM Page 672
presented in earlier chapters and suggest practical The hypotheses progress from considering the
applications and implications of our results to potential universality of specific individual
the theory and practice of global leadership. leadership attributes (Hypothesis 1) to the
potential for developing profiles of leadership
attributes for separate cultures and clusters of
cultures (Hypotheses 2a and 2b). That is, while
THE GLOBE CONCEPTUAL the level of analysis remains at the cultural
MODEL AND CLT HYPOTHESES level, it is extended in Hypotheses 2a and 2b
from separate cultures to groupings of cultures
A diagram of the GLOBE integrated leadership that we label societal clusters. Hypotheses 3
theory was presented and discussed in Figure and 4 consider how nine societal and organiza-
2.1, Chapter 2, by House and Javidan. Recall tional cultural dimensions are related to the six
that a central theoretical proposition of this global CLT leadership dimensions identified
theory is that societal culture influences the kind previously in Chapter 8 by Hanges and Dickson.
of leadership found to be acceptable and effec- The strength of specific cultural dimensions
tive in that society. The integrated theory con- (e.g., Individualism–Collectivism) influencing
sists of many relationships and hypotheses CLT dimensions (e.g., Participative leadership)
specified in the system diagram in Figure 2.1. is examined in this chapter by determining the
For purposes of this chapter, Relationships 5 unique contribution of each cultural dimension
and 6 in the figure are critical. Societal and to a specific CLT leadership dimension.
organizational cultural values and practices
influence the process by which people come to
share common beliefs about leaders. Over time,
Hypothesis 1: Two leadership
members of cultures develop leadership proto-
characteristics—Charismatic/Value-Based
types as part of the normal socialization process
leader behavior and leader integrity—
that occurs with respect to both societal and
will be universally perceived as leading
organizational cultures. We argue that the
to effective leadership
GLOBE CLT profiles are culturally defined
prototypes and elucidate important differences We expected the Charismatic/Value-Based
between societal cultures. Research evidence leader behavior to be universally endorsed
will be presented that supports this assertion. because the visions articulated by, and the
The complete research methodology is pre- integrity enacted by, value-based leaders stress
sented in Part III of this book. values that have universal appeal (House,
Leadership questionnaire items in Project Wright, & Aditya, 1997). Charismatic/Value-
GLOBE consisted of 112 behavioral and Based leaders articulate and emphasize end-
attribute descriptors that were hypothesized to values such as dignity, peace, order, beauty,
either facilitate or impede outstanding leader- and freedom. End-values are intrinsically
ship. Items were rated on a 7-point scale that motivating, self-sufficient, and need not be
ranged from a low of 1 (this behavior or charac- linked to other values. They are not exchange-
teristic greatly inhibits a person from being an able for other values and have universal appeal
outstanding leader) to a high of 7 (this behavior (Rokeach, 1973). Thus, the end-values stressed
or characteristic contributes greatly to a person by Charismatic/Value-Based leaders are likely
being an outstanding leader). Items reflecting to be universally accepted and endorsed; conse-
both organizational and societal cultures were quently, their visions are also likely congruent
also a major aspect of GLOBE Phase 2. We with the values stressed in the culture (House &
obtained responses to approximately 17,000 Aditya, 1997). Similarly, we expected leader
questionnaires from middle managers of 951 integrity to be universally endorsed because
organizations in 62 nations. integrity is an end-value that is also universally
The following four hypotheses concern how held in all cultures (Rokeach, 1973).
leadership attributes regarded as critical for effec- We recognize that aspects of Hypothesis 1,
tive leadership are influenced by specific cultures. particularly with respect to the universal
21-House.qxd 3/19/2004 5:08 PM Page 674
Table 21.1 shows the 6 global CLT leadership CLT Autonomous leadership dimension includes
dimensions along with the 21 primary or first- a single primary leadership subscale labeled
order leadership dimensions. These dimensions autonomous.
are summary indices of the characteristics,
skills, and abilities culturally perceived to Self-Protective. From a Western perspective,
contribute to, or inhibit outstanding leadership. this newly defined leadership dimension
They can be thought of as being somewhat focuses on ensuring the safety and security of
similar to what laypersons refer to as leadership the individual or group member. The GLOBE
styles and are defined as follows: CLT Self-Protective leadership dimension
includes five primary leadership subscales
Charismatic/Value-Based. A broadly defined labeled (a) self-centered, (b) status conscious,
leadership dimension that reflects the ability (c) conflict inducer, (d) face saver, and
to inspire, to motivate, and to expect high perfor- (e) procedural.
mance outcomes from others on the basis
of firmly held core values. This GLOBE CLT Given the information presented earlier in
Charismatic/Value-Based leadership dimension this chapter, GLOBE researchers hypothesized
includes six primary leadership subscales labeled that it would be possible to construct profiles
(a) visionary, (b) inspirational, (c) self-sacrifice, of leadership dimensions based on individuals’
(d) integrity, (e) decisive, and (f) performance questionnaire responses aggregated to the cul-
oriented. ture and culture cluster levels. These aggregated
responses would indicate the extent to which
Team Oriented. A leadership dimension that certain leader attributes and leadership dimen-
emphasizes effective team building and imple- sions are perceived to contribute to effective
mentation of a common purpose or goal among leadership among cultures and clusters.2
team members. This GLOBE CLT Team-
Oriented leadership dimension includes five
Hypothesis 3: There will be positive
primary leadership subscales labeled (a) collab-
relationships between CLT dimensions and
orative team orientation, (b) team integrator,
societal culture dimensions that are
(c) diplomatic, (d) malevolent (reverse scored),
conceptually similar or clearly related on
and (e) administratively competent.
theoretical grounds
Participative. A leadership dimension that The rationale for this hypothesis is that
reflects the degree to which managers involve cultural dimensions would influence the legiti-
others in making and implementing decisions. macy and acceptance of leader behaviors. More
The GLOBE CLT Participative leadership specifically, we believe that culture would
dimension includes two primary leadership influence the kinds of attributes and behaviors
subscales labeled (a) autocratic (reverse scored) considered to be effective. For GLOBE, this
and (b) non participative (reverse scored). cultural influence hypothesis (Triandis, 1995)
posits that culture will have a pervasive influ-
Humane Oriented. A leadership dimension ence on values, expectations, and behaviors
that reflects supportive and considerate leader- and would, therefore, influence the content of
ship but also includes compassion and gen- the CLT profiles. For instance, because charis-
erosity. The GLOBE CLT Humane Oriented matic leaders motivate followers to achieve
leadership dimension includes two primary high levels of performance, we expect that
leadership subscales labeled (a) modesty and societies that promote high societal expectations
(b) humane oriented. for performance would likewise believe in
the efficacy of charismatic leaders and the CLTs
Autonomous. This newly defined leadership profiles associated with such leaders. In sum,
dimension has not previously appeared in the predicting and finding significant relationships
literature. This dimension refers to independent between specific cultural dimensions (e.g.,
and individualistic leadership. The GLOBE Performance Orientation) and conceptually
21-House.qxd 3/19/2004 5:08 PM Page 676
NOTE: The numbered, italicized topics are global CLT leadership dimensions. They consist of primary CLT leadership
subscales. The only exception is dimension 6 (Autonomous), which consists of a single subscale of four questionnaire items.
It is considered both a specific subscale and global dimension. Numbers represent mean values for the 62 societal cultures on
a 7-point scale ranging from 1 (greatly inhibits) to 7 (contributes greatly to) outstanding leadership.
relevant CLT leadership dimensions (e.g., hypothesis. Note that Hypothesis 4 concerns
Charismatic/Value-Based) constitutes a test of organizational influence and associated levels
the cultural influence hypothesis. Because we of analyses, whereas Hypothesis 3 concerns
also expect that national culture would also societal influence and levels of analyses.
influence organizational values and practices, We present our results in the following order.
we form the following hypothesis. First, we provide evidence related to Hypothesis
1—specific attributes that are universally seen
Hypothesis 4: There will be positive as facilitators or inhibitors of effective leader-
relationships between organizational culture ship, or as inhibitors in some cultures and facil-
dimensions and CLT leadership dimensions itators in other cultures. We then present results
that are conceptually similar or related on related to Hypotheses 2a and 2b. For Hypothesis
theoretical grounds 2a, we investigate whether respondents from
62 different cultures differ from each other in
We believe that in addition to the effect of terms of CLT leadership dimensions that
societal culture, organizational cultural values enhance effective leadership. The focus of
and practices will influence the expectation that Hypothesis 2b was whether clustering societies
certain leadership attributes lead to effective into the 10 culture categories (reported by
leadership. For instance, organizational cultures Gupta and Hanges in Chapter 10) results in
that stress performance orientation should hold differences in CLT profiles for these clusters.
in high regard leaders with high expectations Finally, we summarize the hierarchical linear
of their followers. Determining the relationships modeling (HLM) evidence that addresses Hypo-
between organizational culture dimensions and theses 3 (societal influence on CLT leadership
conceptually relevant CLT leadership dimensions dimensions) and 4 (organizational influence on
constitutes a test of this organizational influence CLT leadership dimensions).
21-House.qxd 3/19/2004 5:08 PM Page 677
Trustworthy Integrity
Just Integrity
Honest Integrity
Foresight Charisma 1: Visionary
Plans ahead Charisma 1: Visionary
Encouraging Charisma 2: Inspirational
Positive Charisma 2: Inspirational
Dynamic Charisma 2: Inspirational
Motive arouser Charisma 2: Inspirational
Confidence builder Charisma 2: Inspirational
Motivational Charisma 2: Inspirational
Dependable Malevolent (reverse score)
Intelligent Malevolent (reverse score)
Decisive Decisiveness
Effective bargainer Diplomatic
Win–win problem solver Diplomatic
Administrative skilled Administratively competent
Communicative Team 2: Team Integrator
Informed Team 2: Team Integrator
Coordinator Team 2: Team Integrator
Team builder Team 2: Team Integrator
Excellence oriented Performance oriented
Loner Self-protective
Asocial Self-protective
Noncooperative Malevolent
Irritable Malevolent
Nonexplicit Face saver
Egocentric*
Ruthless*
Dictatorial Autocratic
The visionary and inspirational leadership of culture scores on the item be less than 3 on a
dimensions are critical aspects of Charismatic/ 7-point scale.
Value-Based leadership (Yukl, 2002) and con- These combined criteria would indicate that
tain the most number of attributes universally the attribute is universally perceived as inhibit-
perceived as contributors to effective leadership ing outstanding leadership. Results are pre-
(see Table 21.2). However, attributes comprising sented in Table 21.3, which shows that many
the self-sacrificial dimension (e.g., risk taking) attributes in the primary CLT dimensions
of Charismatic/Value-Based leadership were not labeled self-protective and malevolent are uni-
universally endorsed. Please note that three of versally viewed as impediments to effective
the positively endorsed items concerned leadership by the GLOBE international sample
aspects of integrity. In addition, most of the of middle managers.
other universal positively endorsed attributes,
which did not relate to the Charismatic/ Culturally Contingent Endorsement of Leader
Value-Based leadership dimension, were found Attributes. Most interesting, from a cross-
in the Team-Oriented dimension (see Table 21.2). cultural viewpoint, were attributes that in some
The portrait of a leader who is universally cultures are considered to enhance outstanding
viewed as effective is clear: The person should leadership and in other cultures are considered to
possess the highest levels of integrity and impede outstanding leadership. We present
engage in Charismatic/Value-Based behaviors in Table 21.4 those aggregated attributes (ques-
while building effective teams. These question- tionnaire responses) that yielded scores above
naire results strongly support the hypothesis and below the scale midpoint of 4, contingent on
that charismatic–visionary and charismatic– culture-specific responses aggregated to the soci-
inspirational attributes of the Charismatic/ etal level of analysis. Many of these attributes fell
Value-Based leadership dimension are univer- into the self-centered and individualistic leader-
sally endorsed as contributing to outstanding ship dimensions (see Table 21.4). For instance,
leadership. although the attribute individualistic had a grand
culture mean of 3.11 (slightly inhibits outstand-
Universal Impediments to Leadership Effecti- ing leadership), individual culture scores ranged
veness. Our criteria for specific attributes to from a low of 1.67 (somewhat inhibits) to a high
be universally considered as impediments to of 5.10 (slightly contributes). Similarly, the
effective leadership required that (a) the attribute status conscious ranged in value from
attribute or item grand mean for all cultures a low of 1.92 (somewhat inhibits) to a high
be less than 3 on a 7-point scale and (b) 95% of 5.77 (moderately contributes). Even more
21-House.qxd 3/19/2004 5:08 PM Page 679
NOTE: CLT = culturally endorsed implicit leadership theory. Numbers represent minimum and maximum values for the 62
societal cultures on a 7-point scale ranging from 1 (greatly inhibits) to 7 (contributes greatly to) outstanding leadership.
* These items did not load on any factor.
striking was the risk taker attribute, a component and charismatic–inspirational attributes, but
of the primary CLT dimension labeled Charis- not for charismatic–self-sacrificial. Team-
matic III: Self-Sacrificial, which, in turn, is part Oriented leadership is strongly correlated with
of the Charismatic/Value-Based global CLT Charismatic/Value-Based leadership, and
dimension. The risk taker attribute ranges in attributes comprising this dimension are also
value from 2.14 (somewhat inhibits) to 5.96 universally endorsed. Humane–Oriented and
(contributes somewhat). Participative leadership dimensions are gener-
ally viewed positively, but significant variabil-
Summary of Hypothesis 1: CLT Results. For ity exists across cultures. The remaining CLT
GLOBE, an initial question concerned whether dimensions of Self-Protective and Autonomous
the six global CLT leadership dimensions leadership are generally viewed as neutral or
would be viewed positively or negatively. negative but still vary significantly by culture.
Given the prior research on charismatic leader-
ship, summarized by House et al. (1997) among
Hypotheses 2a and 2b: CLT
others (Dorfman, 1996; Yukl, 2002), it should
Dimensions Associated With GLOBE
not be surprising that one GLOBE hypothesis
Societal Cultures and Societal Clusters
was that attributes comprising the Charismatic/
Value-Based leadership dimension would be For each culture we developed a profile that
universally endorsed. This hypothesis is reported the extent to which the six CLT leadership
strongly supported for charismatic–visionary dimensions were believed to contribute to
21-House.qxd 3/19/2004 5:08 PM Page 680
NOTE: CLT leadership scores are absolute scores aggregated to the cluster level.
effective leadership.3 For each culture we also Table 21.5 presents absolute scores for all
used both absolute measures obtained directly 10 clusters. Analysis of Variance (ANOVA)
from the raw score questionnaire data, and a was used to determine if the cultures and clus-
relative measure constructed from the former ters differed with respect to the CLT leadership
to further illustrate similarities and differ- dimensions reported to be contributors to effec-
ences among cultures. Both measures are tive leadership. Results indicate that cultures
described in the following sections. (i.e., 62 societal cultures) and clusters (i.e., 10
groups consisting of the 62 societal cultures)
Absolute Measure differed with respect to all six CLT leadership
dimensions (p < .01).
Recall that the scores for each CLT leader-
ship dimension (such as Team Oriented)
Relative Measure
could range from 1 (greatly inhibits outstanding
leadership) to 7 (contributes greatly outstanding In addition to the absolute measure described
leadership). A profile can be developed and above, we calculated a relative measure that
presented in graph form indicating aggregated controlled for some of the potential cultural
responses of informants in a specific culture on response biases that could influence our leader-
each of the six CLT leadership dimensions. As ship survey results. As discussed in Chapters 4
indicated previously, we called the profile of the and 8, cultural values can bias the questionnaire
CLT leadership dimensions for each cluster the responses of individuals. For example, a general
cluster culturally endorsed leadership profile acquiescence response bias occurs if respon-
(cluster CLT profile). This cluster CLT profile dents use a general positive or negative
is our best quantitative summary of the content response set. In this case, the responses would
of the leadership belief system shared by be positively or negatively skewed, assuming
members in each of these clusters. For practical the “true” responses are different from the actual
reasons, we present graphs representing the responses. Second, research has shown that there
CLT profiles for the 10 clusters as opposed to are societal-wide differences in individuals’
graphs presenting findings for all 62 cultures. responses to closed-ended items in Likert-type
21-House.qxd 3/19/2004 5:08 PM Page 681
scales, either exclusively using or not using end important for each cluster. Also, we should
points, or only midpoints, of the scale (Dorfman, point out that the correlation between the
2004; Triandis, 1994). A third problem concerns absolute and relative measures is close to
the rating of concepts that may not be exactly perfect—above .90 for all of the CLT leadership
comparable due to differences in meaning cre- dimensions. Furthermore, using the relative
ated during language translation. These response measure along with the absolute measure guards
biases can cause difficulty if comparing cultures against the possibility that our results are tainted
using questionnaire items. by response biases, and specifically lends credi-
As discussed in Chapter 8, we went to extra- bility to the validity of our ranking cultures on
ordinary lengths when designing the GLOBE CLT leadership dimensions.
project and questionnaire to minimize response Although the concept of standardization
bias. To give one example, after obtaining the used to create the relative measure is simple,
data we used a statistical procedure to check the mechanics are more complex (fully
on the actual amount of response bias in our described in Chapter 8). First, we conducted
results by creating a response bias corrected a “within-person” standardization procedure
measure. This procedure is described in detail and aggregated those responses to the culture
in Chapter 8 by Hanges and Dickson and is level and then to the cluster level. Responses to
advocated by cross-cultural researchers (van de all items in each respondent’s questionnaire
Vijver & Leung, 1997). Whereas one might use were standardized around the person’s average
this measure as simply a check on bias, in this response score and standard deviation. To illus-
chapter we will use the response bias corrected trate, think of two people who respond exactly
measure as an integral part of the analysis strat- the same to all questions but with one signifi-
egy. We are going to refer to this measure here- cant difference. One person always responds
after as the relative measure because of a to items using the extreme positive end of
unique property attributed to this procedure. the scale, whereas the other is more conserva-
These relative CLT scores indicate the relative tive and responds positively, but more toward
importance of each CLT leadership dimension the midpoint of the scale. Assume that the
within a person, culture, or culture cluster.4 leadership attribute integrity is the most impor-
Computing relative scores is critical because tant leadership attribute to each respondent.
they are an important check on the accuracy Although the mean score for this attribute
and validity of the absolute scores. Recall that will vary between these two people because of
the latter are problematic if interpreting ques- their acquiescence response bias, their standard
tionnaire responses, particularly if indicating scores will not (assuming equal variance
differences among cultures or clusters. That is, around each mean score).
this procedure not only removed the cultural After computing the relative scores, we
response biases, but it also had the advantage of assessed whether cultures and clusters differed
illustrating the differences among the cultures with respect to both absolute and relative CLT
and the clusters. It is analogous to using a measures. As expected, all Analysis of Variance
microscope. It sensitized us to meaningful dif- (ANOVA) results were significant using either
ferences within each cluster as to the relative culture or cluster as the independent variable
importance of each CLT leadership dimension. and CLT leadership dimension scores as the
We also used this relative measure to compare dependent variable (for both absolute and rela-
the relative importance of each CLT dimension tive scores, p < .01). These tests confirmed the
among cultures.5 differences among clusters using both measures.
We should note that in most cases, the rela- For ease of presenting and communicating the
tive CLT measures confirmed the significant results, cluster data are shown in several for-
differences among the clusters with respect to mats. Figures 21.1–21.10 show the absolute
using the original (i.e., absolute) scores. As a CLT scores for each cluster in a graphic form.
caveat, both the relative and absolute measures The graphs enable us to compare the importance
should be viewed in tandem to provide an accu- of various CLT dimensions within each cluster.
rate portrait of the CLT leadership dimensions Table 21.5 presents the absolute scores for each
21-House.qxd 3/19/2004 5:08 PM Page 682
NOTE: The placement of each societal cluster indicates the relative rank of this cluster compared with other clusters with
regard to the size of the absolute scores on this dimension. For example, the Anglo cluster is the highest in rank for
Charismatic/Value-Based leadership, indicating that this leader dimension had the highest score (absolute measure) compared
with other clusters. Using the Tukey HSD analysis, clusters in the top band are significantly different from those in the bottom
band. The clusters in the middle band are placed between these extremes for heuristic purposes. Societal clusters within each
block are not significantly different from each other.
a Societal clusters in this column are ranked in order; however, there are no significant differences among them.
of the 10 clusters in a format that enables us to includes the United States, we note from Table
compare the CLT dimension scores across clus- 21.5 and Figure 21.6 that the Anglo cluster has
ters. Finally, three additional tables indicate the high absolute scores for Charismatic/Value-
rank, from high to low, of each cluster with regard Based, Team-Oriented, and Participative CLTs,
to the CLT dimension scores. Table 21.6 ranks and a moderately high score for the Humane-
each cluster using absolute scores. Table 21.7 Oriented CLT. The Autonomous CLT score
ranks the clusters using relative scores, and indicates that this factor is about neutral (mean
Table 21.8 summarizes the cluster rankings. score = 3.82, which is slightly below the mid-
An example should help clarify this “two- point of the scale). The Self-Protective leader-
step look” using both absolute and relative ship score is low (mean = 3.08), indicating that
scores for the six CLT leadership dimensions. this factor is perceived to inhibit effective
By examining the societal cluster Anglo, which leadership. Examining the ranking of clusters
21-House.qxd 3/19/2004 5:08 PM Page 683
Table 21.7 Ranking of Societal Clusters Using Relative (i.e., Standardized) CLT Scores
Anglo Southern Asiaa Germanic E. Southern Asia Germanic E.b Middle East
Germanic E. E. Europe Anglo Anglo E. Europe Confucian A.
Nordic E. Confucian A. Nordic E. Sub-Sahara A. Confucian A. Southern Asia
Southern Asia L. America Confucian A. Nordic E. E. Europe
L. Europe Sub-Sahara A. Southern Asia L. America
L. America L. Europe Anglo
Nordic E. Middle East
Confucian A. Anglo L. Europe Germanic E. L. Europe Sub-Sahara A.
Sub-Sahara A. Middle East L. America Middle East Sub-Sahara A. L. Europe
E. Europe Germanic E. Sub-Sahara A. L. America L. America
E. Europe
NOTE: The placement of each societal cluster below a leadership dimension indicates the relative importance of this dimension
compared with the other leadership dimensions within a particular societal cluster. For example, the Anglo cluster is the highest
in rank for Charismatic/Value-Based leadership, indicating that this leader dimension was extremely important (relative measure)
in comparison to the other five leadership dimensions within the Anglo cluster. The size of the relative score for each societal
cluster is thus compared to the size of the relative scores for other societal clusters. Using the Tukey HSD analysis, clusters in
the top band are significantly different from those in the bottom band. The clusters in the middle band are placed between these
extremes for heuristic purposes. Societal clusters within each block are not significantly different from each other.
a, b Societal clusters in these columns are ranked in order; however, there are no significant differences among them in each
column.
for the absolute scores (Table 21.6), it is present two comparisons simultaneously. The
apparent that the Anglo cluster is among the first comparison shows the strength of each CLT
highest ranks for Charismatic/Value-Based, leadership dimension compared with the other
Participative, and Humane-Oriented CLT lead- CLT dimensions within the same cluster. Recall
ership dimensions, but among the lowest for that this is an ipsative measure. The second com-
Self-Protective. parison shows how each CLT leadership dimen-
Now let’s examine the same six CLT leader- sion’s relative score compares with other relative
ship dimensions using the relative scores (see scores from other clusters. By examining Table
Table 21.7). Recall that each relative score 21.7, note that the Anglo cluster has the highest
contrasts the importance of a global leadership ranking for Charismatic/Value-Based leader-
dimension in each cluster by comparing it ship. This cluster is also highly ranked for
to other leadership dimensions within that Participative and Humane-Oriented CLT dimen-
same cluster. Table 21.6 allows us to make and sions. Furthermore, the Anglo cluster is ranked
21-House.qxd 3/19/2004 5:08 PM Page 684
Table 21.8 Summary of Societal Cluster Rankings for CLT Leadership Dimensions
NOTE: For letters separated by a “/”, the first letter indicates rank with respect to the absolute score, second letter with respect
to relative score.
H = high rank; M = medium rank; L = low rank.
H or L (bold) indicates Highest or Lowest cluster score for a specific CLT dimension.
about average on the Autonomous CLT dimen- profiles for each of the 62 societal Cultures
sion, but somewhat lower on the Team Oriented considered independently of their cluster, these
dimension and among the lowest on the Self- CLT scores are presented in Appendix 21.1.
Protective CLT dimension. Taken together, this
presents a compelling image of facilitating and Eastern Europe Cluster. This cluster includes
inhibiting leadership characteristics for the Albania, Georgia, Greece, Hungary, Kazakhstan,
Anglo cluster relative to the hierarchy of the Poland, Russia, and Slovenia. In examining
other clusters in each of the CLT leadership Table 21.5 and Figure 21.1 (absolute scores),
dimensions. The Anglo cluster strongly endorses note that the Charismatic/Value-Based and
Charismatic/Value-Based leadership qualities Team-Oriented CLT leadership dimensions are
along with Humane-Oriented and Participative reported to be the most important dimensions
leadership. However, Anglo respondents report for effective leadership. Participative leader-
that Self-Protective leadership is an impediment ship is viewed positively, as is Humane-
to effective leadership and the Autonomy CLT Oriented leadership, but not as important as
dimension is about average relative to other the first two CLT leadership dimensions.
societal clusters. Autonomous leadership is viewed in a neutral
to slightly positive manner, whereas Self-
Protective leadership is slightly negative.
Ten Societal Clusters
Table 21.6 provides a ranking of country clus-
and Associated CLT Profiles
ters using the absolute CLT scores. The Eastern
We will now examine each cluster and Europe cluster has the highest rank of all clus-
identify its distinguishing CLT leadership ters for Autonomous leadership. It is among the
dimension characteristics with respect to lowest grouping of clusters for Participative
culturally endorsed leadership profiles at the leadership, and is among the highest clusters
cluster level.6 Although we will not discuss the for Self-Protective leadership.
21-House.qxd 3/19/2004 5:08 PM Page 685
Charismatic/Value-Based
7.00
5.74
5.00
1.00
4.20
5.08
Autonomous Participative
4.76
Humane Oriented
Countries: Albania, Georgia, Greece, Hungary, Kazakhstan, Poland, Russia, and Slovenia
Examining the relative CLT scores in These rankings indicate that a leadership
Table 21.7 provides a view that compliments exemplar for the Eastern Europe cluster would
the findings with the absolute scores presented be one who is somewhat Charismatic/Value-
above. Supporting the findings with the absolute Based, Team-Oriented, and Humane-Oriented,
scores, the most noteworthy result is that in but is his or her own person, does not particu-
comparison to other clusters, the Eastern Europe larly believe in the effectiveness of Participative
cluster is ranked as one of the lowest clusters for leadership, and is not reluctant to engage in
Participative leadership and one of the highest Self-Protective behaviors if necessary.
clusters for Self-Protective leadership. Similar
to the rank obtained using absolute scores, this Latin America Cluster. The Latin America cluster
cluster has the second highest rank for includes the largest number of cultures in the
Autonomous leadership. Thus, ranking the clus- GLOBE sample (Argentina, Bolivia, Brazil,
ters using relative CLT scores reinforces the Colombia, Costa Rica, Ecuador, El Salvador,
information provided by the absolute scores. Guatemala, Mexico, and Venezuela). In examin-
Both absolute and relative findings for the 10 ing Table 21.5 and Figure 21.2 (absolute scores),
culture clusters and six CLT leadership dimen- we note that the CLT leadership dimensions
sions are summarized in Table 21.8. contributing the most to outstanding leadership
21-House.qxd 3/19/2004 5:08 PM Page 686
Charismatic/Value-Based
7.00
5.99
5.00
3.00 5.96
3.62
1.00
3.51
5.42
Autonomous Participative
4.85
Humane Oriented
Countries: Argentina, Bolivia, Brazil, Colombia, Costa Rica, Ecuador, E1 Salvador, Guatemala, Mexico, and Venezuela
include Charismatic/Value-Based and Team- (Table 21.6). The Latin America cluster moves
Oriented leadership, followed by the Participative from the middle block of clusters with
and Humane-Oriented CLT dimensions. Autono- respect to the Self-Protective CLT dimension
mous and Self-Protective leadership are viewed (absolute scores) to the highest block of clusters
as slightly inhibiting outstanding leadership. (relative scores). It remains among the highest
Ranked by absolute CLT scores in Table 21.6, the ranked clusters for Charismatic/Value-Based
Latin America cluster has the highest rank for the leadership, and its Team-Oriented leadership
Team-Oriented CLT leadership dimension, the scores keep it among the top three ranks after
second highest rank for Charismatic/Value-Based removing potential response bias. As might be
leadership, and ranks lowest with respect to expected given the high Team-Oriented CLT
the Autonomous CLT leadership dimension. It score, Autonomous leadership was rated
occupies the middle ranks for the remaining CLT lowest using relative scores, supporting the
leadership dimensions. low rank for absolute scores. The Latin
The ranking of clusters using relative CLT America cluster rankings for Participative and
scores (Table 21.7) is very similar, but not Humane-Oriented leadership remain about
identical to the ranking using absolute scores average in comparison to other clusters. A
21-House.qxd 3/19/2004 5:08 PM Page 687
Charismatic/Value-Based
7.00
5.78
5.00
3.00 5.73
3.19
1.00
3.66
5.37
Autonomous Participative
4.45
Humane Oriented
summary of CLT leadership dimension findings examining Table 21.5 and Figure 21.3, we note
is presented in Table 21.8. that the CLT leadership dimensions viewed as
For the Latin America cluster, an exemplar most contributing to outstanding leadership
of effective leadership would be a person who include Charismatic/Value-Based and Team-
practices Charismatic/Value-Based and Team- Oriented leadership. Participative leadership is
Oriented leadership, and would not be adverse to viewed positively but is not as important as the
some elements of Self-Protective leadership. first two dimensions. Humane-Oriented leader-
Independent action would not be endorsed. ship is viewed as slightly positive, whereas
Participative and Humane-Oriented leadership Autonomous leadership is viewed as slightly
behaviors would be viewed favorably, but not to negative and Self-Protective is viewed nega-
the highest level as in other clusters. tively. By examining the rankings using
absolute scores in Table 21.6, note that the Latin
Latin Europe Cluster. France, Israel, Italy, Europe cluster is in the middle rank for all CLT
Portugal, Spain, and Switzerland (French- leadership dimensions except the Humane-
speaking) are included in this cluster. By Oriented and Autonomous CLT dimensions,
21-House.qxd 3/19/2004 5:08 PM Page 688
in which it ranks among the lowest scoring of the paternalistic but autocratic father in
clusters. families. Autonomous leadership is viewed neu-
As with the other clusters, differences between trally, and Self-Protective leadership is seen as
the Latin Europe cluster and others are further a slight impediment to effective leadership. In
revealed by examining the relative CLT scores in examining the rankings in Table 21.6, note that
Table 21.7. When comparing relative CLT scores the Confucian Asia cluster is ranked low with
with those of other clusters, it is noteworthy that respect to Participative and high with respect to
the Latin Europe cluster is ranked similarly to the Self-Protective leadership dimensions.
Latin America cluster, but with a few differences. The relative ranking of clusters (Table 21.7)
One difference occurs with respect to the Self- reveals that the Confucian Asia cluster is
Protective CLT leadership dimension. In the Latin characterized by relatively high scores for
Europe cluster, Self-Protective leadership is Self-Protective, Team-Oriented, and Humane-
viewed as more of an impediment to effective Oriented leadership. Again, what is striking is
leadership than in the Latin America cluster. This the very low relative ranking for Participative
is true for both the absolute and relative scores. leadership. Thus, we find a high level of
Another interesting contrast is reflected in agreement between the relative and absolute
the Humane-Oriented CLT dimension. The Latin measures for several CLT dimensions. For
Europe cluster occupies the lowest band for the instance, the absolute CLT scores for Partici-
Humane-Oriented CLT in contrast to the Latin pative leadership rank the Confucian Asia clus-
America cluster, which is found in the middle ter as one of the lowest of all clusters; it is also
band (for both absolute and relative scores). ranked low on the basis of relative CLT mea-
In sum, by integrating all CLT leadership sures. Although the absolute scores indicate
dimension results for the Latin Europe cluster that Self-Protective leadership is approximately
(see Table 21.8), an exemplar of an effective neutral as a contributor to effective leadership
leader would be a person who endorses (mean = 3.72), it is among the highest scores in
Charismatic/Value-Based and Team-Oriented all the clusters. However, the relative measure
leadership. Autonomous action would not be shows more relevance for the Team-Oriented
endorsed and Humane-Oriented behaviors and Humane-Oriented CLT dimensions than
would not play a particularly important role. does the absolute measure. In addition, the
Although Participative leadership would be Confucian Asia cluster is among the highest
viewed favorably, the Latin Europe cluster ranks for Self-Protective leadership along
would not be noted for it. with the Southern Asia and Middle East clus-
ters. We can account for the relatively high
Confucian Asia Cluster. The Confucian Asia Self-Protective scores by examining specific
cluster includes China, Hong Kong, Japan, scores for the 21 first-order leadership factors
Singapore, South Korea, and Taiwan. In exam- (also called the primary leadership factors,
ining Table 21.5 and Figure 21.4, we note that which, are not shown here), where it is apparent
the CLTs contributing to outstanding leadership that modesty and face saving are viewed posi-
include Charismatic/Value-Based and Team- tively for the Confucian Asia cluster, as might
Oriented leadership, even though the latter is be expected for the collective cultures in the
not scored particularly high. Humane-Oriented Far East (Bond & Hwang, 1986). Finally, as
leadership is viewed favorably, but it is not as will be apparent when discussing the Southern
important as the first two CLT dimensions. Asia cluster later in this section, the Confucian
Although Participative leadership is also viewed Asia and Southern Asia CLT dimension rank-
positively, it is about equal to the lowest-scoring ings are very similar to each other with regard
clusters. In fact, the Confucian Asia and to effective leadership (see Table 21.8).
Southern Asia clusters share the interesting An example of effective leadership for
characteristic that they are the only two clusters the Confucian Asia cluster would include
in which the Participative CLT score is less than Charismatic/Value-Based and perhaps Team-
the Humane-Oriented CLT score for effective Oriented leadership. Self-Protective actions
leadership. This finding likely reflects the role are viewed less negatively than in other
21-House.qxd 3/19/2004 5:08 PM Page 689
Charismatic/Value-Based
7.00
5.63
5.00
3.72 3.00
5.61
1.00
4.04
4.99
Autonomous Participative
5.04
Humane Oriented
Countries: China, Hong Kong, Japan, Singapore, South Korea, and Taiwan
Charismatic/Value-Based
7.00
5.93
5.00
3.00 5.77
2.72
1.00
3.94
5.75
Autonomous Participative
4.42
Humane Oriented
extremely inhibiting to effective leadership. The contrast to most other cluster profiles, it would
Participative CLT rank is among the highest, be particularly noted for high Participative
as might be expected for this cluster. Interes- leadership and low Humane-Oriented and
tingly, the Humane-Oriented leadership score is Self-Protective attributes.
among the lowest of all clusters. This finding
may be surprising, especially because this clus- Anglo Cluster. The Anglo cluster includes
ter’s score on the Humane Orientation cultural Australia, Canada (English-speaking), Ireland,
values score is the highest of all clusters. This New Zealand, South Africa (White sample), the
paradox probably reflects specific attributes rela- United Kingdom, and the United States. In exam-
ted to effective leadership and does not reflect a ining Table 21.5 and Figure 21.6 (absolute
general nonendorsement of humane values. scores), we note that the CLT leadership dimen-
In sum, by examining all CLT leadership sions viewed as contributing to outstanding lead-
dimension findings together (Table 21.8), the ership include Charismatic/Value-Based and
effective Nordic Europe leadership style would Team-Oriented leadership. The Charismatic/
contain elements of Charismatic/Value-Based Value-Based score is the highest for all clusters.
and Team-Oriented leadership. However, in Participative leadership is viewed positively and
21-House.qxd 3/19/2004 5:08 PM Page 691
Charismatic/Value-Based
7.00
6.05
5.00
3.00 5.74
3.08
1.00
3.82
5.73
Autonomous Participative
5.08
Humane Oriented
Countries: Australia, Canada (English-speaking), Ireland, New Zealand, South Africa (White sample),
the United Kingdom, and the United States
has a score equal to that of the Team-Oriented Differences between the Anglo cluster and
CLT. (This is relatively unusual in our data and others can also be seen by examining the
only occurs within the Anglo, Nordic Europe, and relative CLT scores in Table 21.7. When
Germanic Europe clusters; the last is the only comparing relative CLT scores with those of
cluster in which the Participative dimension is other clusters, note that those for Charismatic/
valued more than Team Oriented.) Humane- Value-Based, Participative, and Humane-Oriented
Oriented leadership is considered a facilitator of leadership are very high in the hierarchy of
effective leadership, and Self-Protective leader- scores for the Anglo cluster. Surprisingly, the
ship is considered an impediment. In examining Team-Oriented CLT dimension score is among
the ranking of clusters in Table 21.6, note that the the lower end of the ranks for the clusters
Team-Oriented CLT dimension score is about (using the relative measure), likely reflecting
average in comparison to other clusters, Humane- the strong individualistic characteristic repre-
Oriented leadership is viewed positively and is sented in the Anglo cluster. This cluster is
among the top ranking, and Self-Protective lead- ranked very low on Self-Protective leadership,
ership is among the lowest ranking. indicating that status conscious, face saving,
21-House.qxd 3/19/2004 5:08 PM Page 692
Charismatic/Value-Based
7.00
5.79
5.00
3.00 5.70
3.55
1.00
3.63
5.31
Autonomous Participative
5.16
Humane Oriented
Countries: Namibia, Nigeria, South Africa (Black sample), Zambia, and Zimbabwe
and self-centered attributes strongly inhibit Zimbabwe are included in this cluster. The
effective leadership. The Autonomous CLT absolute scores for the Sub-Saharan Africa clus-
dimension score is about average in comparison ter (see Figure 21.7) are not particularly striking
to other cluster rankings. with the exception of the high score for the
Thus, to summarize this cluster (see Table 21.8), Humane-Oriented CLT leadership dimension
an exemplar of effective Anglo leadership includes and relatively low score for the Autonomous
high Charismatic/Value-Based elements with CLT dimension (see Table 21.5). The Humane-
high levels of Participative leadership enacted in Oriented score was the second highest for
a Humane-Oriented manner. Team-Oriented is all clusters, as seen in Table 21.6. Similar to
valued, but not ranked among the highest CLT other clusters, Charismatic/Value-Based, Team-
dimension. Self-Protective actions would be Oriented, and Participative leadership are viewed
viewed very negatively. as positive contributors to outstanding leader-
ship. The Autonomous and Self-Protective CLT
Sub-Saharan Africa Cluster. Namibia, Nigeria, dimensions are slightly negative, and therefore
South Africa (Black sample), Zambia, and seen as slightly impeding effective leadership.
21-House.qxd 3/19/2004 5:08 PM Page 693
Charismatic/Value-Based
7.00
5.97
5.00
3.00 5.86
3.83
1.00
3.99
5.06
Autonomous Participative
5.38
Humane Oriented
is the highest score for all clusters. Autonomous in contributing to outstanding leadership.
leadership is reported to be a neutral factor. Self- Participative leadership is viewed positively;
Protective leadership is viewed as slightly nega- this is the only cluster in which the Participative
tive to almost neutral and is, therefore, less CLT score exceeded that of the Team-Oriented
negative than indicated in all other clusters. The CLT score. It is almost identical in importance
absolute score for the Self-Protective CLT to Charismatic/Value-Based leadership. The
dimension is the highest of all clusters and sim- Autonomous CLT dimension is viewed in a
ilar to the Middle East and Confucian Asia clus- neutral manner but Self-Protective leadership
ters. The ranking of clusters presented in Table is viewed negatively. The relative ranking of
21.6 places the Southern Asia cluster among the clusters (Table 21.6) places the Germanic
highest ranks for Charismatic/Value-Based, Europe cluster in the highest rank for
Humane-Oriented, and Self-Protective leader- Participative leadership and among the lowest
ship dimensions. for Self-Protective leadership.
Examining the relative CLT scores in Table Examining the relative CLT scores in
21.7 also reveals differences between this cluster Table 21.7 further reveals differences between
and others. When comparing relative CLT lead- the Germanic Europe cluster and others. When
ership dimension scores with those of other clus- comparing this cluster’s relative CLT scores
ters, Team-Oriented leadership is very high in with those of other clusters, the Germanic
the hierarchy for this cluster as is the Humane- Europe cluster profile is striking in its pattern of
Oriented CLT dimension, but Participative lead- leadership dimensions thought favorable for
ership is not. The Southern Asia cluster is also in effective leadership. The Charismatic/Value-
the highest group for the Self-Protective CLT, Based, Participative, and Autonomous CLTs
which also supports the finding for the absolute score highly in the leadership hierarchy for this
scores. Overall, considering both measures cluster. The Self-Protective CLT dimension is
(Table 21.8), one may note the close correspon- one of the most undesirable leadership dimen-
dence between the CLT leadership dimension sions in the hierarchy for this cluster in com-
profile for the Southern Asia cluster and the parison to other clusters’ rankings. Note that
Confucian Asia cluster profiles. Both profiles Charismatic/Value-Based leadership is about
are characterized by high Team-Oriented, equal in importance to that given for the Anglo
Humane-Oriented, and Self-Protective CLT cluster. Most striking is that the Participative
dimensions, and relatively low Participative CLT dimension is scored highest for all clusters
leadership. Charismatic/Value-Based leadership for both absolute and relative measures.
is higher in importance for the Southern Asia According to this CLT profile (Table 21.8),
cluster than for the Confucian Asia cluster as people in the Germanic Europe cluster would
reflected in both the relative and absolute scores. seek out Charismatic/Value-Based leaders who
The profile of an effective Southern Asia leader believe in Participative leadership but who also
would be a person who exhibits Charismatic/ support independent thinking yet reject elements
Value-Based, Team-Oriented, and Humane- of Self-Protectiveness.
Oriented leadership attributes. The person would
be relatively high on Self-Protective behaviors, Middle East Cluster. This cluster includes
and would not be noted for high levels of Egypt, Kuwait, Morocco, Qatar, and Turkey.
Participative leadership. There are a number of striking differences in
comparison to other clusters. Using the absolute
Germanic Europe Cluster. The Germanic scores given in Table 21.5, Table 21.6, and
Europe cluster includes Austria, Germany (for- Figure 21.10, we note that although the CLT
mer GDR–East), Germany (former FRG–West), leadership dimensions viewed as contributing to
the Netherlands, and Switzerland. In examining outstanding leadership include Charismatic/
the CLT leadership dimensions in Table 21.5 Value-Based and Team-Oriented leadership,
and Figure 21.9 (absolute scores), we note that these dimensions have the lowest scores and
Charismatic/Value-Based and Team-Oriented ranks relative to those for all other clusters.
leadership are reported to be very important Participative leadership is viewed positively,
21-House.qxd 3/19/2004 5:08 PM Page 695
Charismatic/Value-Based
7.00
5.93
5.00
3.00
5.62
3.03
1.00
5.86
4.16
Autonomous Participative
4.71
Humane Oriented
Countries: Austria, Germany (former GDR East), Germany (former FRG West), the Netherlands, and Switzerland
but again scores low compared with other in which the Middle East cluster has the highest
clusters’ absolute score and ranks. Humane- relative CLT score. Thus, the relative scores sup-
Oriented leadership is perceived positively, but port the findings with regard to absolute scores.
only about equally to other cluster scores. The This similarity is pointed out particularly for this
Self-Protective CLT dimension is viewed as cluster given the very low absolute scores. The
an almost neutral factor; however, it has the possibility of response biases affecting the find-
second-highest score and rank of all clusters. ings is always a possibility whenever absolute
The relative CLT scores paint a picture that is scores are used to compare cultural data.
consistent with the absolute scores described Clearly, this CLT profile is the most unusual
above. When comparing these relative CLT among all 10 clusters. Two explanations,
leadership scores with other clusters’ scores, although speculative, come to mind. First,
almost all Middle East CLT scores rank at the Middle Easterners may not require the same
low end of the leadership comparisons. The amount of leadership from their leaders as do
exception is the Self-Protective CLT dimension, other clusters. However, this is unlikely, which
21-House.qxd 3/19/2004 5:08 PM Page 696
Charismatic/Value-Based
7.00
5.35
5.00
3.00
3.79 5.47
1.00
3.68
4.97
Autonomous Participative
4.80
Humane Oriented
leads to the alternative explanation that critical profile. These additional leadership dimensions
leadership attributes for this cluster were not were labeled familial, humble, and faithful. It is
part of the GLOBE attribute list. This alterna- likely that the pervasive influence of the Islamic
tive explanation gains credence from the find- religion is a key to understanding the Arab
ings of GLOBE researchers who administered a world, and presumably leadership in the Arab
version of the research instrument in the Middle world (Hagan, 1995).
East containing additional leadership attributes We believe that for an accurate portrayal
not found in the final GLOBE-administered of the Middle East cluster, it is particularly
questionnaire (Dastmalchian, Javidan, & Alam, important to consider both the relative and
2001). Somewhat fortuitously, factor analyses absolute measures simultaneously (Tables 21.6,
of their data indicated leadership dimensions 21.7, and 21.8). That is, even with lower absolute
similar to the six GLOBE dimensions, but in and relative CLT scores for most of the leader-
addition there were several more that may ship dimensions, an effective leader still exhibits
help explain the enigma of the Middle East CLT Charismatic/Value-Based and Team-Oriented
21-House.qxd 3/19/2004 5:08 PM Page 697
Table 21.9 Amount of Variance Accounted for in CLTs by Societal Culture and Societal Cluster
NOTE: All ANOVAs using societal cultures and societal clusters as independent variables were significant at p < .001 for
each of the six CLT leadership dimensions.
a Numbers are eta squares that reflect amount of variance accounted for.
leadership dimensions as well as Participative However, given that clusters vary in leadership
and Humane-Oriented leadership, but not nearly profiles, which implies that differences among
to the extent indicated for other clusters. As cultures and individuals within cultures do not
pointed out in previous culture dimension overwhelm cluster differences, we are justified
chapters, Middle East respondents expect a in thinking of clusters as viable entities that
visionary and future-oriented leader to direct fol- reveal interesting leadership prototypes across
lowers to embrace a more performance-oriented the world.
culture. In addition to the culture-common find- Our findings raise several important questions
ings, we see culture-unique elements of a more for future research. For instance, if CLT leader-
traditional leadership profile in the Middle East ship dimensions are differentially endorsed
countries that include the endorsement of leader- among nations or cultures as contributors to effec-
ship attributes such as familial, humble, faithful, tive leadership (as indicated by our analyses), are
self-protective, and considerate. these CLT leadership dimensions equally com-
pelling and influential? Do the behaviors of
leaders typically reflect the CLT profiles of their
Summary of Hypotheses 2a and 2b:
cultures? If the CLT profiles of the cultures are
Culturally Endorsed Implicit Leadership
not enacted by the leader, will the leader be less
Theory and Future Research.
accepted? Less effective? Project GLOBE Phase
The previously described results support the 3, which is currently under way, was designed to
belief that clusters can be viewed as coherent answer these questions. Even for CLT leadership
entities of cultures. Each cluster is characterized dimensions found to be equally influential among
by a specific CLT configuration of leadership cultures, results from GLOBE country-specific
dimensions that vary in importance. Further, the studies (to be published in 2004) along with con-
concept of shared mental models is a useful the- siderable cross-cultural research suggest we
oretical mechanism by which cultures and clus- should expect considerable variability in how
ters develop CLTs. We also found that cultures managerial leadership is actually enacted across
and clusters capture differing amounts of vari- cultures. Researchers are also likely to investigate
ability in CLT leadership dimensions depending the mechanisms by which psychological and
on the specific CLT profile in question, with more sociological processes link the six CLT leadership
variance accounted for by individual cultures dimensions to dominant cultural values. One
than combined clusters (see Table 21.9). We neglected area of inquiry in cross-cultural studies
realize that not only is there considerable concerns the degree of variability within societies.
variability among cultures, there also is varia- For GLOBE, this issue might be captured by the
bility among individuals within each society and question, “Are CLT leadership dimensions more
that the clusters mask some of this variability. rigidly set for homogeneous societies, such as
21-House.qxd 3/19/2004 5:08 PM Page 698
Japan, than for culturally diverse societies, such as culture and leadership. This link was stated in
the United States?” Hypotheses 3 and 4. A critical issue concerns the
We also have a great deal to learn about strength of the relationships between culture and
gender differences across cultures. In the CLT leadership dimensions.
Gender Egalitarian chapter (14) by Emrich and Hypothesis 3 asserts that there will be sig-
coauthors, we demonstrated statistically signi- nificant positive relationships between societal
ficant, but not substantial, gender differences cultural dimensions (e.g., Individualism and
with Charismatic/Value-Based, Participative, Collectivism) and conceptually related CLT
Team-Oriented, and Self-Protective leadership leadership dimensions (e.g., Team Oriented).
CLT dimensions, but not with Autonomous or Hypothesis 4 asserts there will be significant
Humane-Oriented leadership CLTs. Female positive relationships between organizational
managers rated the first three leadership CLT culture dimensions and conceptually related
dimensions higher than did the male managers CLT leadership dimensions. The second column
for contributing to outstanding leadership, and in Table 21.10 shows our predictions concerning
Self-Protective leadership as more inhibiting to the relationships between the nine GLOBE-
outstanding leadership than did male managers. specific cultural dimensions and the six CLT
Gender differences were more apparent in cer- leadership dimensions. These predictions were
tain cultures than in others, indicated by a sig- generated by examining each cultural dimension
nificant gender-by-nation (i.e., societal culture) and CLT leadership dimension scale and identi-
interaction for all but the Humane-Oriented fying the conceptual similarities or likely causal
leadership CLT dimension. For instance, gender relationships among them. For example, the per-
differences for the Team-Oriented CLT were ceived effectiveness of the Charismatic/Value-
much smaller in the United States than in Hong Based leadership dimension was expected to
Kong or Guatemala. A dissertation by Paris be associated with the Performance Orientation
(2003) explores the antecedent cultural mecha- cultural dimension, because high performance
nisms that lead to these gender differences: expectations and goal attainment are defining
Gender differences were moderated by the level characteristics of charismatic/value-based leaders
of gender egalitarianism in the culture and were and performance-oriented cultures.
minimized in gender egalitarian cultures. Yet, We tested Hypotheses 3 and 4 using hierar-
we know very little about gender differences chical linear modeling (HLM), a procedure
with respect to the effectiveness of these leader- (discussed by Hanges, Dickson, & Sipe in
ship dimensions among nations. Clearly the Chapter 11) that identifies the total amount of
wealth of information provided by this phase of variance in a dependent variable that is
Project GLOBE research raises interesting accounted for by forces at the individual, organi-
questions regarding gender differences. zational, industrial, and societal levels. This pro-
cedure also allows us to determine the amount of
Hypotheses 3 and 4: Linking variance accounted for at each of these levels of
Leadership and Culture—Results analysis considered independently. That is,
of Hierarchical Linear Modeling GLOBE research is primarily concerned with
how cultural values and practices influence CLT
As just described, evidence indicates that CLT profiles at the organizational and societal levels.
leadership dimensions are associated with cul- Because GLOBE is focused on shared leader-
tures and clusters in unique combinations. These ship belief systems, the individual level of analy-
findings are consistent with the hypothesis that sis is not of direct concern for the present
cultural differences strongly influence important chapter. There are four critical points to keep in
ways in which people think about leaders, as well mind while discussing these results.
as the societal norms concerning the status, influ- First, the total amount of CLT leadership
ence, and privileges granted to leaders. However, dimension variance accounted for is a joint func-
to use a culturally based explanation for these tion of forces at the individual-, organizational-,
results instead of a nation-based explanation, it industrial-, and societal-level cultural values.
is necessary to determine the link between However, this chapter is mostly concerned with
(Text continues on page 701)
21-House.qxd 3/19/2004 5:08 PM Page 699
Table 21.10 A Priori Hypotheses and Results Predicting CLT Leadership Style From Societal and
Organizational Culture
HLM Tests
CLT Predicted Single Culture Dimension Multiple Culture Dimension
Leadership & Actual Culture Societal Organizational Societal Organizational
Dimension Dimensions Value Practice Value Practice Value Practice Value Practice
Charismatic/ Performance −.48** −.60** .42**
Value-Based Orientation
Future −.42** .17**
Orientation
Humane −.37** .17**
Orientation
Institutional −.35** .11**
Collectivism
In-Group −.41* −.69** .25* .26**
Collectivism
Assertiveness◊ −.18** −.15**
Gender .41** .36**
Egalitarianism
Power Distance −.57**
Uncertainty −.20*
Avoidance
Team Institutional −.20** .13** .12**
Oriented Collectivism
In-Group .17** −.47** .18**
Collectivism
Humane −.35** −.10* .13** −.12**
Orientation
Assertiveness◊ −.14** −.16** −.15**
Uncertainty −.12* −.12* −.17** −.06* .19** .14** −.07**
Avoidance*
Performance −.31** .15*
Orientation
Future −.15* −.28** .13**
Orientation
Gender .10** .07**
Egalitarianism
Power .23*
Distance
Participative Assertiveness◊◊ −.13** −.14* −.12**
Power Distance◊◊ −.85** −.35* −.32**
Humane −.62* −.32** .21**
Orientation
Uncertainty◊◊ −.49** −.13** −.08* −.23** −.10**
Avoidance◊◊
(Continued)
21-House.qxd 3/19/2004 5:08 PM Page 700
HLM Tests
HLM Tests
NOTE: Numbers in each column are unstandardized HLM coefficients. Predicted dimensions are noted in bold. Dimensions
followed by “◊” are hypothesized to be inversely related to leadership factors.
* p < .05.
** p < .01.
the total CLT variance and the variance single and multiple HLM results. The results may
accounted for by forces at the organizational differ somewhat depending on which perspective
and societal levels. is used—noncompetitive or competitive tests.
Second, the following HLM results reflect It is for this reason that Figures 21.11 to 21.16
competitive tests among the nine cultural differ from those in the individual dimension
dimensions, showing the specific efficacy of a chapters (12–19) as well as from the figures in
cultural dimension in relation to the other cul- the Nontechnical Summary (Chapter 3).
tural dimensions in predicting CLT leadership Third, results of HLMs based on multiple
dimension scores. We will refer to the competi- dimensions of culture indicated that endorsement
tive tests as the multiple HLM tests because of CLT dimensions is most associated with the
multiple culture dimensions are included in the cultural or organizational values orientation
analysis. Please refer to the individual culture (Should Be questionnaire responses), and less so
dimension chapters (Chapters 12–19) for a dis- with cultural or organizational practices (As Is
cussion of the single HLM tests related to CLT questionnaire responses). This makes sense given
leadership dimensions. In a sense, this competi- that values reflect an idealized state of what
tive multiple-HLM procedure is analogous to a should be and, therefore, ought to correspond to
multiple-regression procedure using several individuals’ implicit beliefs regarding idealized
independent variables, but with the addition of leadership attributes. Thus, we are much more
multiple levels of analysis. However, readers confident that the cultural dimension values are
are provided a complete picture of all HLM more important than practices—they are more
analyses in Table 21.10, which shows both the conceptually and theoretically related to CLT
21-House.qxd 3/19/2004 5:08 PM Page 702
Charismatic/Value-Based Leadership
Figure 21.11 Cultural Value Drivers of the Charismatic/Value-Based CLT Leadership Dimension
* Only statistically significant relationships are shown. The most important cultural dimensions for this leadership CLT are
bolded (i.e., relationship is significant at both society and organization levels of analyses or highest HLM coefficient within
each level of analysis).
CLT = culturally endorsed leadership theory
O = Organizational level
S = Societal level
A second major finding, certainly less style to embody performance-oriented skills and
expected than the first regarding Performance behaviors, and to be composed of visionary,
Orientation, was that In-Group Collectivism inspirational, self-sacrificial, integrity-based,
values (Collectivism II) is an important predictor and decisive attributes. This important leader-
of Charismatic/Value-Based leadership for both ship style can be demonstrated in several ways
societal and organizational levels of analysis. as described in the Den Hartog and colleagues
Other cultural values positively related to this (1999) article. The forthcoming GLOBE book
CLT include Gender Egalitarianism, Future consisting of country-specific chapters will fur-
Orientation, and Humane Orientation dimen- ther describe how certain leadership behaviors,
sions. Assertiveness practices were seen as an including Charismatic/Value-Based leadership,
impediment to effective leadership; however our are enacted differently among cultures. The
discussion throughout this section concentrates on compelling results for this CLT dimension
cultural values.7 should allow us to expect that members of
organizations and societies reporting high per-
Cultures Endorsing Charismatic/Value-Based formance-oriented, collectivist (in-group), and
Leadership. Respondents in societies report that gender egalitarian values are likely to have lead-
Charismatic/Value-Based behaviors contribute to ership prototypes that particularly emphasize
effective leadership: It is universally perceived as Charismatic/Value-Based leadership attributes.
important, yet there is variability among cul-
tures. The Charismatic/Value-Based leadership
Team-Oriented Leadership
scale values reported by a culture were best pre-
dicted by the levels of Performance Orientation Multiple Culture Dimension HLM Tests. As
and In-Group Collectivism cultural values. predicted, the In-Group Collectivism and
Absolute and relative CLT scores yield interest- Humane Orientation cultural value dimensions
ing similarities and differences among findings were significantly and positively related to the
for Charismatic/Value-Based leadership. The endorsement of the Team-Oriented CLT dimen-
mean of absolute CLT scores for all cultures sion as contributors to effective leadership at the
exceeds 5.00, an important element of identify- organizational level of analysis (Table 21.10
ing this as a universally endorsed CLT leader- and Figure 21.12). We also found that
ship dimension. In addition, all clusters except Performance Orientation and Future Orientation
Confucian Asia and Middle East have mean values were significant predictors at the organi-
absolute scores above 5.70 for this leadership zational level. Unexpectedly, Uncertainty
dimension; however, clusters do differ some- Avoidance values were particularly important
what regarding the importance of this CLT as they were significant at both the societal and
dimension in contributing to effective leadership organizational levels of analysis. Although a
(see Tables 21.6, 21.7, and 21.8). number of cultural dimension practices were
When considering the CLT scores, clusters also related to this leadership dimension, the
such as the Anglo cluster particularly endorse interpretation of these results is more problem-
Charismatic/Value-Based attributes as contribut- atic. We are much more confident that the cul-
ing to effective leadership. In fact, more than tural dimension values are more important than
half of the clusters have high scores for this CLT practices—they are more conceptually and
dimension. Members of societies and organiza- theoretically related to CLT dimensions.8
tions who value performance improvement and
have ambitious goals that demand excellence are Cultures Endorsing Team-Oriented Leadership.
more likely to accept and expect leaders who This CLT leadership dimension is universally
enact value-based charismatic leader behaviors. perceived to be important in enhancing effective
It should be stressed, however, that the GLOBE leadership. All cultures had absolute scores
view of charismatic leadership is not synony- exceeding 5.00, and most exceeded 5.60.
mous with the notions portrayed in the popular However, when examining the results using both
press—being flamboyant, glib, and attractive. absolute and relative CLT measures, the Southern
Instead, we empirically found this leadership Asia, Confucian Asia, Eastern Europe, and Latin
21-House.qxd 3/19/2004 5:08 PM Page 704
Team-Oriented Leadership
Figure 21.12 Cultural Value Drivers of the Team Oriented CLT Leadership Dimension
* Only statistically significant relationships are shown. The most important cultural dimensions for this leadership CLT are
bolded (i.e., relationship is significant at both society and organization levels of analyses or highest HLM coefficient within
each level of analysis).
CLT = culturally endorsed leadership theory
O = Organizational level
S = Societal level
Participative Leadership
• Participative
Uncertainty Avoidance (S,O) • Autocratic
(reverse scored)
Assertiveness (O)
Figure 21.13 Cultural Value Drivers of the Participative CLT Leadership Dimension
* Only statistically significant relationships are shown. The most important cultural dimensions for this leadership CLT are
bolded (i.e., relationship is significant at both society and organization levels of analyses or highest HLM coefficient within
each level of analysis).
CLT = culturally endorsed leadership theory
O = Organizational level
S = Societal level
are very strong considering the summary picture confirms preconceived notions about these
of cultural values leading to Participative cultures (Bass, 1990). To sum up, members of
leadership. Clearly, the two most important societies reporting high performance oriented
cultural variables positively related to Partici- and gender egalitarian values who also have a
pative leadership were Performance Orientation high tolerance for uncertainty will likely have
and Gender Egalitarianism, in which the most Participative leader attributes as part of their
important cultural value negatively related was effective CLT leadership prototype.
Uncertainty Avoidance. (Because the Power
Distance single HLM results were so com-
Humane-Oriented Leadership
pelling, they are included as a relevant negative
predictor in the Nontechnical Summary.)9 Multiple Culture Dimension HLM Tests. One
obvious prediction was that the Humane-
Cultures Endorsing Participative Leadership. Oriented CLT leadership dimension should be
Overall, Participative leadership was viewed related to the endorsement of the Humane
favorably by respondents from all cultures, Orientation cultural dimension values. We also
but their endorsement of the Participative predicted that Gender Egalitarianism values
CLT leadership dimension as contributing to would be related to Humane Oriented leader-
effective leadership also varied considerably ship. The former, but not the latter, prediction
among cultures. The Germanic Europe, Anglo, proved to be true: Humane Orientation cultural
and Nordic Europe clusters reported in the values were a strong predictor of Humane-
GLOBE survey were particularly supportive of Oriented leadership for organizations. Gender
Participative leadership. There was complete Egalitarianism values were not. We also found
consensus between the absolute and relative that Performance Orientation, Future Orienta-
scores for this leadership dimension. Given the tion, and Uncertainty Avoidance were all posi-
popularity of Participative leadership in the tively related to Humane Oriented leadership
Western and European leadership literature, this (see Table 21.10 and Figure 21.14).10
21-House.qxd 3/19/2004 5:08 PM Page 706
Humane-Oriented Leadership
Figure 21.14 Cultural Value Drivers of the Humane-Oriented CLT Leadership Dimension
* Only statistically significant relationships are shown. The most important cultural dimensions for this leadership CLT are
bolded (i.e., relationship is significant at both society and organization levels of analyses or highest HLM coefficient within
each level of analysis).
CLT = culturally endorsed leadership theory
O = Organizational level
S = Societal level
Cultures Endorsing Humane-Oriented Leader- expect that members of societies and organiza-
ship. As a group, all clusters report this CLT tions that value humane orientation and reduc-
leadership dimension as slightly positive in con- tions in uncertainty will likely have CLT
tributing to effective leadership, yet the ranking leadership prototypes that emphasize humane
indicates that four clusters may be singled out. attributes.
Aggregated responses in the Southern Asia,
Anglo, Sub-Saharan Africa, and Confucian Asia
Autonomous Leadership
clusters particularly endorse this characteristic as
enhancing effective leadership. Multiple Culture Dimension HLM Tests. Two
As expected, the most important cultural cultural dimensions were related to Autono-
dimension predicting the Humane-Oriented mous leadership. As predicted, collectivism
leadership dimension is the Humane Orien- values (specifically Institutional Collectivism)
tation cultural dimension. These societies and were negatively related to the Autonomous
organizations are reported to value attributes leadership CLT at both the societal and organi-
such as concern, sensitivity, friendship, toler- zational levels of analysis. In addition, Per-
ance, and support for others that are extended formance Orientation values were positively
at the societal and organizational levels. related to Autonomous leadership (see Table
Obviously, the Humane Orientation of societies 21.10 and Figure 21.15).11
is closely related to an overall concern about
the well-being of their members. This, in turn, Cultures Endorsing Autonomous Leadership.
contributes to the endorsement of the Humane- Autonomous leadership attributes were gener-
Oriented leadership dimension leading to effec- ally viewed within each cluster as being neutral
tive leadership. Examining the two strongest to slightly negative with respect to contributing
cultural dimensions predictors, we should to or impeding effective leadership. The
21-House.qxd 3/19/2004 5:08 PM Page 707
Autonomous Leadership
Autonomous
+ Leadership
Performance Orientation* (O) CLT
• Individualistic
• Independent
• Autonomous
Institutional Collectivism (S,O)
• Unique
Figure 21.15 Cultural Value Drivers of the Autonomous CLT Leadership Dimension
* Only statistically significant relationships are shown. The most important cultural dimensions for this leadership CLT are
bolded (i.e., relationship is significant at both society and organization levels of analyses or highest HLM coefficient within
each level of analysis).
CLT = culturally endorsed leadership theory
O = Organizational level
S = Societal level
absolute and relative scores converge in are often associated with face saving and status
portraying the Sub-Saharan Africa, Middle consciousness, both of which are elements of
East, Latin Europe, and Latin America clusters the self-protective leadership dimension. In
as rejecting the Autonomous CLT as a contrib- addition, being self-protective is one means to
utor to effective leadership. In contrast, the reduce uncertainty.12
Eastern Europe and Germanic Europe clusters
were the two highest-ranking clusters for this Cultures Endorsing Self-Protective Leadership.
CLT. In sum, members of societies and organi- Almost all cluster respondents viewed the Self-
zations with high performance-oriented and Protective CLT dimension as an impediment to
individualistic values will likely have auto- effective leadership—some strikingly so as
nomous attributes as part of their effective CLT reflected by the low absolute and relative
leadership prototype. scores for the Anglo, Germanic Europe, and
Nordic Europe clusters, yet the Confucian Asia
and Southern Asia clusters viewed Self-
Self-Protective Leadership
Protective leadership in an almost neutral man-
Multiple Culture Dimension HLM Tests. HLM ner (with some attributes of this factor being
results supported most predictions related to viewed positively, such as face saving).
Self-Protective leadership. Power Distance and Richard Brislin (personal communication,
Uncertainty Avoidance values were strong pos- 2000) suggested that the concept for Asian cul-
itive predictors of this CLT, whereas In-Group tures actually reflects “group-protective” rather
Collectivism and Performance Orientation val- than “self-protective” elements and, therefore,
ues were negatively related to this CLT (see would be viewed more positively in the
Table 21.10 and Figure 21.16). The high power Confucian Asia and Southern Asia clusters. We
distance values and practices of Asian societies expect that Self-Protective leadership attributes
21-House.qxd 3/19/2004 5:08 PM Page 708
Self-Protective Leadership
Self-Protective
Power Distance* (S,O) + Leadership
CLT
Uncertainty Avoidance (S,O)
• Self-centered
• Status conscious
• Conflict inducer
In-Group Collectivism (S) • Face saver
• Procedural
Performance Orientation (O) −
Figure 21.16 Cultural Value Drivers of the Self-Protective CLT Leadership Dimension
* Only statistically significant relationships are shown. The most important cultural dimensions for this leadership CLT are
bolded (i.e., relationship is significant at both society and organization levels of analyses and/or highest HLM coefficient
within each level of analysis).
CLT = culturally endorsed leadership theory
O = Organizational level
S = Societal level
would be part of an effective leadership proto- we clearly found that organizational values were
type of members of organizations and societies more frequently predictive of leadership CLTs
valuing high power distance and reductions in than societal values (see Figures 21.11–21.16).
uncertainty. A likely explanation for this is that the question-
naire asked respondents to think about outstand-
Summary of Hypotheses 3 and 4 ing leadership in their organization or industry,
therefore providing a strong organizational
Hypotheses 3 and 4 are generally supported frame of reference in the minds of respondents.
with respect to the relationships between soci- Organizational leadership and culture should be
etal culture dimensions and endorsement of all more immediate, tangible, and salient in the
six CLT leadership dimensions. Overall, our minds of respondents than societal leadership
findings show that GLOBE-defined societal and culture. The reason for this approach was
and organizational cultural values have signifi- our intention to focus on business leadership and
cant influence on the content of CLT profiles; not on leadership in other endeavors, such as
societal and organizational cultural practices religious or political leadership.
generally do not. As an additional issue, recall
that Hypothesis 3 concerns relationships
between societal culture and CLT leadership MANAGERIAL IMPLICATIONS
dimensions whereas Hypothesis 4 concerns the
relationships between organizational culture and The rapid development of regional and global
CLT leadership dimensions. Although we have economic integration of nations through entities
not made an a priori assertion of the relative such as the North American Free Trade
predictive power of societal versus organiza- Agreement (NAFTA), the European Union
tional values (Hypothesis 3 vs. Hypothesis 4), (EU), and the World Trade Organization
21-House.qxd 3/19/2004 5:08 PM Page 709
(WTO) has created a need for culturally while simultaneously striving to preserve their
knowledgeable managers who can work in cultural heritage. The stability and continuing
multicultural environments. Project GLOBE significance of societal values have important
findings should be of particular value to man- implications for organizations and managers. As
agerial leaders in multinational companies that Smith (1997) points out, “Euromanagers” who
are affected by globalization. These companies want to bridge cultural gaps in Europe must
are increasing the number of their expatriate consider the full range of cultural variability
managerial leaders throughout the world within contemporary Europe. Global managers
(Cullen, 2002). Managers placed in cross- who want to work effectively across continents
cultural situations face problems associated not and cultures still require specific cultural infor-
only with rapid change in their industries, but mation to lead successfully. The knowledge of
also those associated with multicultural misun- core cultural values and associated leadership
derstandings. Although a thorough comprehen- profiles can help managers enact effective
sion of cultural values and behaviors will not management practices.
guarantee organizational success, it should pro- We suggest that knowledge of culturally
vide managers with an initial awareness and endorsed implicit leadership profiles (CLT
understanding of critical aspects of effective profiles) should be useful for organizations
leadership within a specific culture. to select, counsel, and train individuals who
Before discussing specific management will work with members of other cultures. For
practices and how these relate to GLOBE find- example, once adapted to the individual level of
ings, it may be useful to reflect on the process analysis, GLOBE results will assist in designing
of cultural change and the stability of cultural expatriate personnel assessment and selection
values. After all, if cultural values change as tools. Organizations may select expatriate man-
quickly as organizations and industries change, agers partly on the basis of how strongly their
why be concerned with particular cultural val- values, belief systems, and leadership concepts
ues if globalization eventually leads to a “one- overlap with those predominantly held in the
world managerial culture”? Stated another way, host culture. The selection process may use a
if cultural convergence is assured through simple match such as having participative
modern industrialization, globally effective leaders selected for cultures that endorse partic-
and ineffective management practices will ipative leadership. Alternatively, it may use
inevitably surface. Although this argument has sophisticated statistical procedures such as
some merit, a contrasting position of cultural profile analysis and discriminant function
stability with enduring cultural patterns is more analysis using data from a host of differing cul-
likely. Historians and social psychologists note tural dimensions. A similar matching process
the fundamentally stable nature of the values may be used for leaders who manage culturally
and beliefs of citizens in various countries diverse work teams. For instance, military and
(Inkeles, 1981; Smith & Bond, 1993). The civilian service members who enforce UN-
Project GLOBE results presented in this book, mandated peacekeeping operations should find
along with research findings from other large- it useful to understand indigenous cultural
scale studies (see House et al.’s 1997 review of dimensions and their effective leadership pro-
Hofstede’s [1980, 2001] and others’ seminal files to lead and function successfully within a
research projects) lead us to reject the culture foreign population. It seems that this informa-
convergence hypothesis, particularly in its most tion would be especially helpful to them
extreme deterministic form. Although we because they act not only in the cultural context
acknowledge that global communication, tech- of their member nation, but also have the extra
nical innovation, and industrialization can cre- burden of a military culture to uphold.
ate a milieu for cultural change, a convergence Nonetheless, empirical validation will be neces-
of cultural values is by no means assured. In sary to demonstrate the usefulness of this type
fact, cultural differences among societies may of selection tool for intercultural managerial
be exacerbated as they adapt to modernization effectiveness.
21-House.qxd 3/19/2004 5:08 PM Page 710
Selecting the right manager to become an CLT leadership dimensions may be useful
expatriate is merely the first step in improving for developing a range of training situations
the success rate for expatriate managers. Selec- likely to generate cross-cultural fit or misfit
tion is not a substitute for cross-cultural training in leader–follower relationships. For example,
or developing mutual respect for differences in respondents in Germany and Spain report two
conducting collaborative work (cf. Smith, 1997). different concepts related to the CLT dimension
GLOBE findings can be used for informing and integrity (honest, sincere, just, trustworthy). In
training managers about leadership behaviors line with the Germanic Europe leadership
and organizational practices that are viewed prototype, a German manager favors a more
as acceptable and effective and those that are autonomous and interpersonally direct approach,
viewed as unacceptable and ineffective in for- an approach that will likely collide with the
eign cultures. The particular leadership styles expectations of Latin European managers, who
and attributes that were shown to be universal, favor a team integrative and interpersonally
and those shown to be culturally specific, should less-direct approach. Thus, in situations with
be used as a starting point. Consider participa- inherent conflict between individual and group
tive leadership, which is generally viewed across interests, it is most likely that dissent will occur
cultures as a positive leadership style; neverthe- if particular leadership prototype misfits
less, there is considerable variability among become salient. It should be possible to trans-
cultures in its endorsement. Participative leader- late potential leader prototype misfits into train-
ship is reported to be a particularly effective ing scenarios relating to real-world situations.
leadership style for the Anglo, Nordic Europe, GLOBE results would lead us to predict prob-
and Germanic Europe clusters, but is much less lems; conversely, if leadership profiles match
so for the Eastern Europe, Southern Asia, cultural requirements, fewer problems will likely
Confucian Asia, and Middle East clusters. occur.
Even for cultures that endorse participative lead- For the development of cross-cultural train-
ership, the manner in which it is effectively ing, GLOBE delivers empirically grounded
enacted will likely differ in day-to-day manage- information for any combination of target
rial situations (Smith & Peterson, 1994). countries. Whereas this chapter concerns the six
Because of these culture-specific variations, it global leadership dimensions, in the future all
would be useful to develop a range of situations 21 primary (first-order) leadership dimensions
likely to generate cross-cultural misunderstand- consisting of more than 100 leadership attrib-
ing in leader–follower relationships involving utes can be analyzed and transformed into
delegation, consultation, and normal, everyday diagnostic training situations. We suggest that
decision making. In a future publication cover- understanding culturally endorsed differences in
ing culture-specific chapters, this additional hand- leadership concepts is a first step managers can
book will provide valuable detailed information take to adjust their leadership behavior to that
regarding effective leadership actions consistent required in a host society. Knowledge about
with cultural norms. particular cultural variations in leadership
GLOBE findings are relevant to the critical prototypes should help expatriate managers to
issue of how much prior training, coaching, more accurately anticipate potential problems
and actual experience in the host society is in cross-cultural business interactions. More
necessary to ensure effective cross-cultural broadly speaking, knowledge of cultural and
leadership. We believe that the question of organizational norms and practices can inform
amount will depend on the magnitude of meaningful prescriptions for cross-cultural
differences among the cultures. The cultural strategy and policy formulation, organizational
proximity between an organization and a poten- improvement interventions, human resource
tial host culture should determine the amount management practices, the design of organiza-
and type of training materials and methods tional structures and incentive and control
for cross-cultural management preparations. systems, and a multitude of business and
Ordering cultures according to the GLOBE management issues.
21-House.qxd 3/19/2004 5:08 PM Page 711
were also important predictors of Charismatic/ Orientation of societies is closely related to the
Value-Based leadership. Other cultural values overall concern about the well-being of their
positively related to this CLT dimension include members. This, in turn, contributes to the
the Future Orientation and Humane Orientation endorsement of Humane-Oriented leadership
dimensions. These findings in their totality sug- dimension values leading to effective leader-
gest that if an organization wishes to enhance ship. Less anticipated was the large influence of
charismatic/value-based leadership, they might the Uncertainty Avoidance cultural dimension
consider developing an organizational or soci- as an important predictor of a societal inclina-
etal culture that is more performance oriented, tion toward Humane-Oriented leadership.
organizational collectivistic, gender egalitarian, Similar to the endorsement of Team-Oriented
humane, and future oriented. leadership, higher levels of Uncertainty
Team-Oriented leadership was best predicted Avoidance are associated with the endorsement
by In-Group Collectivism values and Uncertainty of humane leadership.
Avoidance values. The results regarding In-Group Finally, we also found cultural dimensions
Collectivism were expected, given the concep- that were related to the two neutral-to-nega-
tual overlap between the two constructs; collab- tively perceived CLTs. Two cultural dimensions
orative team orientation would be expected in were related to Autonomous leadership. As pre-
organizations that value pride, loyalty, and cohe- dicted, Institutional Collectivism values were
siveness in their organizations. Less expected negatively related to the Autonomous leader-
was the importance of Uncertainty Avoidance— ship at both societal and organizational levels of
the more the society and organization values the analysis. In addition, Performance Orientation
reduction of uncertainty, the more they report values were positively related to Autonomous
to endorse team-oriented leadership. Humane- leadership. Regarding Self-Protective leader-
oriented, performance–oriented, and future- ship, a general impediment to effectiveness for
oriented organizations would also likely seek all clusters, Power Distance and Uncertainty
team-oriented leaders. Avoidance were strong positive predictors of
Several societal and organizational values this leadership dimension. In contrast, In-Group
were found to relate to the endorsement and, Collectivism and Performance Orientation were
conversely, rejection of Participative leadership. negatively related to it.
From the positive side, two dimensions stand To summarize, the leadership profiles identi-
out in importance: Performance Orientation and fied in the GLOBE studies should be useful in a
Gender Egalitarianism. The most important very practical sense. Foremost, the societal and
negative cultural value was Uncertainty Avoi- organizational values help delineate culture-
dance. Assertiveness and Power Distance were specific boundaries of acceptable, effective
also negatively related to the Participative CLT. leader behaviors and practices. Leaders who are
Thus, organizational members from high power aware of a culture’s values and practices can
distance and assertive organizations, and organi- make conscious, educated decisions regarding
zations that are intolerant of uncertainty, are not their leadership practices and likely effects on
likely to use participative leadership effectively. the day-to-day operations and crisis manage-
Conversely, organizational members from gender ment within an organization. Acknowledgment
egalitarian and performance-oriented organiza- and explanation from a leader to his followers
tions are likely to use participative leadership. that a customary cultural practice will be
As might be expected, the most important breached, and why, can help avoid or diminish
cultural dimension predicting the effectiveness problems and complications. We also suggest
of Humane-Oriented leadership is the Humane that knowledge of culturally endorsed implicit
Orientation cultural dimension. Members of leadership profiles of the 10 culture clusters
humane organizations value attributes such as should be useful for selecting, counseling, and
concern, sensitivity, friendship, tolerance, training individuals who work in intercultural
and support for others. Obviously, Humane environments.
21-House.qxd 3/19/2004 5:08 PM Page 713
(Continued)
21-House.qxd 3/19/2004 5:08 PM Page 714
a Switzerland (French-speaking) CLT leadership scores in this table are absolute scores
b Canada (English-speaking) aggregated to the societal level.
c South Africa (White sample)
d South Africa (Black sample)
e Germany (East): Former GDR
f Germany (West): Former FRG
variables across differing levels of analysis, such as Nevertheless, we will discuss results using both; the
assuming that characteristics or relationships that exist relative metric has the heuristic advantage of reveal-
at the societal level apply to the individual level. ing differences among societal clusters. The specific
Conversely, the reverse ecological fallacy is said to mechanics used to compute the relative CLT measure
occur if we assume that relationships at the individual are presented in Chapter 8.
level reflect relationships at the societal level. What Because the standardizing procedure maximizes
applies for individuals may or may not apply for the differences among societal clusters, we need to
groups. We argue, however, that both ecological fal- remain cognizant of the “absolute” scores that are
lacy problems can be minimized by paying careful taken directly from the responses in the question-
attention to the level of analysis issue as was discussed naire. There are leadership attributes that are seen as
in detail in Chapter 8. Further, GLOBE researchers universal facilitators and universal inhibitors to effec-
were specific in phrasing culture questions to explicitly tive leadership. Our standardized procedure focuses
refer to organizational and societal entities, in contrast on relative comparisons. We feel both are necessary
to phrasing questions with individuals as referents. to get a true picture of leadership processes across
Further, GLOBE responses represent organizational cultures. The descriptions for each culture cluster use
and societal level measures that are more than simple both the absolute (raw scores) and the relative (stan-
aggregations of individual responses. Regarding CLT dardized scores) to achieve the most meaningful
profiles, it seems sensible to expect that if individuals interpretation of qualities important for outstanding
indicate that certain characteristics lead to effective leadership.
leadership, an aggregation of responses to the societal 5. At first glance one might incorrectly conclude
level is appropriate given sufficient generalizability that because this relative measure is by definition an
or “sameness” of responses within that society. The ipsative measure, reporting high scores on one CLT
same should hold true for societal clusters. Evidence leadership dimension automatically creates lower
supporting these assertions is presented throughout the scores on the other CLTs. However, recall that we
chapter. used the entire domain of constructs and items when
3. Recall that for three nations it was important creating our ipsative scores (i.e., we used both the
to differentiate societal cultures within each nation: leadership items as well as the items assessing the
East and West Germany, White and Black samples in various cultural dimensions). Thus, all the ipsatively
South Africa, and Germanic and French Switzerland scored CLT items do not necessarily sum to zero.
constitute separate samples. Although the possibility still exists that high scores
4. An interesting outcome of the standardization on one CLT leadership dimension affected the stand-
procedure is that our newly created measure is an ing on the other dimensions, it is incorrect to conclude
ipsative measure that has a particularly important that this automatically occurred. In fact, we have
quality. It allows us to see how important a specific confidence in the validity of the relative CLT scores
CLT leadership dimension is, such as Charismatic/ given similar, but not identical, results for the
Value-Based leadership, in comparison to the hierar- absolute and relative CLT measures.
chy of leadership dimensions for that same person. 6. It is important to discuss the validity of think-
Similarly, by aggregating to the societal level, we ing of societal cultures and societal clusters as having
determine how important a specific CLT leadership “shared mental models” of leadership. Although cul-
dimension is for the society in comparison to the ture is a macro variable, there is evidence that the
hierarchy of leadership dimensions within that responses to the GLOBE scales can be meaningfully
society. The rationale for using this standardization aggregated to the macro level (i.e., societal level).
process is that even for CLT dimensions that are Similarly, we argue that individual schemas of lead-
universally endorsed such as Charismatic/Value- ership can be aggregated to yield shared mental mod-
Based leadership, some societal cultures may endorse els for a societal culture and for the societal cluster.
this leadership style more strongly than others in To consider the validity of our shared mental model
relation to the other CLT leadership dimensions. Thus, construct for the CLT data, we might ask and answer
this second procedure using standardized scores the following questions: First, do societal cultures
constitutes a relative metric, rather than an absolute differ with respect to each of the six CLT leadership
metric. We believe these types of comparisons will dimensions reported as contributing to effective lead-
reveal useful differences among societal cultures. ership? Second, do societal clusters differ with
21-House.qxd 3/19/2004 5:08 PM Page 716
respect to each of the six CLT dimensions? cluster data. However, the order of impact on specific
Obviously, it would be fruitless to examine differ- CLT leadership dimensions is roughly the same, with
ences among clusters for those CLT dimensions that a large amount of variance being captured for the Self-
do not vary among societal cultures. Third, what Protective CLT and the least for the Team-Oriented
percentage of the variance in CLT leadership dimen- and Autonomous CLT dimensions. Lastly, the third
sions is accounted for by societal cultures? The result column of Table 21.9 indicates the overlap in
will answer the question of whether societal cultures explained variance by the clusters as compared to the
are meaningfully different with respect to differences societal cultures. The size of the overlap indicates that
among cultures. Fourth, what percentage of the vari- societal clusters capture almost all of the information
ance in CLT leadership dimensions do societal clus- distinguishing societal cultures for Self-Protective
ters account for? In this case, we are investigating the leadership, but not for the Team-Oriented or
importance of societal clusters in contrast to the ques- Autonomous CLTs. This result suggests that societal
tion that reflects the importance of individual societal clusters can provide a very handy shortcut, or sum-
cultures. Fifth, what is the amount of overlap between mary, for societal cultures within the cluster for the
information provided by the societal cultures in con- Self-Protective CLT dimension, but less so for Team-
trast to the societal clusters? The latter question Oriented or Autonomous CLTs.
speaks to the issue of whether the societal clusters 7. The five cultural dimension values and prac-
capture the variance in leadership dimensions that are tices (see Figure 21.11) explain a total of 14% of
accounted for by societal differences. the variance for the Charismatic Value-Based CLT
To answer the above questions, we conducted a leadership dimension. Approximately 47.5% of this
series of one-way Analysis of Variance (ANOVA) total explained variance was associated with forces
statistical tests using societal codes and societal operating at the organizational level of analysis. The
cluster codes as the between-subject factors on remaining portion of the explained variance (52.5%)
the original individual respondent data (i.e., original was associated with forces operating at the societal
data). Our results provide strong evidence that the level of analysis. It should be noted that the total
concept of shared mental models of leadership is variance accounted for gives a very conservative
affected by societal culture. It is clear from examin- impression of the impact of the culture variables
ing Table 21.9 that all leadership CLTs are predicted because individual variation in the dependent vari-
by using societal cultures and societal clusters as able (CLT) has not been removed. Because GLOBE
independent variables. In addition, the amount of hypotheses have been conceptualized at the culture
variance accounted for, as indicated by eta squares, level of analysis, individual variation is considered
varies by leadership dimension. Also, eta-squared measurement error. Therefore we also followed the
statistics speak to the issue of the actual amount of procedure discussed by Hanges et al. in Chapter 11 to
influence of societal cultures in contrast to societal provide explained variance estimates separately at
clusters, and overlap of societal cultures versus the organizational or societal level of analysis. The
societal clusters. Results indicate that societal culture organizational R2 for the Charismatic/Value-Based
differences account for the most variance in the Self- CLT was 48.2% and the societal R2 was 44.5%. See
Protective CLT and the least variance in the Chapter 11 by Hanges et al. for a detailed discussion
Autonomous CLT. The specific order in terms of of this level-specific R2. These percentages corre-
variance accounted for by societal culture is: Self- spond to the traditional regression R2 done at each
Protective leadership followed by Participative, level of analysis separately.
Charismatic/Value-Based, Humane-Oriented, Team- 8. The five cultural dimension values and
Oriented, and Autonomous leadership. practices (see Figure 21.12) explain a total of 13.1%
As indicated in the previous paragraph, when of the variance for the Team-Oriented CLT leader-
examining how the CLTs vary according to societal ship dimension. Approximately 65.5% of this total
clusters, each CLT leadership dimension is impacted explained variance was associated with forces oper-
significantly by this between-groups societal-cluster ating at the organizational level of analysis. The
factor. The total variance accounted for, however, is remaining portion of the explained variance (34.5%)
less than when considering the effect of societal cul- was associated with forces operating at the societal
tures. This is to be expected because information is level of analysis. The organizational R2 for the Team-
lost when the societal data are collapsed into societal Oriented CLT was 54.2% and the societal R2 was
21-House.qxd 3/19/2004 5:08 PM Page 717
34.7%. See Chapter 11 by Hanges et al. for a detailed variance was associated with forces operating at
discussion of this level-specific R2. These percentages the organizational level of analysis. The remaining
correspond to the traditional regression R2 done at portion of the explained variance (91.8%) was
each level of analysis separately. associated with forces operating at the societal level
9. The four cultural dimension values and prac- of analysis. The organizational R2 for the Self-
tices (see Figure 21.13) explain a total of 15.5% of Protective CLT was 51.2% and the societal R2 was
the variance for the Participative CLT leadership 82.2%. See Chapter 11 by Hanges et al. for a detailed
dimension. Approximately 13.1% of this total discussion of this level-specific R2. These percent-
explained variance was associated with forces operat- ages correspond to the traditional regression R2 done
ing at the organizational level of analysis. The at each level of analysis separately.
remaining portion of the explained variance (86.9%)
was associated with forces operating at the societal
level of analysis. The organizational R2 for the REFERENCES
Participative CLT was 39.6% and the societal R2 was
54.3%. See Chapter 11 by Hanges et al. for a detailed Bass, B. M. (1990). Bass and Stogdill’s handbook of
discussion of this level-specific R2. These percent- leadership: Theory, research, and managerial
ages correspond to the traditional regression R2 done applications (3rd ed.). New York: Free Press.
at each level of analysis separately. Bass, B. M. (1997). Does the transactional-transfor-
10. The four significant cultural dimension val- mational leadership paradigm transcend organi-
ues and practices (see Figure 21.14) explain a total zational and national boundaries? American
of 12.7% of the variance for the Humane-Oriented Psychologist, 52(2), 130–139.
CLT leadership dimension. Approximately 45.2% of Bond, M., & Hwang, K. K. (1986). The social psy-
this total explained variance was associated with chology of the Chinese people. In M. H. Bond
forces operating at the organizational level of (Ed.), The psychology of the Chinese people (pp.
analysis. The remaining portion of the explained 213–266). Hong Kong: Oxford University Press.
variance (54.8%) was associated with forces operat- Brodbeck, F. C., Frese, M., Akerblom, S., Audia,
ing at the societal level of analysis. The organiza- G., Bakacsi, G., Bendova, H., et al. (2000).
tional R2 for the Humane-Oriented CLT was 51.3% Cultural variation of leadership prototypes
and the societal R2 was 60.5%. See Chapter 11 by across 22 European countries. Journal of
Hanges et al. for a detailed discussion of this level- Occupational and Organizational Psychology,
specific R2. These percentages correspond to the 73,1–29.
traditional regression R2 done at each level of Cullen, J. B. (2002). Multinational management: A
analysis separately. strategic approach (2nd ed.). Cincinnati, OH:
11. The two cultural dimension values and South-Western Thomson Learning.
practices (see Figure 21.15) explain a total of 3.9% of Dastmalchian, A., Javidan, M., & Alam, K. (2001).
the variance for the Autonomous CLT Leadership Effective leadership and culture in Iran: An
Dimension. Approximately 37.2% of this total empirical study. Applied Psychology: An
explained variance was associated with forces operat- International Review, 50(4), 532–551.
ing at the organizational level of analysis. The Den Hartog, D., House, R. J., Hanges, P. J., Ruiz-
remaining portion of the explained variance (62.8%) Quintanilla, S. A., Dorfman, P. W., & GLOBE
was associated with forces operating at the societal Associates. (1999). Culture specific and cross
level of analysis. The organizational R2 for the culturally generalizable implicit leadership
Autonomous CLT was 30.4% and the societal R2 theories: Are attributes of charismatic/transfor-
was 18.0%. See Chapter 11 by Hanges et al., for a mational leadership universally endorsed?
detailed discussion of this level-specific R2. These Leadership Quarterly, 10(12), 219–256.
percentages correspond to the traditional regression Dickson, M. W., Hanges, P. J., & Lord, R. G. (2001).
R2 done at each level of analysis separately. Trends, developments and gaps in cross-cultural
12. The four cultural dimension values (see research on leadership. In W. H. Mobley &
Figure 21.16) explain a total of 33.1% of the vari- M. W. McCall (Eds.), Advances in global leader-
ance for the Self-Protective CLT Leadership dimen- ship (Vol. 2, pp. 75–100). Oxford, UK: Elsevier
sion. Approximately 8.2% of this total explained Science.
21-House.qxd 3/19/2004 5:08 PM Page 718
Dorfman, P. W. (1996). International and cross- Gupta, V., et al. (1999). Cultural influences on
cultural leadership research. In B. J. Punnett & leadership and organizations: Project GLOBE.
O. Shenkar (Eds.), Handbook for international In W. F. Mobley, M. J. Gessner, & V. Arnold
management research (pp. 267–349). Oxford, (Eds.), Advances in global leadership (Vol. 1,
UK: Blackwell. pp. 171–233). Stamford, CT: JAI.
Dorfman, P. W. (2004). International and cross- House, R. J., Wright, N. S., & Aditya, R. N. (1997).
cultural leadership research. In B. J. Punnett & Cross-cultural research on organizational lead-
O. Shenkar (Eds.), Handbook for international ership: A critical analysis and a proposed theory.
management research (2nd ed., pp. 265–355). In P. C. Earley & M. Erez (Eds.), New perspec-
Ann Arbor: University of Michigan Press. tives in international industrial/organizational
Duncan, P. C., Rouse, W. B., Johnston, J. H., psychology (pp. 535–625). San Francisco: New
Cannon-Bowers, J. A., Salas, E., & Burns, J. J. Lexington Press.
(1996). Training teams working in complex Howell, J. P., & Costley, D. L. (2001). Understanding
systems: A mental model-based approach. behaviors for effective leadership. Upper Saddle
Human/Technology Interaction in Complex River, NJ: Prentice Hall.
Systems, 8, 173–231. Inkeles, A. (1981). Convergence and divergence in
Gerstner, C. R., & Day, D. V. (1994). Cross-cultural industrial societies. In M. O. Attir, B. Holzner,
comparison of leadership prototypes. Leader- & Z. Suda (Eds.), Directions of change:
ship Quarterly, 5(2), 121–134. Modernization theory, research and realities
Hagan, C. M. (1995). Comparative management: (pp. 3–38). Boulder, CO: Westview.
Africa, the Middle East, and India (Working Keller, T. (1999). Images of the familiar: Individual
Paper). Boca Raton: Florida Atlantic University differences and implicit leadership theories.
Press. Leadership Quarterly, 10(4), 589–607.
Hains, S. C., Hogg, M. A., & Duck, J. M. (1997). Klimoski, R., & Mohammed, S. (1994). Team mental
Self-categorization and leadership: Effects of model: Construct or metaphor? Journal of
group prototypicality and leader stereotypical- Management, 20, 403–437.
ity. Personality and Social Psychology Bulletin, Kraiger, K., & Wenzel, L. H. (1997). Conceptual
23(10), 1087–1099. development and empirical evaluation of mea-
Hanges, P. J., Higgins, M., Dyer, N. G., Smith-Major, sures of shared mental models as indicators of
V., Dorfman, P. W., Brodbeck, F. C., et al. team effectiveness. In M. T. Brannick, E. Salas,
(2001, April). Influence of cultural values on & C. Prince (Eds.), Team performance assess-
leadership schema structure. Paper presented ment and measurement (pp. 63–84). Mahwah,
at the 16th Annual Society of Industrial and NJ: Lawrence Erlbaum.
Organizational Psychology Conference, San Lord, R. G., Foti, R. J., & DeVader, C. L. (1984). A
Diego, CA. test of leadership categorization theory: Internal
Hanges, P. J., Lord, R. G., & Dickson, M. W. (2000). structure, information processing, and leadership
An information processing perspective on leader- perceptions. Organizational Behavior and Human
ship and culture: A case for connectionist archi- Performance, 34, 343–378.
tecture. Applied Psychology: An International Lord, R. G., & Maher, K. J. (1991). Leadership and
Review, 49, 133–161. information processing: Linking perceptions
Hofstede, G. (1980). Culture’s consequences: Inter- and performance (Vol. 1). Cambridge, MA:
national differences in work-related values. Unwin Hyman.
London: Sage. Mohammed, S., & Dumville, B. (2001). Team men-
Hofstede, G. (2001). Culture’s consequences: Comparing tal models in a team knowledge framework:
values, behaviors, institutions, and organizations Expanding theory and measurement across dis-
across nations. Thousand Oaks, CA: Sage. ciplinary boundaries. Journal of Organizational
House, R. J., & Aditya, R. N. (1997). The social Behavior, 22(2), 80–106.
scientific study of leadership: Quo vadis? Journal Mohammed, S., Klimoski, R., & Rentsch, J. R.
of Management, 23(3), 409–473. (2000). The measurement of team mental mod-
House, R. J., Hanges, P. J., Ruiz-Quintanilla, S. A., els: We have no shared schema. Organizational
Dorfman, P. W., Javidan, M., Dickson, M., Research Methods, 3(2), 123–165.
21-House.qxd 3/19/2004 5:08 PM Page 719
Nye, J. L., & Forsyth, D. R. (1991). The effects of Shaw, J. B. (1990). A cognitive categorization model
prototype-based biases on leadership appraisals: for the study of intercultural management.
A test of leadership categorization theory. Small Academy of Management Review, 15(4),
Group Research, 22(3), 360–379. 626–645.
O’Connell, M. S., Lord, R. G., & O’Connell, M. K. Smith, P. B. (1997). Cross-cultural leadership: A path
(1990, August). Differences in Japanese and to the goal? In P. C. Earley & M. Erez (Eds.), New
American leadership prototypes: Implications perspectives on international industrial/organiza-
for cross-cultural training. Paper presented at the tional psychology (pp. 626–639). San Francisco:
meeting of the Academy of Management, San New Lexington Press.
Francisco. Smith, P. B., & Bond, M. H. (1993). Social psychol-
O’Neill, O. A., & Hanges, P. J. (2001). Individual ogy across cultures: Analysis and perspectives.
values and the structure of leadership schemas: London: Harvester Wheatsheaf.
Verification of a connectionist network (Working Smith, P. B., & Peterson, M. K. (1994 July).
Paper). University of Maryland, College Park. Leadership as event-management: A cross-
Paris, L. (2003). The effect of gender and culture on cultural survey based upon middle managers
implicit leadership theories: A cross-cultural from 25 nations. Paper presented at the Cross-
study. Unpublished doctoral dissertation, New Cultural Studies of Event Management at the
Mexico State University, Las Cruces. International Congress of Applied Psychology,
Rentsch, J. R., & Klimoski, R. (2001). Why do “great Madrid, Spain.
minds” think alike? Antecedents of team Triandis, H. C. (1994). Cross cultural industrial and
member schema agreement. Journal of Organi- organizational psychology. In H. C. Triandis,
zational Behavior, 22(2), 107–120. M. D. Dunette, & L. M. Hough (Eds.),
Rokeach, M. (1973). The nature of human values. Handbook of industrial and organizational
New York: Free Press. psychology (2nd ed., Vol. 4, pp. 103–172).
Schein, E. H. (1992). Organizational culture and Palo Alto, CA: Consulting Psychologists.
leadership: A dynamic view (2nd ed.). San Triandis, H. C. (1995). Individualism and collec-
Francisco: Jossey-Bass. tivism. Boulder, CO: Westview.
Schwartz, S. H. (1992). Universals in the content and van de Vijver, F., & Leung, K. (1997). Methods and
structure of values: Theoretical advances and data analysis for cross-cultural research.
empirical tests in 20 countries. In M. P. Zanna Thousand Oaks, CA: Sage.
(Ed.), Advances in experimental social psychol- Yukl, G. A. (2002). Leadership in organizations
ogy (Vol. 25, pp. 1–66). New York: Academic (5th ed.). Upper Saddle River, NJ: Prentice
Press. Hall.
21-House.qxd 3/19/2004 5:08 PM Page 720
22-House.qxd 3/19/2004 5:08 PM Page 721
PART V
CONCLUSION
721
22-House.qxd 3/19/2004 5:08 PM Page 722
22-House.qxd 3/19/2004 5:08 PM Page 723
22
CONCLUSIONS AND
FUTURE DIRECTIONS
MANSOUR JAVIDAN
ROBERT J. HOUSE
PETER W. DORFMAN
VIPIN GUPTA
PAUL J. HANGES
723
22-House.qxd 3/19/2004 5:08 PM Page 724
724– • –CONCLUSION
complexities and extended the timeline for of instrument design from the beginning. The
completion, we felt that it was the only viable scales were designed to independently measure
way to report our work. cultural practices and values using a sound
The preceding chapters have provided theoretical basis, and items were developed and
substantial information on what we set out to validated against exacting standards. The instru-
do and what we have found. We have found ments used isomorphic structures at the societal
answers to many questions. This volume is filled and organizational levels of culture. We found
with a wide variety of findings, some confirming high degrees of interrater agreement within
conventional wisdom and others questioning it, cultures and a high degree of discrimination
but like any other research project our work has across societies. The levels of interitem consis-
also led to many new questions. Although we tency were well above conventional standards.
have advanced the current state of knowledge on Two independent pilot tests produced sound
several fronts, we have uncovered many new and psychometric properties.
puzzling questions that require additional research. We developed new methods for identifying
In this chapter, we provide a brief summary of and controlling for response bias and built
GLOBE’s main theoretical and methodological nomological nets to test the construct validity
contributions and present a series of new ques- of the scales. Multilevel confirmatory factor
tions to help direct future research on the important analysis was used to confirm the internal con-
issues of cross-cultural management. sistency and cross-cultural viability of the cul-
ture and leadership scales. Contextual analysis
was also performed to verify that the aggre-
MAIN FEATURES OF GLOBE’S gated phenomena were operating at the appro-
THEORETICAL FRAMEWORK priate organizational and societal levels. In
essence, we addressed the reverse ecological
GLOBE developed an integrated and cross-level fallacy error by showing that the GLOBE mea-
theory of the relationship between cultural values sures operate at higher levels than simply an
and practices and leadership, organizational, and aggregation of the individual responses. We
societal effectiveness. The theory is based on the further validated cultural value scales against
integration of four important theoretical perspec- outcropping measures and used unobtrusive
tives: The implicit leadership theory expanded to measures to validate cultural practice scales.
the cultural level, the strategic contingency The validation of the GLOBE societal-level
theory, McClelland’s achievement theory of practices and values scales using independently
human motivation, and Hofstede’s culture collected, unobtrusive, and outcropping mea-
theory. We have extended the current knowledge sures data, as well as the relationships with
base by a more comprehensive conceptualization other variables as theoretically predicted, all
of cultural dimensions and by introducing new attest to the validity of the GLOBE societal
dimensions. We further conceptualized and mea- variables and indicate that the findings reflect
sured culture in terms of practices and values. At the societies in which middle managers work
the organizational level, we introduced nine new rather than the cultures of middle managers
dimensions of organizational culture. alone.
To avoid common source bias, two forms
of the survey were developed and administered
MAIN FEATURES OF GLOBE’S to two different groups of respondents. Form
METHODOLOGICAL FRAMEWORK Alpha contained items measuring leadership
effectiveness and organizational culture. Form
The instruments developed by GLOBE resear- Beta contained items measuring leadership
chers are based on the existing literature but effectiveness and societal culture. When analyz-
they are truly cross-cultural in the sense that ing the relationship between leadership and cul-
the country co-investigators (CCIs) are either ture, we correlated the responses from one group
natives of their cultures or are very knowledge- on leadership with those from another group on
able about them. They participated in the process culture, thus preventing common source bias.
22-House.qxd 3/19/2004 5:08 PM Page 725
In total, more than 17,000 managers from 951 societies could be grouped into their predicted
organizations that disclosed their identity in three clusters on the basis of their cultural scores (see
industries participated in our surveys. Employees Chapter 10, by Gupta & Hanges). One possible
of local offices of multinational firms were explanation for the difference between Phase 1
excluded to avoid cultural contamination. These and Phase 2 results about the required number
industries were selected in a poll of the CCIs and of respondents may be the quality of scales. On
because they exist in all countries. At the same the basis of the feedback received in Phase 1
time, the three industries face different industry consisting of two pilot studies, we made signif-
contingencies. The telecommunications industry icant improvements in leadership as well as
is more severely regulated than the food industry societal culture scales, as discussed in Chapter 8
and tends to be more dynamic. The finance indus- by Hanges and Dickson. Therefore, it is likely
try entails a set of universal rules that go beyond that high quality scale measures may reduce the
individual countries to ensure global consistency. number of respondents needed to obtain accu-
Previous cross-cultural scholars have used rate estimates of the cultural-level constructs.
samples such as first-line supervisors in mar- We further show that it is appropriate to
keting and service occupations (Hofstede, generalize about the national-level cultural
1980), school students and teachers (Schwartz, constructs on the basis of a single sample of indi-
1999), and the general population (Inglehart, viduals, even in highly diverse nations where
Basanez, & Moreno, 1998) for developing multiple subcultures coexist. We studied sub-
cultural-level constructs. Our findings suggest cultures in three nations: South Africa, Germany,
that middle-level managers are also an appro- and Switzerland. In other countries, we studied
priate group for studying cultural constructs. only one sample in each nation. In the societies in
The responses of the middle-level managers which the subcultures were studied, our results
appear to be both reliable (high within-culture indicated predictable patterns of societal-cultural
response consistency, and high between-culture constructs (see Chapter 10, by Gupta & Hanges).
response variation) and valid (strong predictive In South Africa, the Black sample belonged to
relationship with convergent constructs, and the Sub-Saharan Africa cluster, whereas the
weak secondary relationship with divergent White sample belonged to the Anglo cluster. In
constructs). Germany, both former East and former West
The CCIs collected the data in their respective German samples were part of the Germanic
countries in a culturally sensitive way and, to the Europe cluster. Finally, in Switzerland, the
extent possible, sampled the dominant cultures French-speaking subculture was part of the Latin
in their countries. Furthermore, we developed Europe cluster, whereas the German-speaking
a theoretically driven comprehensive database of sub-culture belonged to the Germanic Europe
information regarding economic prosperity and cluster. Our findings affirm the significance of
the human condition in GLOBE countries. sampling and studying subcultures. At the same
Our findings, based on a pilot study in Phase time, our findings also suggest that by sampling
1 of the GLOBE study, suggested that the aver- from a dominant subculture within each society,
age number of respondents needed for a cross- one may indeed be able to predict several
cultural construct of a target reliability of 0.85 national-level behaviors (see Chapter 9, by
is 45 (see Chapter 8, by Hanges & Dickson). In Gupta, Sully de Luque, & House).
Phase 2, the number of respondents on the soci- Last, but not least, we show that it is
etal culture scale ranged from 13 (El Salvador) appropriate to use the survey data for measur-
to 300 (Iran) and as many as 895 respondents in ing cultural-level practices, values, and implicit
Sweden. Eight societies were able to sample leadership effectiveness constructs. Our find-
fewer than 45 respondents for the societal cul- ings indicate that survey data yield reliable and
ture constructs. Our estimates of societal culture valid estimates of cultural-level constructs, pro-
scales for even these societies appear to be in vided the survey is based on a sound research
the true ballpark. There is little evidence of design and a rigorous measurement instrument.
response bias in the scores of these societies To add rigor to our instrument, we used Q-sort
(see Appendix B, this volume). Further, these methodology to confirm that at least 80% of the
22-House.qxd 3/19/2004 5:08 PM Page 726
726– • –CONCLUSION
respondents across societies associate the cultural culturally contingent attributes of leadership.
items with respective constructs. In addition, we Hierarchical linear modeling (HLM) was used
requested the country co-investigators to evalu- to explore the relationships at different levels
ate each cultural item, and to report if the item of analysis using alternate data sources in a
carries the attributed meaning in their society way that eliminates common source bias. Our
and language. These “item evaluation” reports findings show the relationships among cultural
yielded some surprising findings. An item “beg- dimensions, organizational practices, and cul-
ging is banned in the city centers” was indicated turally endorsed (i.e., CLT) leadership dimen-
by most respondents to be an indicator of the sions. They also show the relationship between
Humane Orientation construct. However, in some cultural dimensions and measures of societal
societies, the respondents held that the item achievement such as economic prosperity and
connotes a high Humane Orientation, because human condition.
their society provides institutional mecha- In terms of the relationship between leader-
nisms, such as nongovernment organizations, ship and culture, we have identified the cultural
for rehabilitating beggars as a result of the dimensions that can best predict CLT dimen-
ban. In other societies, the respondents held that sions and assessed the differential strength of
the item connotes a low Humane Orientation association between each cultural dimension in
because it prevents the poor people from seek- relation to each CLT dimension. We have also
ing support from others and may even encour- determined the differential explanatory power of
age a higher crime rate. Because the item did not the organizational-level culture versus societal-
carry a common meaning across cultures, it had level culture in terms of outstanding leadership.
to be deleted from the survey to allow for cross- In terms of the linkage between societal and
cultural comparability of the scores on the organizational culture, we show a strong rela-
Humane Orientation scale. tionship between the two—organizations mirror
societies from which they originate. Most
important, the analysis demonstrating this fact
GLOBE’S MAJOR eliminated potential common source bias.
EMPIRICAL CONTRIBUTIONS Although we showed the interactive effects of
society and industry on organizational culture,
GLOBE country culture scores are reported organizational cultures seem to be more of a
using a banding technique to improve the inter- reflection of their societal context rather than
pretation of country rankings. We have identified their industry context. In fact, the industry con-
the universally desirable cultural dimensions like text seems to have limited, if any, influence on
performance orientation and the universally the assessment of societal-level cultural dimen-
undesirable dimensions like power distance. We sions, organizational-level cultural dimensions,
used the existing literature to develop a concep- and leadership attributes.
tual clustering of societies and empirically vali- Our findings also show that the difference
dated them. The clusters are the first empirical between the cultural value score and cultural
attempt to use a holistic approach to understand practice score for each dimension is larger at the
and verify cultural differences across societies. societal level than at the organizational level,
We now have cluster scores on cultural values, perhaps pointing to the broader range of vari-
practices, and implicit leadership theories. ables involved at the society level than at the
Our findings will help enhance the theoreti- organizational level. It also leads to the conclu-
cal development of the cross-cultural field. We sion that changing culture is likely to be easier at
have identified 21 primary and 6 global leader- the organizational level than at the societal level.
ship dimensions. From these, we created We further showed the differential associa-
endorsed leadership profiles across the 10 cul- tion between various cultural dimensions and
ture clusters, which, in turn, follow from the measures of societal achievement and con-
development of our culturally implicit leader- firmed a clear cultural underpinning to the way
ship theory (CLT). We have also identified uni- societies generate and distribute wealth and take
versally desirable, universally undesirable, and care of their people. As a result, any decision or
22-House.qxd 3/19/2004 5:08 PM Page 727
action to change the way governments operate leadership theory (CLT). It will study the extent
or the way societies allocate their resources has to which CEO behavior conforms to CLTs. It
to take into consideration the cultural issues and will also examine the consequences of CEO
their implications. behaviors that violate CLTs. In this book, we
have shown that culturally implicit leadership
theories are shaped by societal and organizational
FUTURE DIRECTIONS cultures. Our explanation was that leaders grow
up in their cultures and build their worldview on
The wealth of findings provided in this book
the basis of their own learning and development.
sets the stage for a more sophisticated and com-
Furthermore, they have to motivate and energize
plex set of questions. Our enhanced knowledge
employees who are also culturally conditioned.
about societal practices and values, organiza-
But what if leaders violate societal cultural norms
tional practices and values, leadership attrib-
or leadership expectations represented by a cul-
utes, and societal accomplishments enables us
ture’s CLT profile? Under what conditions do
to pursue even more complicated questions. In
they violate these norms? Can they violate the
his Foreword to the book, Triandis wrote that
norms and succeed? Which norms are more crit-
“Thousands of doctoral dissertations in the
ical for leaders—The societal norms or the
future will start with these findings.” In this
CLTs? Can they violate one and not the other?
section, we intend to speed up this process by
Although these are important academic ques-
posing a series of questions to help direct and
tions, they do have significant managerial impli-
energize further research on important issues in
cations. For instance, several GLOBE countries
cross-cultural management.
are reported to score low on the Performance
Orientation cultural dimension. Suppose a corpo-
Culture and Leadership
ration from a high performance-oriented culture
To begin with, Phase 3 of GLOBE, which decides to start an operation in a low perfor-
is currently under way, is focused on several mance-oriented country. What kind of a busi-
important leadership questions. Although ness-unit leader should they assign to this new
GLOBE findings so far have shown the various position? If they assign an executive with high
attributes of leadership, they have not identified performance orientation, he or she will be in vio-
the behavioral manifestations of such attributes. lation of the local norms and CLTs. How good
For example, it is clear that integrity is a uni- are the new leader’s chances of success? If a local
versally desirable attribute. But does it mean person with low performance orientation is
the same thing to a Chinese as it does to an appointed, can he or she succeed in achieving the
American? How do people in different cultures company’s goals? He or she may be in tune with
conceptualize, perceive, and exhibit behavior the local employees and ensure peace and stabil-
that reflects integrity? What specific behaviors ity, but can he or she deliver results? GLOBE
comprise high integrity leadership, and do they Phase 3 is designed to address these questions.
have the same function and impact across cul- Another issue relating to leadership CLTs is
tures? Can a leader’s integrity be adequately the potential confounding effect of unexpected
measured across cultures with a single survey contingencies that will likely pressure the leader
instrument? Related to this, we need to under- to operate outside his or her comfort level.
stand the nonverbal and emotional manifesta- Presumably, leaders enact CLTs because they
tions of the CLT leadership dimensions. In what have learned that acting according to CLTs is
ways other than visible behavior do leaders desirable and rewarded in their cultures. For
connect to others in their organizations? And example, a leader in a society in which the CLT
to what extent are these nuances, nonverbal emphasizes team orientation has learned over
behaviors, and emotional expressions universal time to be a team player. But a typical leader
or culturally contingent? faces a variety of situations and challenges, each
Phase 3 research will delve deeper into the with its own contingencies. For example, there
relationship between the behavior of chief exec- may be urgent challenges to attend to—teams
utive officers (CEOs) and the culturally implicit may be ineffective and slow to respond, or
22-House.qxd 3/19/2004 5:08 PM Page 728
728– • –CONCLUSION
reductions in personnel due to poor economic human needs, such as the need for achievement,
conditions could have a detrimental effect on that go beyond cultural boundaries (Fyans, Salili,
team spirit. What should a team-oriented leader Maehr, & Desai, 1983; Javidan & Carl, in press;
do in such a situation? It is important to under- McClelland, 1961).
stand the interactions between societal norms, Another possible driver of universality may
CLTs, and situation-based contingencies. How lie in ethical values. Several authors have
does a leader reconcile the broader cultural and argued that some types of leadership, namely
CLT prescriptions with the narrower and transformational leadership, are rooted in
shorter-term situation-based contingencies? strong ethical values (Bass & Steidlmeier, 1999;
Which one ends up shaping the leader’s actions Kanungo, 2001; Kanungo & Mendonca, 1996;
and decisions? And why? Mendonca, 2001). Mendonca (2001) speculated
There are several other important issues that that “good moral character is the essence of
need attention. For example, our leadership pro- every human being” (p. 270) and transforma-
files are based on factor analyses of the leader- tional leadership provides the opportunity to
ship items that were included in the instruments. “fully realize the unique and unrepeatable
Although we were meticulous in the design of potential that is in each person” (p. 270).
the leadership instrument, we make no claim On the basis of these assertions, Javidan and
that it is all encompassing and exhaustive. There Carl (in press) have proposed that
are other aspects of leadership that could be
included in the leadership model. For instance, the . . . reason for commonality of charismatic
we have learned that family and religion play a leadership across cultures may be due to its moral
bigger role for leaders in Arab countries than for and ethical foundations. It is plausible that charis-
those in other societies. As another example, we matic leadership is embedded in a universal implicit
had fewer leadership attribute items related to theory of leadership because of the human desire
Participative leadership than was desirable. for autonomy and a sense of self-control.
Additional leadership items regarding differing
aspects of participation (e.g., consultation vs. On the other hand, a desire for autonomy
delegation) would clarify the distinctions that and self-control may reflect societal norms found
likely exist among cultures regarding this lead- in high individualistic and low collectivistic
ership dimension. societies.
Another major finding of GLOBE is that Although the various explanations and specu-
there are indeed universal attributes of leader- lations for universal attributes of leadership are
ship. GLOBE’s theoretical model postulated that all plausible, none are empirically verified. It is
each culture develops its own culturally implicit important to test the various competing argu-
theory (CLT) of leadership. We empirically veri- ments and propositions and reach a clear and def-
fied a significant relationship between culture inite conclusion on the drivers of universality in
and leadership. The implication of this assertion leadership. Furthermore, if it is true that universal
is that to the extent that cultures are different, needs drive universal leadership attributes, then a
their CLTs will be different. At the same time, we related question concerns the interaction between
have also shown a set of leadership attributes that universal and cultural drivers of leadership. How
are universally desirable and universally undesir- do they interact? Which one is more important?
able. Many reasons have been suggested for the Under what conditions?
universality of leadership attributes. Common
technological imperatives (Woodward, 1958),
Measuring Cultures
common industrial logic (Adler, Doktor, &
Redding, 1986), and globalization of institutions Thus far the discussion has been centered on
and practices have all been suggested as drivers the generalizations we could make about soci-
of common management practices and processes etal culture and its influence on organizational
(Child & Tayeb, 1983). Others have suggested culture and practices, and on implicit leadership
that some leadership dimensions such as charis- theories prevalent in the culture. To the extent
matic leadership may satisfy universal and basic that the measuring instruments contain items
22-House.qxd 3/19/2004 5:08 PM Page 729
that are not behavior specific, we minimize the academic questions, they also have significant
risk of nonequivalence of measures across the managerial implications. The current cross-
cultures studied. Even so, we must recognize cultural literature advises managers on what to
the fact that two cultures with the same levels of do in high power distance cultures or in high
a cultural dimension may exhibit different collectivist cultures. What is the advice for high
behaviors associated with that dimension. This power distance and high collectivist cultures? Is
aspect is well expounded by d’Iribarne (1997). it different from the advice for high power dis-
Culture-specific behaviors are difficult to inter- tance and high performance-oriented cultures?
pret from outside a culture, except in terms In other words, which cultural dimensions, if
of the outside observer’s culture. There is a any, should be attended to most in different cul-
distinction between “knowing” a culture and tures? What is the impact of different bundles of
“internalizing” a culture: the former belongs to cultural dimensions? Which combination of cul-
the realm of rational thought, the latter to sub- tural dimensions should be studied in each cul-
jective experience. Although GLOBE provides ture? In short, although the current literature
a profile of cultural dimensions for each society, informs us about different cultural strands, it
it does not present a behavioral profile. Further fails to shed light on the cultural fabric. The
research is required to build an in-depth under- GLOBE findings presented in this book should
standing of how people actually function and help pave the way for new thinking and analysis
manifest different cultural attributes. to move us in this direction.
The current cross-cultural literature, influ- One final point about measuring cultures is
enced by Hofstede’s (1980) seminal work, has the GLOBE finding that there is a negative cor-
been focusing on cultures as a collection of cul- relation between cultural values and practices
tural dimensions. Much has been written about in seven out of nine cultural dimensions. As a
the various dimensions and their implications. typical example, cultures with high values on a
GLOBE has introduced several new cultural dimension are likely to exhibit low practices
dimensions and has studied each dimension (e.g., as seen with the Performance Orientation
individually. It is easy to suggest that no partic- scores). This is contrary to the conventional
ular set of cultural dimensions is all encompass- wisdom in the literature. Much of the writing on
ing or exhaustive. There are of course new culture suggests implicitly or explicitly that cul-
dimensions that researchers could study. tural practices are driven by cultural values and
Cultures are obviously very complex multidi- that there is a linear and positive relationship
mensional phenomena that extend beyond any between them (Hofstede, 1980; Schein, 1992).
particular box of categories. But an even more One potential explanation for our findings
important issue, and one that has received little is that they are an artifact of the questionnaire
attention in the literature, is the gestalt of cul- design. Perhaps the simple act of asking the
tures. Our cross-cultural understanding will be respondents about “as is” and “should be”
substantially enhanced by taking a holistic view cultural items triggers this type of response
of cultures. Cultures are not a set of independent because people usually want more than they
self-standing dimensions, but instead are have, particularly if the element is positive (e.g.,
formed as a confluence of cultural attributes. It pay). However, the “as is” and “should be” items
is of course easier to study each dimension were physically separated on the questionnaire.
independently, as various researchers, including But even if they were not, it is not clear why the
GLOBE, have done. But such an approach relationship should be negative rather than posi-
leaves many questions unanswered. How do dif- tive. In general, it is unclear as to whether a high
ferent cultural dimensions interact? What is the or low level of any dimension is necessarily
relative importance of each dimension in under- desirable.
standing each culture? A country like Iran is dis- Another possible explanation is that the way
tinguished by its strong in-group collectivism GLOBE conceptualized the construct of cultural
and power distance (Javidan & Dastmalchian, in values is different from the literature and is not
press). What roles do the other dimensions play appropriate. Schwartz (personal communica-
in such a setting? Although these are important tion, August 4, 2003), for instance suggested
22-House.qxd 3/19/2004 5:08 PM Page 730
730– • –CONCLUSION
that respondents might be able to accurately in Switzerland, Sweden, and Denmark report
report their own values, but inaccurately report some of the highest practice scores and the
a society’s values. We have addressed this issue lowest value scores on this dimension.
by showing that GLOBE scales have strong In short, our findings point to the need for a
external and internal validity, they are validated more complex understanding of this relation-
by other independent measures of cultural values, ship which views it as dynamic and double
and they relate in meaningful ways to leadership directional rather than static and uni-directional.
dimensions and aspects of societies. What exacerbates the situation is that our find-
A third explanation, and one that we would ings show that attributes of societal success are
subscribe to, is that the relationship between strongly related to cultural practices, but attrib-
values and practices is nonlinear and more com- utes of outstanding leadership are strongly
plex than initially assumed. GLOBE is the first related to cultural values. Unless we can better
rigorous attempt to empirically measure and understand the relationship between cultural
verify the relationship between cultural values practices and values, we are unable to explain
and practices and it shows counterintuitive this complex situation and have little to offer to
results. The extant view is that people behave in leaders who are trying to improve their
a particular way because they hold particular societies’ well being.
views on how things should be. Our findings
show that the opposite relationship may be at
Cross-Cultural Contact
work; people may hold views on what should be
based on what they observe in action. For An important reason for the interest in
instance, in the case of Future Orientation, with understanding cultures is the increasing rate
very few exceptions, all societies have higher of contact among cultures. Today, people and
values scores than practices scores—societies organizations from different cultures come into
report a desire for a more future oriented regular contact as customers, competitors, part-
society. The negative correlation between prac- ners, and suppliers. As a result, there is an
tices and values occurs because for societies academic and practical need to understand what
with higher practices scores, the difference happens when cultures connect. The current
between values and practices scores (i.e., the literature typically examines each culture in iso-
increment) is much smaller than it is for those lation from others but reaches conclusions about
with low practices scores. This can be seen by cultural contact. For example, the field of cross-
comparing Tables 13.5 and 13.6 which show border mergers is populated by studies that
societal level scores on the GLOBE Future measure cultural distance and reach conclusions
Orientation practices and values scales, respec- on whether it is functional or dysfunctional
tively. Singapore respondents report the highest (Conn & Connell, 1990; Datta & Puia, 1995;
practice score of 5.07 and their value score is Doukas & Travlos, 1988; Morosini, Shane, &
only slightly higher (i.e., 5.51). Russian respon- Singh, 1998). The typical approach to measur-
dents, on the other hand, reported the lowest ing cross-cultural differences is to calculate the
practice score of 2.88 but their value score was mathematical difference on Hofstede’s country
dramatically higher (i.e., 5.41). scores. Despite somewhat conflicting findings,
It is important to recognize, however, that the the conventional wisdom is that cultural dis-
direction of people’s reaction to what should be tance is dysfunctional and leads to failure of
compared to what is may be dimension specific. cross-border mergers.
For example, if one considers another culture It is important to consider the possibility that
dimension, namely Uncertainty Avoidance, the cultural contact between two or more separate
more societal uncertainty avoiding practices cultures entails unique dynamics that go beyond
reported by respondents, the less uncertainty these objective measures of each culture.
avoiding values were reported. This can be seen Objective measures of each culture may not be
by comparing the Uncertainty Avoidance cul- sufficient to understand what happens during
tural practices and values scores in Tables 20.7 the contact. What is probably important is not
and 20.8. As shown in these tables, respondents how a culture is objectively measured but how it
22-House.qxd 3/19/2004 5:08 PM Page 731
is subjectively perceived by those from another to help improve the state of knowledge in this
culture. Furthermore, in cross-cultural settings important field. We used our collective experi-
in which more than two cultures are repre- ence over the past decade to generate these
sented, should we study cultural difference as a thoughts. It is now up to the readers to take up
dyadic phenomenon or as a complex web of the action. We look forward to the upcoming
multilateral issues? For example, do Germans work in this area.
and Austrians behave the same toward each
other if Americans are present? Do they behave
the same if the French are present? Do people REFERENCES
from the Middle Eastern countries behave the
same with and without the presence of Adler, N. J., Doktor, R., & Redding, S. G. (1986).
Europeans? From the Atlantic to the Pacific century: Cross-
Another related issue is the relative and dif- cultural management reviewed. 1986 Yearly
ferential effect of cultural dimensions. When Review of Management of the Journal of
two cultures come in contact, which cultural Management, 12(2), 295–318.
dimensions are key in that relationship? Are Bass, B., & Steidlmeier, P. (1999). Ethics, character,
they all important or are some more important and authentic transformational leadership
than others? Under what conditions? Unless we behavior. Leadership Quarterly, 10, 181–217.
can understand this, we are unable to conceptu- Child, J. D., & Tayeb, M. (1983, Winter). Theoretical
alize cross-cultural contact and its implications. perspectives in cross-national research. Inter-
We also need to study the direction of cultural national Studies of Management and Organiza-
distance. The existing literature assumes sym- tion, 32–70.
metry in cultural difference. This is an impor- Conn, R., & Connell, F. (1990). International merg-
tant and not necessarily logical assumption. ers: Returns to U.S. and British firms. Journal of
There may in fact be asymmetry in such rela- Finance and Accounting, 17(5) 689–711.
tionships. The German and Russian cultures are Datta, D. K., & Puia, G. (1995). Cross-border
different on many dimensions. Can we reach the acquisitions: An examination of the influence of
same conclusions in a German–Russian corpo- relatedness and cultural fit on shareholder value
rate acquisition when the German firm is the creation in U.S. acquiring firms. Management
acquirer and when it is the acquiree? International Review, 35, 337–359.
In addition, there is a clear need to study D’Iribarne, P. (1997). The usefulness of an ethno-
the consequences of cross-cultural differences. graphic approach to the international compari-
New research needs to study the conditions son of organizations. International Studies of
under which cross-cultural differences can be Management and Organization, 26(4), 30–47.
positive or negative. Furthermore, we know Doukas, J., & Travlos, N. G. (1988). The effects of
little of how to deal with these consequences. corporate multi-nationalism on shareholders’
The current writing on this topic is based mostly wealth: Evidence from international acquisi-
on anecdotes and limited case studies. Given the tions. Journal of Finance, 43(5), 161–175.
increasing globalization of markets, it seems Fyans, L. J., Jr., Salili, F., Maehr, M. L., & Desai, K.
obvious that there is both a theoretical and A. (1983). A cross-cultural exploration into the
practical need for rigorous research in this area. meaning of achievement. Journal of Personality
A related issue is the role of leadership. We and Social Psychology, 44(5), 1000–1013.
have shown in this book that societies can have Hofstede, G. (1980). Culture’s consequences: Inter-
similar leadership CLT profiles. It is not clear national differences in work-related values.
whether similar CLT profiles necessarily mean Beverly Hills, CA: Sage.
that leadership styles can easily be transported Inglehart, R., Basanez, M., & Moreno, A. (1998).
between the two cultures. Although this is an Human values and beliefs: A cross-cultural
intuitively obvious and appealing conclusion, it sourcebook. Ann Arbor: University of Michigan
needs empirical verification. Press.
In the previous paragraphs, we have Javidan, M., & Dastmalchian, A. (in press.). Culture
attempted to raise new and interesting questions and leadership in Iran: The land of individual
22-House.qxd 3/19/2004 5:08 PM Page 732
732– • –CONCLUSION
achievers, strong families, and powerful elite. Morosini, P., Shane, S., & Singh, H. (1998). National
Academy of Management Executive. cultural distance and cross-border acquisition
Javidan, M., & Carl, D. (in press.). East meets West: performance. Journal of International Business
Searching for the etic in leadership. Journal of Studies, 29(1), 137–158.
Management Studies. Morrison, A. J. (2000). Developing a global leader-
Kanungo, R. N. (2001). Ethical values of transac- ship model. Human Resource Management
tional and transformational leaders. Canadian Journal, 39(2 & 3), 117–131.
Journal of Administrative Sciences, 18(4), Schein, E. (1992). Organizational culture and leader-
257–265. ship (2nd ed.). San Francisco: Jossey-Bass.
Kanungo, R. N., & Mendonca, M. (1996). Ethical Schwartz, S. H. (1999). A theory of cultural values
dimensions of leadership. Thousand Oaks, CA: and some implications for work. Applied
Sage. Psychology: An International Review, 48,
McClelland, D. C. (1961). The achieving society. 23–47.
Princeton, NJ: D. Van Nostrand. Woodward, J. (1958). Management and technology.
Mendonca, M. (2001). Preparing for ethical leader- London: HMSO.
ship in organizations. Canadian Journal of
Administrative Sciences, 18(4), 266–276.
Appendix A.qxd 3/19/2004 5:09 PM Page 733
APPENDIX A
SOCIETAL-LEVEL
CORRELATIONS AMONG GLOBE
SOCIETAL CULTURE SCALES
PAUL J. HANGES
733
734
Appendix A.qxd
Table A.1 Correlations Among GLOBE Cultural Dimension Practice (As Is) Scales at the Societal Level of Analysis
3/19/2004
Institutional −.42*
Collectivism
* Indicates that the correlation was statistically significant (p < .05). These correlations are based on N = 61 societies.
Appendix A.qxd
Table A.2 Correlations Among GLOBE Cultural Dimension Values (Should Be) Scales at the Societal Level of Analysis
3/19/2004
Institutional −.21
Collectivism
Collectivism
* Indicates that the correlation was statistically significant (p < .05). These correlations are based on N = 61 societies.
735
736
Appendix A.qxd
Table A.3 Correlations Among GLOBE Cultural Dimension Practices (As Is) and Cultural Values (Should Be) Scales at the Societal Level of Analysis
3/19/2004
Assertiveness −.26* .12 −.24 .09 .18 .35* −.01 −.11 .04
5:09 PM
Institutional .37* −.61* −.16 −.25* −.25* −.01 −.39* .38* −.19
Collectivism
In-Group .28* .43* .21 .62* −.44* −.09 .12 .02 .80*
Page 736
Collectivism
Future Orientation .17 −.30* −.42* −.41* .19 .22 −.23 .12 −.53*
Gender −.01 −.31* .16 −.24 .32* −.01 .02 .10 −.06
Egalitarianism
Humane .45* −.02 .20 .26* −.39* −.32* .00 .30* .29*
Orientation
Performance .37* −.22 −.36* −.24 −.11 −.06 −.28* .34* −.24
Orientation
Power Distance −.05 .41* .33* .60* −.08 .21 .40* −.43* .48*
Uncertainty .07 −.32* −.45* −.57* .12 .11 −.14 .19 −.62*
Avoidance
* Indicates that the correlation was statistically significant (p < .05). These correlations are based on N = 61 societies.
Appendix B.qxd 3/19/2004 5:10 PM Page 737
APPENDIX B
PAUL J. HANGES
737
Appendix B.qxd 3/19/2004 5:10 PM Page 738
Although useful, there are several limitations of this statistical correction procedure.
One limitation with the current correction procedure is that although it is apparently
successful in removing cultural response bias, the values of the corrected scales are no
longer interpretable. More specifically, if participants use a 7-point rating scale to
respond to a series of items tapping a particular construct, the meaning of the average
responses produced by these participants is directly interpretable simply by referring
back to the adjectives anchoring the original rating scale. However, the corrected scores
are no longer directly interpretable because the correction procedure creates scores that
are not bound to the original 7-point scale. Indeed, the aggregated corrected scale val-
ues frequently include negative values. Thus, interpretation of the aggregated corrected
score by referring to the original response scale is impossible.
Another limitation with the current correction procedure is that it only provides a
global indication of whether cultural response bias is present in the data. Currently,
cross-cultural researchers correlate the corrected scale scores with the original (i.e.,
“uncorrected”) scores. If the magnitude of the correlations between these two scores
is large, the data are declared to be relatively free from response bias. However, if the
correlations are small, then the data appear to have problems with response bias, and
conclusions drawn from the survey data are suspect. Although useful, this omnibus
assessment of response bias does not provide for the possibility that the cultural
response bias found in the data might be attributable to a few societies rather than the
entire sample. Thus, it would be useful to have a procedure that identifies potentially
response bias contaminated observations.
One final limitation of the current response bias procedure is that it produces ipsative
scores. In other words, the corrected scores reflect the average response to a scale
relative to how all individuals in a society rate all of the scales. There is a long history
in the psychometric literature about the difficulty of using ipsative scores to make
between-group comparisons.
In this appendix, I describe a modification of the traditional cultural response statis-
tical correction procedure that addresses these aforementioned limitations. This modi-
fied procedure not only addresses the first aforementioned limitation by rescaling
the corrected values to make them directly interpretable, but it also enables the use of
well-established diagnostic tools that address the second aforementioned limitation.
Specifically, by using simple regression analysis to predict the uncorrected scale scores
from the corrected scale scores, the corrected scores can be rescaled into the original
7-point scale. Further, one can use residual analysis to identify data points that are out-
liers in this regression analysis (Draper & Smith, 1981). Residual analysis can be used
to identify countries with substantial degrees of response bias.
To provide the reader with a better understanding of the modification that was
developed and used for the GLOBE procedure, Table B.1 shows the societal level
Uncertainty Avoidance cultural practices scores for some of the societies in the
GLOBE database. The uncorrected and corrected average scores (using the traditional
correction procedure) for these societies are shown in this table. The zero-order corre-
lation for the entire dataset between these two scores was .98. As can be seen in this
table, the original scale scores are directly interpretable and range from 2.88 to 5.37.
The corrected scores range are ipsative scores ranging from −.72 to +.56. Thus, they
are not directly interpretable because they are no longer on the 1 to 7 anchored rating
Appendix B.qxd 3/19/2004 5:10 PM Page 739
Appendix B– • –739
scale. All that is known is that relative level of support for uncertainty avoidance
cultural practices in a particular society compared with how people in that society rated
the other cultural and CLT scales.
The first modification that was made to the traditional procedure was to use regres-
sion analysis to rescale these corrected scores back onto the original 7-point anchored
scale. Specifically, we performed a simple ordinary least squares regression analysis
using the corrected Uncertainty Avoidance cultural practices scores to predict the
uncorrected Uncertainty Avoidance cultural practices scores. The obtained regression
equation was significant and the unstandardized predicted values from this equation are
shown in Table B.1 under the column labeled Regression Predicted Scores.1 As can be
seen from this table, the regression predicted scores are back on the original 7-point
anchored scale. Because only a single predictor was used in this regression analysis,
these “regression-predicted” values are simply an isomorphic transformation of the cor-
rected scores (i.e., there is a perfect positive correlation between the corrected and the
regression-predicted scale scores). Thus, the unstandardized regression predicted val-
ues represent our best estimate of what the society Uncertainty Avoidance cultural
practices averages would have been if the respondents were not influenced by cultural-
response bias. Thus, by performing this extra step, the first limitation of the traditional
correction procedure is addressed. The magnitude of the corrected scale scores can now
be directly interpreted.
The second modification to the traditional response bias correction procedure was
the use of well-established diagnostic tests to identify societies whose scale scores
exhibit substantial response bias. The last column in Table B.1, labeled Studentized
Residuals, basically compares the difference (i.e., residual) between the uncorrected
and the regression-predicted Uncertainty Avoidance cultural practices scales and assesses
whether the discrepancy between these two is substantial enough for a particular society’s
data to be considered an outlier. Basically, these studentized residuals are t values. If
the absolute value of the t value is greater than 2, the data point is considered a poten-
tial outlier (i.e., the difference between the uncorrected and the regression-predicted
score is larger than would have been expected by chance).
Examining the studentized residuals column in Table B.1 identifies four countries
that are potential outliers (i.e., Taiwan, Morocco, Qatar, and Indonesia) for this scale.
In other words, the response bias correction procedure substantially changed the scale
scores for these four societies. These four societies can be interpreted as exhibiting sub-
stantially more response bias than the other societies for this scale. Another interpreta-
tion is that the rank order of these four societies significantly changes after correction
for response bias.2
We used this modified correction procedure on the GLOBE culture data and discov-
ered some interesting insights into our data. Table B.2 shows the regression-based
response bias corrected scores for all of the GLOBE societal culture scales, and
Table B.3 shows the societies that were identified as outliers for each of the 18 soci-
etal culture scales. As can be seen, only a few societies are listed as outliers for a par-
ticular rating scale. As discussed in Chapter 8 by Hanges and Dickson (this volume),
this is consistent with the conclusions obtained by applying the traditional response-
bias correction procedure. Overall, very little evidence for response bias was present
in our data.
Appendix B.qxd 3/19/2004 5:10 PM Page 740
Table B.1 Societal Means for Uncertainty Avoidance Cultural Practices Scale Scores and Response Bias
Corrected Values
Appendix B– • –741
742
Appendix B.qxd
Albania 4.57 4.28 5.51 3.69 3.48 4.40 4.57 4.44 4.45
Argentina 4.18 3.66 5.51 3.10 3.44 3.94 3.63 5.56 3.63
Australia 4.29 4.31 4.14 4.09 3.41 4.32 4.37 4.81 4.40
5:10 PM
Austria 4.59 4.34 4.89 4.47 3.18 3.77 4.47 5.00 5.10
Bolivia 3.78 3.96 5.44 3.55 3.45 3.99 3.57 4.46 3.32
Brazil 4.25 3.94 5.16 3.90 3.44 3.76 4.11 5.24 3.74
Page 742
Canada 4.09 4.36 4.22 4.40 3.66 4.51 4.46 4.85 4.54
(English-speaking)
China 3.77 4.67 5.86 3.68 3.03 4.29 4.37 5.02 4.81
Colombia 4.16 3.84 5.59 3.35 3.64 3.72 3.93 5.37 3.62
Costa Rica 3.83 3.95 5.26 3.64 3.56 4.38 4.10 4.70 3.84
Denmark 4.04 4.93 3.63 4.59 4.02 4.67 4.40 4.14 5.32
Ecuador 3.98 3.82 5.55 3.66 3.09 4.45 4.06 5.29 3.63
Egypt 3.91 4.36 5.49 3.80 2.90 4.60 4.15 4.76 3.97
El Salvador 4.49 3.74 5.22 3.73 3.23 3.69 3.72 5.56 3.69
England 4.23 4.31 4.08 4.31 3.67 3.74 4.16 5.26 4.70
Finland 4.05 4.77 4.23 4.39 3.55 4.19 4.02 5.08 5.11
France 4.44 4.20 4.66 3.74 3.81 3.60 4.43 5.68 4.66
Georgia 4.15 4.03 6.18 3.45 3.52 4.17 3.85 5.15 3.54
Germany 4.77 3.67 4.59 4.04 3.17 3.45 4.16 5.70 5.19
(former East)
Appendix B.qxd
Germany 4.66 3.97 4.16 4.41 3.25 3.30 4.42 5.48 5.35
3/19/2004
(former West)
Greece 4.55 3.41 5.28 3.53 3.53 3.44 3.34 5.35 3.52
Guatemala 3.96 3.78 5.54 3.35 3.14 3.91 3.85 5.47 3.44
5:10 PM
Hong Kong 4.53 4.03 5.33 3.88 3.26 3.72 4.69 4.94 4.17
Hungary 4.71 3.63 5.31 3.31 4.02 3.39 3.50 5.57 3.26
India 3.70 4.25 5.81 4.04 2.89 4.45 4.11 5.29 4.02
Page 743
Indonesia 3.70 4.27 5.50 3.61 3.04 4.47 4.14 4.93 3.92
Ireland 3.93 4.57 5.12 3.93 3.19 4.96 4.30 5.13 4.25
Israel 4.19 4.40 4.63 3.82 3.21 4.07 4.03 4.71 3.97
Italy 4.12 3.75 4.99 3.34 3.30 3.66 3.66 5.45 3.85
Japan 3.69 5.23 4.72 4.29 3.17 4.34 4.22 5.23 4.07
Kazakhstan 4.51 4.38 5.50 3.72 3.87 4.15 3.72 5.40 3.76
Kuwait 3.56 4.32 5.70 3.18 2.59 4.44 3.79 4.97 4.02
Malaysia 3.77 4.45 5.47 4.39 3.31 4.76 4.16 5.09 4.59
Mexico 4.31 3.95 5.62 3.75 3.50 3.84 3.97 5.07 4.06
Morocco 4.72 4.18 6.37 3.50 3.08 4.52 4.31 6.14 3.95
Namibia 3.81 4.02 4.39 3.32 3.69 3.83 3.52 5.29 4.09
Netherlands 4.46 4.62 3.79 4.72 3.62 4.02 4.46 4.32 4.81
New Zealand 3.53 4.96 3.58 3.46 3.18 4.43 4.86 5.12 4.86
Nigeria 4.53 4.00 5.34 3.95 3.04 3.96 3.79 5.53 4.14
743
(Continued)
Table B.2 (Continued)
744
Appendix B.qxd
Poland 4.11 4.51 5.55 3.23 3.94 3.67 3.96 5.09 3.71
Portugal 3.75 4.02 5.64 3.77 3.69 3.96 3.65 5.50 3.96
Qatar 4.39 4.78 5.07 4.08 3.86 4.79 3.76 5.05 4.26
5:10 PM
Russia 3.86 4.57 5.83 3.06 4.07 4.04 3.53 5.61 3.09
Singapore 4.06 4.77 5.66 4.88 3.52 3.29 4.81 4.92 5.16
Slovenia 4.01 4.09 5.49 3.56 3.84 3.75 3.62 5.32 3.76
Page 744
South Africa 4.43 4.47 5.18 4.66 3.78 4.46 4.72 4.31 4.64
(Black sample)
South Africa 4.49 4.54 4.42 4.08 3.25 3.45 4.07 5.10 4.06
(White sample)
South Korea 4.36 5.20 5.71 3.90 2.45 3.73 4.53 5.69 3.52
Spain 4.39 3.87 5.53 3.52 3.06 3.29 4.00 5.53 3.95
Sweden 3.41 5.26 3.46 4.37 3.72 4.09 3.67 4.94 5.36
Switzerland 4.58 4.20 4.04 4.80 3.12 3.73 5.04 5.05 5.42
Switzerland 3.61 4.31 3.82 4.36 3.46 3.98 4.36 5.00 5.05
French-speaking
Taiwan 3.70 4.30 5.45 3.65 2.92 3.82 4.27 5.00 4.04
Thailand 3.58 3.88 5.72 3.27 3.26 4.87 3.84 5.62 3.79
Turkey 4.42 4.02 5.79 3.74 3.02 3.92 3.82 5.43 3.67
U.S. 4.50 4.21 4.22 4.13 3.36 4.18 4.45 4.92 4.15
Venezuela 4.26 3.96 5.41 3.43 3.60 4.19 3.41 5.22 3.55
Zambia 4.00 4.41 5.72 3.55 2.88 5.12 4.01 5.23 3.92
Zimbabwe 4.04 4.08 5.53 3.76 3.09 4.38 4.20 5.54 4.12
Regression Predicted Scores for Societal Cultural Values Scales
Appendix B.qxd
Argentina 3.18 5.29 6.07 5.73 4.89 5.50 6.28 2.30 4.62
Australia 3.83 4.47 5.82 5.21 5.02 5.60 5.99 2.77 3.99
Austria 2.85 4.78 5.32 5.15 4.83 5.68 6.12 2.52 3.65
5:10 PM
Bolivia 3.68 5.03 5.91 5.56 4.65 5.11 5.98 3.31 4.64
Brazil 3.06 5.57 5.17 5.60 4.91 5.52 5.98 2.59 5.00
Canada 4.15 4.20 5.94 5.34 5.04 5.58 6.13 2.73 3.73
(English-speaking)
Page 745
China 5.52 4.52 5.12 4.70 3.73 5.34 5.72 3.01 5.34
Colombia 3.45 5.27 5.99 5.52 4.85 5.43 6.15 2.21 4.92
Costa Rica 4.04 5.14 5.94 5.10 4.59 5.08 5.78 2.66 4.58
Denmark 3.59 4.41 5.71 4.49 5.20 5.59 5.82 2.96 4.01
Ecuador 3.57 5.19 5.81 5.62 4.42 5.13 5.95 2.36 4.95
Egypt 3.22 4.72 5.39 5.60 3.34 5.13 5.71 3.20 5.24
El Salvador 3.67 5.60 6.28 5.89 4.66 5.38 6.37 2.76 5.27
England 3.76 4.39 5.66 5.15 5.20 5.52 6.03 2.82 4.17
Finland 3.91 4.34 5.60 5.24 4.47 5.80 6.23 2.46 4.04
France 3.57 5.27 5.88 5.35 4.71 5.91 6.10 2.96 4.65
Georgia 4.29 3.79 5.58 5.45 3.83 5.48 5.63 2.86 5.23
Germany 3.24 4.86 5.38 5.36 4.97 5.56 6.24 2.74 4.02
(former East)
Germany 3.21 5.07 5.46 5.06 5.06 5.63 6.27 2.66 3.38
(former West)
745
(Continued)
Table B.2 (Continued)
746
Appendix B.qxd
Greece 3.05 5.41 5.47 5.17 4.84 5.28 5.79 2.57 5.16
Guatemala 3.65 5.16 5.95 5.78 4.49 5.24 5.96 2.49 4.85
Hong Kong 4.80 4.35 5.11 5.52 4.27 5.38 5.71 3.00 4.52
5:10 PM
Hungary 3.42 4.57 5.58 5.74 4.65 5.48 5.97 2.59 4.74
India 4.65 4.59 5.22 5.43 4.40 5.20 5.87 2.58 4.58
Indonesia 4.50 4.96 5.46 5.48 3.71 5.06 5.54 2.38 5.04
Page 746
Ireland 4.00 4.55 5.72 5.18 5.07 5.45 5.99 2.66 3.94
Israel 3.74 4.25 5.69 5.17 4.66 5.51 5.71 2.72 4.34
Italy 3.87 5.20 5.76 6.01 4.88 5.57 6.11 2.51 4.52
Japan 5.84 4.01 5.44 5.42 4.41 5.53 5.37 2.76 4.40
Kazakhstan 3.88 4.16 5.62 5.22 4.85 5.66 5.57 3.19 4.52
Kuwait 3.61 5.04 5.32 5.62 3.50 5.06 5.89 3.02 4.65
Malaysia 4.73 4.78 5.77 5.84 3.72 5.43 5.96 2.75 4.81
Mexico 3.67 4.77 5.78 5.74 4.57 5.10 6.00 2.75 5.18
Morocco 3.68 5.34 6.03 6.33 4.07 5.73 6.12 3.30 5.77
Namibia 3.76 4.26 6.13 6.30 4.20 5.47 6.52 2.59 5.19
Netherlands 3.13 4.76 5.39 5.24 5.10 5.41 5.71 2.61 3.34
New Zealand 3.52 4.31 6.54 5.90 4.32 4.85 6.24 3.56 4.17
Nigeria 3.14 4.86 5.31 5.80 4.16 5.71 5.99 2.66 5.45
Philippines 4.93 4.55 5.86 5.66 4.36 5.19 6.00 2.54 4.92
Regression Predicted Scores for Societal Cultural Values Scales
Appendix B.qxd
Portugal 3.61 5.40 5.97 5.50 5.12 5.40 6.41 2.45 4.50
Qatar 3.72 5.10 5.55 5.92 3.49 5.31 5.94 3.18 4.82
Russia 2.90 4.01 5.90 5.60 4.34 5.62 5.68 2.73 5.26
5:10 PM
Singapore 4.28 4.42 5.46 5.46 4.43 5.66 5.70 2.84 4.08
Slovenia 4.61 4.36 5.71 5.43 4.78 5.31 6.41 2.50 5.03
South Africa 3.97 4.46 5.14 5.25 4.43 5.23 5.09 3.80 4.92
Page 747
(Black sample)
South Africa 3.65 4.36 5.82 5.59 4.54 5.53 6.13 2.67 4.65
(White sample)
South Korea 3.69 3.84 5.50 5.83 4.23 5.61 5.41 2.39 4.74
Spain 4.01 5.25 5.82 5.66 4.82 5.63 5.85 2.23 4.80
Sweden 3.49 3.91 6.25 4.96 5.19 5.72 6.01 2.49 3.45
Switzerland 3.31 4.87 5.16 4.93 5.01 5.63 6.00 2.54 3.20
Switzerland 3.83 4.42 5.54 4.89 4.77 5.68 6.17 2.80 3.84
French-speaking
Taiwan 2.91 4.95 5.30 4.94 3.88 5.15 5.58 2.77 5.14
Thailand 3.43 5.08 5.73 6.26 4.12 5.05 5.76 2.74 5.71
Turkey 2.68 5.18 5.63 5.71 4.46 5.40 5.34 2.52 4.61
U.S. 4.36 4.20 5.79 5.34 5.03 5.51 6.14 2.88 3.99
Venezuela 3.34 5.28 5.92 5.61 4.70 5.24 6.11 2.43 5.19
Zambia 4.24 4.55 5.64 5.76 4.27 5.37 6.08 2.37 4.45
Zimbabwe 4.60 4.84 5.74 6.01 4.40 5.20 6.33 2.65 4.68
747
Appendix B.qxd 3/19/2004 5:10 PM Page 748
Assertiveness • Qatar
• France
• Nigeria
Assertiveness • Taiwan
• Morocco
• Japan
Appendix B– • –749
Table B.4 Frequency for Society to be Identified as Outlier in Response Bias Regression Procedure
Ecuador 3 17
France 13 72
Indonesia 2 11
Japan 1 6
Morocco 14 78
New Zealand 2 11
Nigeria 2 11
Philippines 1 6
Qatar 8 44
Taiwan 8 44
Table B.4 summarizes these findings by indicating the percentage of times that a
particular society was identified as an outlier. As can be seen from this table, most
societies identified as outliers by this regression-based correction procedure were only
identified as such for one or two culture scales. However, a few societies (i.e., France,
Morocco, Qatar, and Taiwan) were repeatedly identified as outliers. The observed fre-
quency of these societies being classified as outliers across the 18 societal cultural
scales is substantially greater than what would be expected by chance, and so these
results suggest that the data from these societies might be exhibiting substantial levels
of cultural response bias. It is the societies that are repeatedly identified as outliers that
are of concern and, once identified, researchers need to seek confirmatory evidence
from their research notes, discussions with individuals who collected the data, or from
some other source that confirms that respondents in these societies were overly sensitive
Appendix B.qxd 3/19/2004 5:10 PM Page 750
to cultural response bias before taking any action with regard to these observations
(e.g., discarding data from these societies).
It is worth repeating that researchers should not automatically discard data from
societies identified as problematic. Outlier status cannot be unambiguously attributed
to cultural response bias. Rather, outlier status could indicate that some novel but the-
oretically meaningful process is occurring in these identified societies. The history of
science has several cases of research progress being slowed by adopting the practice of
automatically excluding outliers. For example, the measurements indicating existence
of the hole in the ozone layer were initially thought to be outliers and so they were auto-
matically discarded. This oversight delayed the discovery of this environmental prob-
lem by several years (Berthouex & Brown, 1994). Thus, labeling an observation as an
outlier should activate researchers to seek additional information that provides possible
explanations for the outlier status of the observation rather than activate researchers to
automatically reject parts of their data.
In the GLOBE project, we spoke to the country co-investigators of these four outlier
societies and learned that although they experienced some unique data collection prob-
lems, there was no clear-cut confirmatory evidence of rampant response bias. Thus, we
followed the protocol outlined above and kept these observations in the analyses.
However, it should be noted that we did explore whether our conclusions would have
substantially differed if these countries were excluded. No substantial or systematic
changes in conclusions were noted when we reran our HLM analyses excluding these
four societies or when we reran the HLM analyses using the response-bias corrected
scores instead of the raw scores.
In summary, I developed a new protocol that modifies the traditional cultural
response statistical correction procedure used in the cross-cultural literature. This
regression-based correction procedure addresses several limitations of the traditional
approach. First, it rescales the corrected values so that they are directly interpretable.
Second, it enables use of well-established diagnostic tools that can identify specific
observations that appear to be problematic.
ENDNOTE
1. It should be noted that data from the Czech Republic was not included in this analysis.
Prior application of this regression-based procedure revealed that the Czech Republic data exhib-
ited substantial and pervasive response bias. Once the Czech Republic was removed from these
analyses, the results reported in this appendix were obtained.
2. A simple test confirms this last interpretation of the regression modification. If the reader
compares the rank order of the societies using the uncorrected scales to the rank order of the
societies using the regression-predicted scores, it will be discovered that the four countries iden-
tified as outliers in Table B.1 are the four countries with the largest change in their rank order-
ing. The outlier analysis identifies the societies that change their rank ordering the most after
controlling for response bias.
REFERENCES
Berthouex, P. M., & Brown, L. C. (1994). Statistics for environmental engineers. London: CRC
Press.
Draper, N. R., & Smith, H. (1981). Applied regression analysis (2nd ed). New York: John
Wiley.
Appendix B.qxd 3/19/2004 5:10 PM Page 751
Appendix B– • –751
Hui, C. C., & Triandis, H. C. (1989). Effects of culture and response format on extreme response
style. Journal of Cross-Cultural Psychology, 20, 296–309.
Stening, B. W., & Everett, J. E. (1984). Response styles in a cross-cultural managerial study.
Journal of Social Psychology, 122, 151–156.
Triandis, H. C. (1994). Cross-cultural industrial and organizational psychology. In H. C. Triandis,
M. D. Dunnette, & L. M. Hough (Eds.), Handbook of industrial and organizational
psychology (2nd ed., Vol. 4, pp. 103–172). Palo Alto, CA: Consulting Psychologists Press.
Appendix B.qxd 3/19/2004 5:10 PM Page 752
Appendix C.qxd 3/19/2004 5:11 PM Page 753
APPENDIX C
EVIDENCE FOR
CONTEXTUAL EFFECTS
PAUL J. HANGES
MINA T. SIPE
ELLEN G. GODFREY
A hierarchical linear model (HLM) analyses to test our hypotheses regarding the
relationship among organizational culture, societal culture, and the six cultur-
ally endorsed leadership theory (CLT) dimensions. The results of these analyses are
discussed in the culture dimension chapters, the organizational culture chapter, and the
CLT cluster chapter (i.e., Chapters 12–21). In this appendix, we discuss the results of
additional HLM analyses conducted to demonstrate that the organizational-level rela-
tionships reported in these aforementioned chapters are truly a function of organiza-
tional-level covariation and not simply a function of individual-level biases. Even
though we used HLMs to test the GLOBE hypotheses about the relationships between
organizational culture and the six CLTs, individual-level covariation could have biased
our findings because the same individuals provided both the CLT dimensions scores
and the organizational culture scores in these analyses. The analyses reported in this
appendix will demonstrate that the obtained organizational culture and CLT relationships
reported in the body of this book go beyond individual-level biases and are a function
of forces operating at the organizational level of analysis.
There are several statistical methods1 that have been developed to evaluate the level of
covariation influencing statistical relationships. In particular, WABA 2 (Within and Between
Analysis; Dansereau & Yammarino, 2000) and contextual analyses (Bliese, 2000) are two
753
Appendix C.qxd 3/19/2004 5:11 PM Page 754
Table C.1 Results of Contextual HLM Analysis for the Relationship Between Organizational Culture
Scales and the Higher-Order CLT Dimensions
Power Distance
Appendix C– • –755
Uncertainty Avoidance
Collectivism 1
Charismatic/Value-Based leadership
Autonomous leadership
Collectivism 2
Gender Egalitarianism
Assertiveness
Charismatic/Value-Based leadership
Participative leadership
Humane Orientation
Appendix C– • –757
Performance Orientation
Autonomous leadership
Future Orientation
Self-Protective leadership
ENDNOTE
1. Elimination of individual-level covariation from the organizational-level relationships
could have been handled empirically by having one set of respondents complete the organiza-
tional culture items and a different set of respondents complete the CLT questions. Indeed, this
is the method used in the HLM analyses reported in Chapters 12–21 to ensure that any obtained
societal-level culture–CLT relationships were a function of societal-level covariation.
REFERENCES
Bryk, A. S., & Raudenbush, S. W. (1992). Hierarchical linear models for social and behavioral
research: Applications and data analysis methods. Newbury Park, CA: Sage.
Dansereau, F., & Yammarino, F. J. (2000). Within and between analysis: The variant paradigm
as an underlying approach to theory building and testing. In K. J. Klein & S. W. J. Kozlowski
(Eds.), Multilevel theory, research, and methods in organizations: Foundations, extensions,
and new directions (pp. 425–466). San Francisco, CA: Jossey-Bass.
Appendix D.qxd 3/19/2004 5:11 PM Page 759
APPENDIX D
Confidence Interval
Estimation of Correlations
PAUL J. HANGES
n this book we reported whether the correlations between variables were statistically
25 −.31 to .48 −.21 to .55 −.11 to .62 .01 to .69 .13 to .75
30 −.27 to .44 −.17 to .52 −.07 to .60 .05 to .66 .17 to .73
35 −.24 to .42 −.14 to .50 −.04 to .58 .08 to .65 .20 to .71
40 −.22 to .40 −.12 to .48 −.01 to .56 .10 to .63 .22 to .70
45 −.20 to .38 −.10 to .47 .01 to .55 .12 to .62 .24 to .69
50 −.18 to .37 −.08 to .45 .02 to .53 .14 to .61 .26 to .68
55 −.17 to .36 −.07 to .44 .04 to .52 .15 to .60 .27 to .68
60 −.16 to .35 −.06 to .43 .05 to .51 .16 to .59 .28 to .67
65 −.15 to .34 −.05 to .42 .06 to .51 .17 to .59 .29 to .66
is reported to be .30 and if this correlation was based on 45 observations, the 95%
confidence interval for the relationship between these two variables would be .01 to .55.
We hope that this table will provide the reader with a feel for the magnitude of the con-
fidence intervals that would have been reported in our study.
REFERENCES
Cohen, J. (1994). The earth is round (p < .05). American Psychologist, 49, 997–1003.
Schmidt, F. L. (1996). Statistical significance testing and cumulative knowledge in psychology:
Implications for the training of researchers. Psychological Methods,1, 115–129.
Schmidt, F. L., & Hunter, J. E. (1997). Eight common but false objections to the discontinuation
of significance testing in the analysis of research data. In L. L. Harlow, S. A. Mulaik, &
J. H. Steiger (Eds.), What if there were no significance tests? (pp. 37–64). Mahwah, NJ:
Lawrence Erlbaum.
Author-Ind.qxd 3/19/2004 5:42 PM Page 761
AUTHOR INDEX
761
Author-Ind.qxd 3/19/2004 5:42 PM Page 762
Hashimoto, T., 452 543, 549, 555, 566, 602, 603, 604, 605, 607,
Hatzichristou, C., 416, 604, 612 608, 609, 610, 611, 612, 613, 614, 616, 617,
Haudek, G., 78, 85 625, 626, 627, 633, 634, 640, 644, 645, 646,
Hausmann, R., 208, 211 663, 709, 725, 729
Hay, D., 396 Hogg, M. A., 670
Hayes, J., 76 Holland, J. L., 85
Hayes, R. H., 296 Hollandsworth, J. G., 396
Hazucha, J. F., 52 Holmberg, I., 414
Hedlund, G., 291 Holtgraves, T., 403, 404, 452
Heilman, M. E., 349, 384, 389, 400 Hom, P. W., 299
Heiman, R. J., 437 Hoornaert, J., 292
Heine, S. J., 604 Hope, C. A., 353, 354
Heliman, M. E., 348 Hoppe, M. H., 54, 345, 346, 406, 530, 543, 611
Heller, F., 535 Horovitz, J. H., 290
Heller, T., 58 Horowitz, R. T., 294
Helmreich, R. L., 533, 543, 611 Hossain, Z., 351, 357, 358
Hempill, J. K., 266, 588 House, R. J., xix, 5, 10, 15, 16, 17, 18, 19,
Henderson, M. C., 449 51, 53, 56, 57, 59, 64, 65, 95, 98, 103, 133,
Hendrix, L., 350, 351, 357, 358 222, 267, 268, 272, 275, 276, 282, 288, 291,
Herder, J. G., 208 301, 343, 359, 381, 395, 430, 460, 514,
Herold, D. M., 604 516, 536, 554, 569, 592, 670, 672, 673, 679,
Hersey, P., 535 703, 709
Herzberg, F., 441 Hovland, C. I., 292
Heusser, C., 400 Howard, A., 399, 400
Hezlett, S. A., 52 Howell, J., 617
Hibino, S., 60, 453, 460, 588 Howell, J. M., 65
Hickson, D. J., 16, 26, 617, 656 Howell, J. P., 58, 59, 60, 61, 62, 64, 65, 453, 455,
Hideg, E., 292, 294, 295, 300 460, 588, 702
Higgins, C. A., 399, 617 Hox, J. J., 222
Higgins, M., 672 Hsee, C., 616
Hills, G. E., 606 Hu, C. T., 283
Hills, J., 358, 369 Huang, X., 212
Hiltrop, J. M., 584, 585 Hui, C. C., 137, 737
Hines, S. C., 605 Hui, C. H., 59, 442, 444, 448, 452, 453, 456, 503
Hinings, C. R., 656 Huismans, S. E., 353, 371, 373, 418, 419, 422,
Hinkin, T. R., 514 432, 449
Hinnings, C. R., 16, 26 Hume, D., 284
Hirsch, J., 255, 568 Humphrey, D. B., 634
Hochschild, A., 350 Hunt, J. G., 57, 61
Hoelter, J., 566 Hunt, R. G., 444
Hoffman, L. W., 240, 566 Hunter, A., 142
Hofmann, D. A., 222, 223, 225 Hunter, J. E., 759
Hofstede, G., xiv, xix, xxiii, 13, 16, 17, 54, 55, 56, Hunter, L., 59
59, 75, 77, 78, 99, 100, 117, 120, 123, 137, Huntington, E., 208, 259
138, 139, 140, 141, 154, 179, 180, 181, 205, Huntington, S. P., xvi, 502
211, 239, 241, 242, 243, 244, 255, 258, 263, Huo, P. Y., 58, 611
266, 284, 286, 287, 289, 290, 295, 296, 300, Huselid, M. A., 534
305, 311, 313, 314, 319, 323, 332, 333, 343, Huxley, A., 188
344, 345, 346, 347, 348, 351, 352, 353, 354, Hwang, K. K., 60, 688
360, 371, 373, 374, 377, 388, 389, 396, 399, Hyden, G., 180
400, 401, 406, 412, 413, 414, 418, 419, 422,
428, 431, 432, 437, 440, 441, 442, 443, 445, Ibarra, H., 79
446, 447, 450, 451, 452, 457, 460, 461, 462, Ibrahim, S. E., 311
463, 474, 475, 476, 478, 480, 484, 502, 516, Ide, E., 604
517, 525, 527, 529, 530, 531, 532, 533, 537, Ilgen, D. R., 606
Author-Ind.qxd 3/19/2004 5:42 PM Page 768
Inglehart, R., xvi, 20, 21, 30, 101, 110, 142, 143, Kakar, S., 60, 62
152, 162, 163, 167, 181, 189, 235, 256, 284, Kakuyama, T., 456
287, 311, 314, 333, 533, 626, 628, 735 Kale, S. H., 405, 613
Inkeles, A., 58, 346, 709 Kalin, R., 400
Inkson, J. H. K., 617 Kammeyer-Mueller, J. D., 605
Inlis, I. R., 602 Kanbur, R., 568
Ino, S., 298 Kant, I., 284
Irani, K. D., 398 Kanter, R., 616
Ivancevich, J. M., 64 Kanungo, R. N., 241, 295, 446, 447, 453, 455, 460,
Iwao, S., 443, 444 462, 501, 566, 571, 584, 728
Iyengar, S. S., 456 Kaplan, R. S., 297
Kappeter, I., 292, 294, 295, 300
Jablin, F. M., 605 Karsh, E., 187
Jaccard, J. J., 443 Karsh, I., 187
Jackson, S. E., 291, 569 Kasch, C. R., 605
Jacobs, J. A., 350 Kashima, E. S., 452
Jaeger, A. M., 295, 447, 453 Kashima, Y., 448, 449, 452, 453, 455, 458, 460, 461
Jago, A., 60, 61, 517 Kastenbaum, R., 292
Jakubowski, P., 164, 396, 397 Kateb, G., 439
James, K., 241, 566 Kattler, D., 348
James, L. R., 128, 132, 133, 146, 222, 229, 232, Katz, D., 83, 266, 588, 656
278, 335, 390, 433, 598, 647 Katz, M. M., 348, 352
Jang, H., 78 Katznelson, D. B., 298
Jarman, J., 350 Kavcic, B., 604
Javidan, M., 51, 268, 282, 343, 359, 381, 395, 554, Keane, T. M., 398
569, 581, 670, 672, 696, 728 Keats, B. W., 606
Jeanquart-Barone, S., 614 Keegan, W. J., 291
Jehn, K. A., 654, 665 Kellam, S., 566
Jessop, B., 255, 567, 568 Kellehear, A., 162
Johannesen-Schmidt, M. C., 384 Keller, T., 670
Johnson, D. M., 299 Kelly, J. A., 398
Johnson, H. T., 297 Kelly, J. R., 605
Johnson, S., 210 Kemmelmeier, M., 437, 474
Johnston, J. H., 671 Kennedy, A. A., 425
Jones, A. P., 51 Kennedy, P., 450
Jones, J. M., 293 Kennis, I., 60
Jones, R. M., 566 Kenny, D. A., 227
Jones, W., 454, 455 Keough, K. A., 285, 299, 300
Jordan, J. O., 348 Kern, J. M., 398
Judd, C. M., 227 Kerr, C., 581, 656
Judge, T., 605 Kerr, H. C., 180
Judge, T. A., 399 Kerr, S., 267
Judge, W. Q., 296 Kerrigan, J. E., 180
Jung, D., 461, 493 Kets de Vries, M. F., 79
Jürgens, U., 582, 584 Khadra, B., 327
Kidder, C., 299
Kabasakal, H., 554, 589 Kiely, E., 294
Kagan, D., 438 Kim, J. S., 60
Kagitçibasi, C., 240, 437, 443, 452, 526, 566, Kim, N. H., 60
578, 608 Kim, U., 437, 439, 443, 447, 448, 449, 453, 461,
Kagono, T., 291 500, 608
Kahn, H., 241 Kipp, K., 419
Kahn, R. L., 83, 266, 604, 656 Kiray, M., 566
Kahneman, D., 291 Kirk, R. E., 658
Kaicheng, Y., 576 Kirkcaldy, B. D., 179, 181, 254, 257
Author-Ind.qxd 3/19/2004 5:42 PM Page 769
Trompenaars, F., 52, 59, 180, 182, 243, 245, 246, Waldman, D. A., 61
247, 248, 253, 264, 265, 286, 288, 289, 290, Walker, H. A., 348
311, 312, 314, 402, 404, 445, 532, 640 Wall, T. D., 125
Truban, D. B., 605 Wallace, M., 292, 299
Tsai, Y., 355 Walsh, T., 608
Tu, W. M., 283 Wanberg, C. R., 605
Turgeon, N., 185 Wang, D. F., 616
Turner, J. H., 605 Wang, Z. M., 62, 77
Tushman, M. L., 18 Warner, R. L., 356
Tversky, A., 291 Wasti, S. A., 455, 456
Tyler, L. E., 293, 295, 298, 300 Watson, D., 399
Watson, J. B., 284
Udas, A., 458 Wayne, S. J., 535
Ueno, S., 627 Webb, E., 153
Uleman, J. S., 462 Webber, M. J., 290
Ulin, R. D., 58 Weber, E. U., 616, 617
Ungson, G. R., 58, 63 Weber, M., xix, 59, 208, 239, 241, 253, 277, 295,
323, 515, 526
Vadi, M., 449, 462 Weber, R. P., 153, 158, 175
Valencia, J., 605 Wehr, T. A., 211, 212
Valkenburg, S., 208 Weick, K. E., 291
Vallibhotama, S., 188 Weigenstein, J., 626
van Deusen, C., 287 Weill, M., 86
Van de Ven, A., 607 Wei-Ming, T., 189
van de Vijer, F., 99, 100, 138, 203, 236, 681 Weingart, L. R., 605
Van de Vliert, E., 212, 353, 371, 373, 418, 419, Weiss, J. W., 64
422, 432, 531 Weiss, L. B., 300
VandeWalle, D., 604 Weissman, M. D., 449
van Gunsteren, H. R., 609 Welch, M. R., 357, 359
van Twillert, B., 416, 604, 612 Wells, S., 348
Van Velsor, E., 145 Wensink, J., 416, 604, 612
Van Velzen, D., 569 Wenzel, L. H., 671
Van Vianen, A. E., 310 Westwood, R. G., 611
Van Yperen, N. W., 353, 373, 531 Whaba, M., 515
Vassiliou, V. A., 454 Wheeler, L., 452
Veblen, T., 556 Whitehill, A. M., 63
Venaki, S., 294 Whyte, M. K., 356
Venkataraman, S., 76, 607, 617, 638 Whyte, W. F., 529
Verderber, K. S., 541 Widick, B. J., 582
Verderber, R. F., 541 Wierda, E., 568
Verma, J., 451, 452 Wieser, G., 604
Veroff, J., 240 William, F. E., 208
Verweij, J., 354, 355, 633 Williams, J. E., 287, 344, 349, 377, 378, 379, 400
Vesala, J. M., 634 Williams, L. J., 222, 232
Vianello, M., 604 Williamson, J. B., 348, 355
Viktor, I., 555 Williamson, O., 606
Villareal, M. J., 165, 166, 444, 448, 503 Williamson, R., 184
Volder, M. M. de, 299 Winer, B. J., 266, 588
Von Glinow, M. A., 59, 60, 63 Wing Sue, D., 295
Vries, M. K., 199 Winnbust, J. A. M., 292
Vroom, V. H., 60, 455, 517, 535 Winter, D. G., 515, 516
Vunderink, M., 346 Winterbottom, M. R., 240, 527
Withaney, S., 611
Wade, R., 54 Witkin, H. A., 608
Wagner, J. A., III, 443, 449, 457 Wlaton, W., 348
Author-Ind.qxd 3/19/2004 5:42 PM Page 777
SUBJECT INDEX
779
Subject-Ind.qxd 3/19/2004 5:51 PM Page 780
power distance scores, 539–540 (tables) response style, assertiveness-training and, 397
response bias scales, 740–749 (tables) sex-trait stereotypes and, 400–401
uncertainty avoidance and, 622–623 (tables) societal component of, 405–406, 405 (table)
Ascribed status, 242–243, 253, 264–265, 404, 532 trust-building process and, 404–405
Asia-Pacific Economic Cooperation (APEC), 204 uncertainty avoidance and, 624 (table)
Asia. See Confucian Asia; Southern Asia values-practices relationship, statistical analysis
Assertiveness, xiv, 13, 395–396 of, 30–32, 31 (table), 33–37 (figures)
achieved vs. ascribed status and, 404 See also Organizational cultural practices
aggressiveness and, 396 Assertiveness training, 397
assertiveness training and, 397–399 cultural differences and, 398–399
climatic influence on, 208, 210–211, 418–422, gender differences and, 398
420–421 (table), 432 individualism and, 397–398
CLT leadership dimensions, assertiveness pragmatism and, 398
cultural values and, 427–430, rationality and, 398
429 (figure), 432 studies on, 398
collectivism scales and, 470, 472, Attraction-selection-attrition (ASA) model,
472–473 (tables) 78, 79, 85
competitiveness, value of, 402–403 Australia, 12, 55, 106, 108, 129, 180, 184, 242, 269,
cultural dimension of, 401–402 304, 306, 326
cultural/societal effectiveness measures, 37–39, assertiveness and, 410–411 (tables)
38–41 (tables) collectivism rankings, 468–471
culture dimensions, correlation with, 413–415, gender egalitarianism and, 365–366 (tables)
415 (table), 432 humane orientation and, 573–574 (tables)
definition of, 12, 30 power distance scores, 539–540 (tables)
domestic political violence, 422, 432 response bias scales, 740–749 (tables)
economic health and, 416, 417 (table) uncertainty avoidance and, 622–623 (tables)
external adaptation/internal integration and, Austria, 12, 61, 106, 109, 129, 185, 269, 304, 306,
402–405, 432 326, 365, 366, 411, 412
future orientation, correlation of, 309–311, collectivism rankings, 468–471
309 (table) humane orientation and, 573–574 (tables)
GLOBE findings on, 408–412, 409 (table) power distance scores, 539–540 (tables)
GLOBE measures of, 406–408, 407–408 (tables) response bias scales, 740–749 (tables)
human condition and, 416–417, 417 (table) Authoritarian leadership, 62, 63, 311, 328, 334
humane orientation and, 575, 576 (table) Autocratic leadership style, 61
individual values/perceptions, cluster/society Automotive industry, 82, 86–87, 582
effects on, 200–201, 202 (table) Autonomous cultures, 312 (table), 313–314,
industry scores, overall societal scores and, 527–528, 528 (table), 627–628, 628 (table)
408–412, 409–411 (tables) Autonomous leadership, xv, xvi, 7, 14, 675
language/communication preferences and, assertiveness cultural practices/values and,
403–404 429 (figure), 430
masculinity/femininity culture dimension and, collectivism practice/values scores and,
412–415, 432 492 (table), 493, 494–496 (tables),
organizational culture and, 425, 432 496 (figure), 498, 499, 501
organizational/societal values-practices cultural dimension values and, 45, 48 (figure)
relationship, 426–427, 426–427 (tables) future orientation cultural practices/values and,
parental investment theory and, 419 330 (figure), 331
personality trait of, 399–400 gender differences and, 384–385
political involvement and, 164 gender egalitarianism cultural practices/values
power distance and, 544, 544 (table) and, 383
psychological perspective on, 397–406 humane orientation cultural practices/values and,
regional clusters and, 422–423, 593 (figure), 595, 597
423–424 (table), 432 leadership dimension scores, 42, 44 (figure)
regional clusters, classification of, 192–200, performance orientation cultural values and,
193–194 (tables), 195 (figure) 274 (figure), 275
response bias scales, 412, 742, 749 (tables) power distance cultural practices and, 553–555
Subject-Ind.qxd 3/19/2004 5:51 PM Page 781
cultural values/cultural practices and, 214–215, See also Economic development; Global
214 (table) competitiveness
culture dimensions and, 208, 210–211 Computer Aided Design/Computer Aided
extreme climate hypothesis, thermoclimate factor Manufacturing (CAD/CAM) systems, 605
and, 211–212 Confidence intervals, 237, 759–760, 760 (table)
future orientation and, 320–322, 321 (table) Conformity, 79, 457
gender egalitarianism and, 352–353 Confucian Asia, xvi, 189
historical development of, 206 assertiveness and, 423, 424 (table), 427,
humane orientation and, 580, 581 (table) 427 (table)
ideal climate, 259 climatic effects and, 212–215, 213–214 (tables)
managerial/theoretical implications of, 215 collectivism practice/values scores and, 476–481,
performance orientation and, 259–261, 477 (table), 479 (table)
260–261 (table) cultural dimension scores, 192–200, 193–194
power distance and, 549–551 (tables), 195–199 (figures)
regional attributes in, 207–208, 209 (table) cultural values-practices relationship, 32,
uncertainty avoidance and, 208, 210–211, 636, 33 (figure)
638–639 (table) future orientation and, 322–323, 322 (table)
See also Regional clustering gender egalitarianism and, 374, 376 (table)
CLTs. See Culturally endorsed implicit theories of humane orientation and, 580–581, 582 (table),
leadership (CLTs) 585, 587, 587 (table)
Cluster phenomenon. See Climatic clustering; humane-oriented leadership and, 591–592,
Regional clustering 592 (table)
Code of Hammurabi, 438 individual values/perceptions, cluster effect on,
Coercive isomorphism, 82–83, 86 200–201, 202 (table)
Cohesion perspective, 79, 80 leadership dimension scores, 42, 42–45 (figures),
Collectivism. See Individualism/collectivism 493, 494 (table)
constructs; In-group collectivism; Institutional metaconfiguration, cultural clusters and, 200,
collectivism 201 (figure)
Collectivism syndromes, 472–474 performance orientation and, 261–264,
Collectivistic cultures. 263 (table), 271 (table)
See Individualism/collectivism constructs power distance and, 547–549, 548 (table),
Colombia, 12, 107, 109, 129, 186, 269, 550 (table)
304, 306 uncertainty avoidance and, 637 (table)
assertiveness and, 411–412 (tables) visionary leadership and, 327–328, 328 (table)
collectivism rankings, 468–471 See also Culturally endorsed implicit theories of
gender egalitarianism and, 365–366 (tables) leadership (CLTs); Regional clustering
humane orientation and, 573–574 (tables) Confucian Dynamism Index, 254, 258, 286–288, 313
power distance scores, 539–540 (tables) Confucianism, 59, 188, 189, 241, 295, 439, 523
response bias scales, 740–749 (tables) Confucian work ethic, 13, 288
uncertainty avoidance and, 622–623 (tables) Congruence hypothesis, 204
visionary leadership in, 325, 326 Considerate leaders. See Charismatic/value-based
Communication, 7, 25–26 leadership
assertiveness and, 397, 403–404 Construct validation process, 152–153
cultural clustering and, 182, 205 archival data, objective correlates and, 153,
individualistic/collectivistic cultures and, 452 154 (table)
uncertainty reduction and, 605, 612 assertiveness values, political participation
Communism, 525, 567 and, 164
Compensation practices, 78, 453, 455 coding process, 156–158
Competition, 80 conceptual/thematic analysis and, 156,
assertiveness and, 402–403 157 (table)
organizational future orientation and, 296 consistency/reliability measures, 158–160,
prosperity/performance orientation practices and, 158 (table)
253–254, 253 (table) construct-inference link, contextual knowledge
regulatory pressures and, 86 and, 155–156
reports on, 104 content analysis and, 153
Subject-Ind.qxd 3/19/2004 5:51 PM Page 783
future orientation scores of, 310 x-inefficiencies, mass production and, 582
Gender egalitarianism and, 365–366 (tables) competitiveness, social cultural practices and, 19
humane orientation and, 569, 573–574 (tables) financial markets and, 4
power distance scores, 539–540 (tables) gender egalitarianism and, 355–356
response bias scales, 740–749 (tables) Islamic banking practices, 295
uncertainty avoidance and, 622–623 (tables) modernization, neo-liberal restructuring and, 568
Dependency theory, 81, 85 neoclassical development theory, 568
Desert climate, 206–207, 207 (table), 209 (table) service economy, 284–285
assertiveness and, 419–422, 420–421 (table) societal clusters and, 180, 181
future orientation and, 320–322, 321 (table) sustainable vs. transient competitive
gender egalitarianism and, 371–373, 372 (table) advantage, 288
humane orientation and, 580, 581 (table) theories of, 17
performance orientation and, 259–261, world market integration, 568
260 (table) See also Climatic clustering; Globalization
uncertainty avoidance and, 638 (table) Economic health, 37–38, 38–39 (tables), 110
Discrimination, 567, 579–580 assertiveness dimension and, 416, 417 (table)
Diversity, 53, 54 competitiveness, societal support for,
Domestic political violence (DPV), 422, 432 111–112 (table), 254–255
Dominant culture systems, 18, 54, 56, 62 future orientation and, 315–316, 315 (table)
Double translations, xiii gender egalitarianism and, 364–367, 368 (table)
humane orientation and, 577–578, 579 (table)
Eastern Europe, xvi, 185–186 individualism/collectivism and, 481–483,
Asian influence and, 186 482 (tables)
assertiveness and, 423, 423 (table), 427, 427 performance orientation and, 252–256,
(table) 253 (table)
climatic effects and, 212–215, 213–214 (tables) power distance and, 556, 557 (table)
collectivism practices/values scores and, productivity factor, 111
476–481, 477 (table), 479 (table) prosperity factors, 111
cultural dimension scores, 192–200, public sector support factors, 111
193–194 (tables), 195–199 (figures) savings rates, 287–288, 319–320, 320 (table)
cultural values-practices relationship, uncertainty avoidance and, 629–630, 631 (table)
32, 35 (figure) Ecuador, 12, 107, 129, 186, 269, 304, 306
gender egalitarianism and, 374, 375 (table) assertiveness and, 410–411 (tables)
humane orientation and, 580–581, 582 (table), collectivism rankings, 468–471
585, 587, 587 (table) future orientation scores for, 310
humane-oriented leadership and, 591–592, gender egalitarianism and, 365–366 (tables)
592 (table) humane orientation and, 573–574 (tables)
individual values/perceptions, cluster effect on, power distance scores, 539–540 (tables)
200–201, 202 (table) response bias scales, 740–749 (tables)
leadership dimension scores, 42, 42–45 (figures), uncertainty avoidance and, 622–623 (tables)
493, 494 (table) visionary leadership in, 325, 326
metaconfiguration, cultural clusters and, Educational levels, 295
200, 201 (figure) Effectiveness, 7
performance orientation and, 261–264, assertive personalities and, 399–400
262 (table), 270 (table) cultural/societal effectiveness, 37–39,
power distance and, 547–549, 548 (table), 38–41 (tables)
550 (table) expatriate failures and, 203
steppe empire in, 186 hierarchy of, leadership styles and, 61
uncertainty avoidance and, 636 (table) international operations and, 10
visionary leadership and, 327–328, 328 (table) leader attributes/behavior and, 14, 19, 51–52
See also Culturally endorsed implicit theories of leadership performance orientation and,
leadership (CLTs); Regional clustering 267–268, 272, 275–276
Ecological fallacy, xiii, 99 regional clusters, effective management and, 203
Ecological influence, 80, 450 Egalitarian practices, 61
Economic development, xv, xvi Egypt, 12, 107, 109, 129, 187, 269, 304, 306, 326
Subject-Ind.qxd 3/19/2004 5:51 PM Page 786
cultural drivers of, 351–359, 351 (table), societal practices/values and, 362–363, 362
359 (table), 386 (table), 364, 365–366 (tables), 370–371
cultural norms, feedback loop and, 350–351, societal values factors, 117
352, 377 uncertainty avoidance and, 624 (table), 625
cultural/societal effectiveness measure, 37–39, values-practices relationship, statistical analysis
38–41 (tables) of, 30–32, 31 (table), 33–37 (figures)
decision-making authority and, 356–357 women’s economic/political participation,
definition of, 12, 30 379–380, 380 (table), 387
economic development and, 355–356, 387 See also Organizational cultural practices
economic health and, 364–367, Gender roles, 295, 343, 349–350
368 (table), 387 Generalizability, 99, 100, 133, 146
future orientation and, 319, 319 (table) Geographic features. See Regional clustering
future orientation, correlation of, 309–311, Georgia, 12, 107, 129, 186, 269, 304, 306, 326
309 (table) assertiveness and, 409, 410–411 (tables)
gender differences, CLT leadership ratings and, collectivism rankings, 468–471
384–385, 388–389 gender egalitarianism and, 365–366 (tables)
gender differences, cultural practices/values and, humane orientation and, 573–574 (tables)
385–386, 389 power distance scores, 539–540 (tables)
gender parity/unity in diversity and, 166 response bias scales, 740–749 (tables)
gender-role ideologies, 349–350, 378–379, uncertainty avoidance and, 622–623 (tables)
378 (table), 388 Germanic Europe, xvi, 185
gender stereotypes and, 349, 377–378, assertiveness and, 423, 424 (table), 427,
378 (table), 388, 400 427 (table)
GLOBE findings on, 360–364 climatic effects and, 212–215, 213–214 (tables)
GLOBE measures of, 359–360, collectivism practices/values scores and,
360–361 (tables), 387 476–481, 477 (table), 479 (table)
human condition and, 367–369, 368 (table), 387 cultural dimension scores, 192–200, 193–194
humane orientation and, 575, 576 (table) (tables), 195–199 (figure)
individual values/perceptions, cluster/society cultural values-practices relationship, 32,
effects on, 200–201, 202 (table) 33 (figure)
industry scores, overall scores and, 363–364, future orientation and, 310, 322–323, 322 (table)
363 (tables), 367 (table) gender egalitarianism and, 374, 375 (table)
masculinity/femininity dimension and, 344–348, humane orientation and, 580–581, 582 (table),
374–377, 388 585, 587, 587 (table)
organizational culture and, 380–381, 388 humane-oriented leadership and, 591–592,
organizational/societal values-practices 592 (table)
relationship and, 381, 383–384 individual values/perceptions, cluster effect on,
parental investment theory and, 351–352, 200–201, 202 (table)
353, 386 leadership dimension scores, 42, 42–45 (figures),
political ideology and, 369–370, 369 (table), 493, 494 (table)
387–388 metaconfiguration, cultural clusters and, 200,
political systems and, 358–359 201 (figure)
power distance and, 544, 544 (table), 545 performance orientation and, 261–264,
power use and, 516–517 262 (table), 271 (table)
production modes/control, social status and, power distance and, 547–549, 548 (table),
357–358 550 (table)
psychological perspective on, 348 uncertainty avoidance and, 637 (table)
regional clusters and, 374, 375–376 (table), 388 visionary leadership and, 327–328, 328 (table)
regional clusters, classification of, 192–200, workplace reforms, 584
193–194 (tables), 198 (figure) See also Culturally endorsed implicit theories of
religion, role of, 353–355, 386–387 leadership (CLTs); Regional clustering
resource control, social structure and, Germany, 12, 61, 106, 129, 184, 185, 266, 269, 304,
356–357, 387 306, 326
resource dependency theory and, 353 assertiveness and, 400, 401, 409, 410–411 (tables)
response bias scales, 742–749 (tables) collectivism rankings, 468–471
Subject-Ind.qxd 3/19/2004 5:51 PM Page 789
Hofstede/GLOBE scales correlation,138–141, collectivist scales and, 470, 472, 472–473 (tables)
140 (table) corporate social responsibility and, 585
Holistic viewpoint, 323 cross-cultural studies on, 569–571
Hong Kong, 12, 54, 58, 106, 108, 129, 242, 269, cultural dimensions and, 575, 576 (table)
286, 287, 304, 306, 326 cultural/societal effectiveness measures, 37–39,
assertiveness and, 409, 410–411 (tables) 38–41 (tables)
collectivism rankings, 468–471 definition of, 13, 30
gender egalitarianism and, 365–366 (tables) economic health and, 577–578
humane orientation and, 573–574 (tables), 589 family practices/norms and, 566–567, 578, 596
power distance scores, 539–540 (tables) future orientation, correlation with, 309–311,
response bias scales, 740–749 (tables) 309 (table)
uncertainty avoidance and, 622–623 (tables) GLOBE findings on, 572–575, 573–574 (tables)
Human condition factors, 113 (table) GLOBE measures of, 571, 571–572 (tables)
assertiveness and, 416–417, 417 (table) government role and, 567–568, 596
future orientation and, 316, 316 (table) helpfulness measures, 577
gender egalitarianism and, 367–369, 368 (table) high vs. low humane orientation societies,
humane orientation and, 577, 579 (table) 569–571, 570 (table)
human health, 113 human condition and, 577
individualism/collectivism and, 484–485, human rights/discrimination and, 567, 579–580
484–485 (tables) individual values/perceptions, cluster/society
life expectancy, 113 effects on, 200–201, 202 (table)
performance orientation and, 256–257, 256 (table) industry scores and, 582–583, 583 (tables), 585,
power distance and, 557–559, 558 (table) 586–587 (tables), 591 (table)
psychological health, 114 labor processes and, 581–582, 584–585, 596
satisfaction, 113–114 leadership characteristic of, 587–592,
societal health, 113 590–592 (tables), 596
Human development, xv, 37, 38–39, 40–41 (tables) masculinity/femininity dimension and, 566
Human Development Index (HDI), 14, 25, 39, modernization measures and, 577–578,
40–41 (tables), 105–106, 106–107 (table), 579 (table)
114, 256–257, 316, 368, 369, 484–485, motivation bases, social norms, 565–566, 596
556–557, 632 organizational culture and, 583–587,
Human Development Report (HDR), 37, 105–106, 586–587 (tables), 596
105 (table), 114 paternalism/patronage and, 566, 571, 584, 589, 596
Human nature as good vs. human nature as bad political systems/ideology and, 576–577
dimension, 13 power distance and, 544, 544 (table), 545
Human resource management (HRM), 78 public morality and, 165, 565
bureaucratic organizations and, 584 regional clusters and, 580–581, 582 (table), 585,
expatriate failures and, 203 587, 587 (table), 591–592, 592 (table), 596
humane orientation and, 584–585 regional clusters, classification of, 192–200,
individualism/collectivism constructs and, 193–194 (tables), 196 (figure)
453–455 religious influence and, 565
regional clustering and, 201–203 response bias scales, 742–749 (tables)
staffing decisions, 201–202, 613–614, 640, self-transcendence/self-enhancement dimensions
709, 710 and, 565–566
strategies in, 82 social support, sources of, 569
training programs, 202–203, 568 uncertainty avoidance and, 624 (table)
See also Career development values-practices relationship, statistical analysis
Human rights, 567, 579–580 of, 30–32, 31 (table), 33–37 (figures)
Humane orientation, xiv, 12, 564, 595–596, 597, 675 wealth-sharing and, 568–569
assertiveness and, 414, 415 (table) See also Organizational cultural practices
climatic influence on, 208, 210–211, 580, Humane-oriented leadership, xv, xvi, 14
581 (table) assertiveness cultural practices/values and,
CLT leadership dimensions, humane 429 (figure), 430, 432
orientation cultural values and, 592–595, collectivism practices/values scores and,
593 (figure), 597 496–497 (figures), 498, 499, 501
Subject-Ind.qxd 3/19/2004 5:51 PM Page 792
cultural dimension values and, 45, 47 (figure) strategic planning and, 291
future orientation cultural practices/values and, uncertainty avoidance and, 622–623 (tables)
330 (figure), 331 Individualism/collectivism constructs, 60, 62,
gender differences and, 384–385 64, 437
gender egalitarianism cultural practices and, 383 accountability, organizational members and,
humane societal orientation and, 587–592, 457–458
590–592 (tables), 593 (figure), achievement and, 485–486, 487 (tables)
594–595, 597 behavioral approach to leadership and, 458–460
leadership dimension scores, 41–42, 44 (figure) charismatic leadership and, 461
performance orientation cultural values and, CLT leadership dimensions, collectivism cultural
274–275, 274 (figure) values and, 491–501, 492 (table),
power distance cultural values and, 552 (figure), 496–497 (figures), 496 (table)
553–555 CLT leadership dimensions, culture cluster
regional clusters and, 591–592, 592 (table) rankings and, 493, 494–495 (table)
uncertainty avoidance cultural practices/values cognitive perspectives on leadership and,
and, 642 (figure), 643–644 460–461
See also Culturally endorsed implicit theories of collectivism syndromes/cultural syndromes and,
leadership (CLTs) 472–474, 476
Hungary, 12, 107, 109, 129, 186, 269, 295, 304, conformity and, 457
306, 326 cooperative team processes and, 457
assertiveness and, 410–411 (tables) country collectivism rankings, 467–470,
collectivism rankings, 468–471 468–471 (tables)
gender egalitarianism and, 365–366 (tables) cultural dimensions, correlation with, 470,
humane orientation and, 573–574 (tables) 472–474, 472–473 (tables)
power distance scores, 539–540 (tables) ecological context and, 450
response bias scales, 740–749 (tables) economic health and, 481–483, 482 (tables)
uncertainty avoidance and, 622–623 (tables) embeddedness value scale and, 475, 475 (table)
empirical research on, 440–450
Iceland, 109, 184, 185 employee motivation and, 455–456
Idiocentrism, 62, 444, 448 entrepreneurial activity and, 458
Imperialism, 187 family/friendships and, 485, 486 (tables)
Implicit leadership theory (ILT), 16–17, 18, 56, family systems and, 452
145, 669 features of, 443
conceptual foundation of, 670–673 gender equality and, 489, 490 (tables)
congruence hypothesis and, 204 GLOBE findings, 466–491, 502–503
cross-cultural differences, leadership profiles GLOBE measures of, 463–466, 464–466 (tables)
and, 672–673 GLOBE research on, 461–463
cultural level of analysis and, 669–670 group processes and, 456–457
leadership belief systems, commonalities in, health-related correlates of, 451–452
671, 672 horizontal/vertical social relationships and,
shared mental models and, 671, 697 443–444
See also Culturally endorsed implicit theories of human condition and, 484–485, 484–485 (tables)
leadership (CLTs) human resource management practices and,
Implicit motivation theory, 17 453–455
India, 12, 58, 61, 65, 97, 107, 109, 129, 189, 269, individualism scale and, 474–475,
304, 306, 326 475 (table), 501
assertiveness and, 409, 410–411 (tables) individual-level studies of, 448–449
collectivism rankings, 439, 468–471 industry practices/values scores, 490, 491 (table)
gender egalitarianism and, 365–366 (tables) intellectual/affective autonomy scales and, 475,
humane orientation and, 573–574 (tables) 475 (table)
leadership research in, 62–63 job attitudes and, 456
power distance and, 518, 522, language/communication and, 452
539–540 (tables) leadership, relationship to, 458–461, 462 (table)
response bias scales, 740–749 (tables) legal institutions and, 438
Southeast Asian nations and, 188 national wealth and, 450–451
Subject-Ind.qxd 3/19/2004 5:51 PM Page 793
organizational-level correlates of, 453–458, gender egalitarianism and, 363–364, 363 (tables),
459 (table) 367 (table)
organizational-level studies of, 446–448 leader behaviors and, 17–18
organizational/societal values-practices Nordic/Japanese reform models and, 583–584
relationship and, 491–492 Taylorist-Fordist paradigm in, 581–582, 584, 596
organizational trust, psychological contracts uncertainty avoidance strategies, 605–607
and, 457 See also Institutional systems; Organizational
pace of life and, 451 cultural practices; Organizational culture;
political ideology and, 439–440, 486–487, Organizational leadership
488 (tables)
regional clusters and, 476–481, 477–480 (tables) Inequalities. See Power distance
religious dogma/devotion and, 487, 489, In-group collectivism, xiv, xv, 13, 45
489 (tables) assertiveness and, 415, 415 (table)
religious institutions and, 438–439 climatic influence on, 208, 210–211
science/technology success and, 483, 483 (tables) cultural/societal effectiveness measures, 37–39,
social interaction patterns and, 452–453 38–41 (tables)
social loafing and, 456 definition of, 12, 30
societal collectivism scales, 466–467, 467 (tables) family/national pride and, 165–166
societal-level correlates of, 450–453 future orientation, correlation of, 309–311,
societal-level studies of, 440–446, 454 (table) 309 (table)
Individuals, xiv, 13 humane orientation and, 575, 576 (table)
adaptive capacity among, 299 individual values/perceptions, cluster/society
age, goal selection and, 298 effects on, 200–201, 202 (table)
assertiveness training and, 397–398 power distance and, 544, 544 (table)
contextual correlates of future orientation, regional clusters, classification of, 192–200,
299–300 193–194 (tables), 197 (figure)
educational levels and, 295 response bias scales, 742–749 (tables)
future orientation and, 292–293, 298–300 uncertainty avoidance and, 623, 624 (table), 625
general correlates of future orientation, 298–299 values-practices relationship, statistical analysis
gratification, delay of, 299–300 of, 30–32, 31 (table), 33–37 (figures)
individual values/perceptions, cluster/society See also Individualism/collectivism constructs;
effects and, 200–201, 202 (table) Organizational cultural practices
internal locus of control and, 244 Innovation, 606–607, 616, 617–618
investment behaviors and, 299 Institutional collectivism, xiv, xv, 13, 45
psychological health of, 300 assertiveness and, 415, 415 (table)
relationship development and, 300 climatic influence on, 208, 210–211
societal variables and, 293 cultural/societal effectiveness measures, 37–39,
socioeconomic condition and, 298–299 38–41 (tables)
time perspective, incentive/motivation of, 299 definition of, 12, 30
uncertainty avoidance and, 603–605 future orientation, correlation of, 309–311,
upbringing factors and, 298 309 (table)
Indonesia, 12, 107, 109, 129, 188, 189, 269, 286, humane orientation and, 575, 576 (table)
304, 306, 326 individual values/perceptions, cluster/society
assertiveness and, 409, 410–411 (tables) effects on, 200–201, 202 (table)
collectivism rankings, 468–471 power distance and, 544, 544 (table)
gender egalitarianism and, 365–366 (tables) regional clusters, classification of, 192–200,
humane orientation and, 573–574 (tables) 193–194 (table), 197 (figure)
power distance scores, 539–540 (tables) response bias scales, 742–749 (tables)
response bias scales, 740–749 (tables) societal respect and, 165
uncertainty avoidance and, 622–623 (tables) uncertainty avoidance and, 623, 624 (table),
Industrial organizations, 10 625–626
x-inefficiencies, mass production and, 582 values-practices relationship, statistical analysis
cultural convergence and, 53–54 of, 30–32, 31 (table), 33–37 (figures)
future orientation scales and, 312 (table), See also Individualism/collectivism constructs;
314, 332 Organizational cultural practices
Subject-Ind.qxd 3/19/2004 5:51 PM Page 794
Personality measures, 399, 566, 603–604 CLT leadership dimensions, power distance
Peru, 109 cultural values and, 551–555, 552 (figure)
Philippines, 12, 107, 109, 130, 189, 269, 304, 306 collectivist scales and, 470, 471, 472–473 (tables)
assertiveness and, 409, 410–411 (tables) communism and, 525
collectivism rankings, 468–471 cross-cultural stream and, 517–526
gender egalitarianism and, 364, 365–366 (tables) cultural dimensions and, 544–545,
humane orientation and, 573–574 (tables) 544 (table), 559
power distance scores, 539–540 (tables) cultural/societal effectiveness measures, 37–39,
response bias scales, 740–749 (tables) 38–41 (tables)
uncertainty avoidance and, 622–623 (tables) definition of, 12, 30
visionary leadership in, 325, 326 democratic principles of government and,
Physical health. See Health status 523–525
Pilot tests, xiii, 127–128, 129–130 (table) family indoctrination, power values and,
Pink-collar ghettos, 364, 387 526–527
Poland, 12, 61, 107, 109, 130, 186, 269, 304, future orientation, correlation of, 309–311,
306, 326 309 (table)
assertiveness and, 410–411 (tables) gender, power use and, 516–517
collectivism rankings, 468–471 GLOBE conceptualization/measures of,
gender egalitarianism and, 365–366 (tables) 537–538, 537–538 (tables)
humane orientation and, 573–574 (tables) GLOBE findings on, 538–543, 539–540 (tables)
power distance scores, 539–540 (tables) high vs. low power distance societies, 527–536,
response bias scales, 740–749 (tables) 536 (table)
uncertainty avoidance and, 622–623 (tables) human condition and, 557–559, 558 (table)
Political ideology, 115–116 humane orientation and, 575, 576 (table)
assertiveness and, 417–418, 418 (table) immigration load and, 525–526
domestic political violence, 422 individualism/collectivism indices and,
future orientation and, 318, 318 (table) 531–532, 533
gender egalitarianism and, 369–370, 369 (table) individual values/perceptions, cluster/society
humane orientation and, 576–577 effects on, 200–201, 202 (table)
individualism/collectivism and, 439–440, industry-level statistics, 545–547,
486–487, 488 (tables) 545–546 (tables)
liberal vs. socialist ideology, 255, 576 isomorphic pressures and, 84
performance orientation practices and, 257, knowledge-intensive organization and, 534
258 (table) leadership strategies and, 535–536
uncertainty avoidance and, 631–632, 632 (table) middle class, role of, 525
Political involvement, 164 monopolistic orientation and, 166
Political stability, 294–295 need for power and, 515–516
Polychronicity, 289, 290, 291 organizational power distance, 541–542,
Portugal, 12, 32, 61, 106, 109, 130, 184, 269, 304, 542 (table), 545–547, 546 (table), 548–549,
306, 326 550 (table)
assertiveness and, 411–412 (tables) organizational/societal practices-values
collectivism rankings, 468–471 relationship, 542–543
gender egalitarianism and, 365–366 (tables) organizational structure and, 533–534
humane orientation and, 573–574 (tables) participation levels and, 61
power distance scores, 539–540 (tables) power distance index and, 528–531, 530 (table),
response bias scales, 740–749 (tables) 543, 543 (table)
uncertainty avoidance and, 622–623 (tables) power sources, 514–515
Power distance, xiii, xiv, xv, 13, 45, 513–514 psychological stream, power and, 514–517
achieved vs. ascribed status and, 532 regional clusters and, 547–549, 547–548 (tables),
assertiveness and, 413, 414, 415 (table) 550 (table), 559–560
autonomy/self-actualization measures, 527–528, regional clusters, classification of, 192–200,
528 (table) 193–194 (tables), 198 (figure)
certification processes and, 86–87 religion/philosophies and, 518–523, 526, 559
Chinese Culture Connection study and, 531 response bias scales, 742–749 (tables)
climatic influence on, 208, 210–211, 549–551 role overload and, 531
Subject-Ind.qxd 3/19/2004 5:51 PM Page 801
visionary leadership and, 327–328, 328 (table) performance orientation and, 256
within-society heterogeneity and, 183 relationships and, 300
work-related variables and, 182–183 uncertainty avoidance and, 631, 632 (table)
See also Climatic clustering Saudi Arabia, 61, 63, 155
Relationship development, 300 Savanna climate. See Tropical wet/dry climate
Relationship-oriented leadership, 60 Savings rates, 287–288, 319–320, 320 (table)
Reliability, xiii, 79 Scales. See GLOBE culture and leadership scales;
See also Construct validation process GLOBE research design; GLOBE scales
Religion, xiv, 59, 60, 63, 116–117, 188 development; GLOBE scales validation;
assertiveness and, 417–418, 418 (table) Questionnaires
future orientation and, 295–296, 318–319, Schwartz Value Survey/GLOBE scales
319 (table) correlation,141–142, 143 (table)
gender egalitarianism and, 353–355 Science success factors, 110, 112, 112 (table)
humane orientation and, 565 future orientation and, 317, 317 (table)
individualism/collectivism and, 438–439, 487, individualism/collectivism and, 483, 483 (tables)
489, 489 (tables) power distance and, 557 (table)
performance orientation practices and, 258–259 uncertainty avoidance and, 633, 633 (table)
power distance and, 518–523 Selectivity theory, 293, 298
uncertainty avoidance and, 633–634, 635 (table) Self-actualization, 242, 527–528, 528 (table)
See also Catholicism; Confucianism; Islamic Self-concept, 268, 272, 275
religion; Protestantism; Protestant work Self-protective leadership, xv, xvi, 7, 14, 675
ethic; Spiritual attitudes assertiveness cultural practices/values and, 430
Research. See Cross-cultural research; GLOBE collectivism practices/values scores and,
research design; GLOBE research project 496–497 (figures), 498, 500, 501
Resource allocation, 289–290, 356–357, 576 cultural dimension values and, 45, 48 (figure)
Resource dependency theory, 80–81, 85, 353 future orientation cultural practices/values and,
Respect, 165, 567 330 (figure), 331
Response bias. See Cultural response bias gender differences and, 384–385
Rewards/punishments, 60, 62 gender egalitarianism cultural values and,
contingent rewards, 453 382 (figure), 383, 384
group behavioral norms and, 79 humane orientation cultural values and,
See also Compensation practices 593 (figure), 595, 597
Risk-taking, 612, 616–617 leadership dimension scores, 41, 45 (figure)
Roman Catholicism. See Catholicism performance orientation cultural values and,
Russia, 12, 55, 61, 107, 109, 130, 186, 269, 304, 274 (figure), 275
306, 326 power distance cultural practices/values and,
assertiveness and, 410–411 (tables) 552 (figure), 553–555
collectivism rankings, 468–471 uncertainty avoidance cultural practices/values
gender egalitarianism and, 365–366 (tables) and, 642 (figure), 644
humane orientation and, 573–574 (tables) See also Culturally endorsed implicit theories of
power distance scores, 539–540 (tables) leadership (CLTs)
response bias scales, 740–749 (tables) Self-transcendence/self-enhancement, 565–566
uncertainty avoidance and, 622–623 (tables) Shared practices, 79
Sikh religion, 634
Salzburg Seminar Alumni Study (SSAS), 345, Singapore, 12, 65, 106, 108, 130, 269, 286, 291,
346, 611 304, 306, 326
Satisfaction, xiv, 38–39, 40–41 (table), assertiveness and, 409, 410–411 (tables)
113–114 (table) collectivism rankings, 468–471
cultural norms, violation of, 64 gender egalitarianism and, 365–366 (tables)
future orientation and, 316 humane orientation and, 573–574 (tables)
gender egalitarianism and, 368, 369 power distance scores, 539–540 (tables)
individualism/collectivism and, 484–485, response bias scales, 740–749 (tables)
484–485 (tables) uncertainty avoidance and, 622–623 (tables)
job satisfaction, 456 Slovakia, 109
leader behaviors and, 60, 62 Slovenia, 12, 106, 130, 186, 269, 304, 306, 326
Subject-Ind.qxd 3/19/2004 5:51 PM Page 803
assertiveness and, 409, 410–411 (tables) Societal health, 38–39, 40–41 (tables),
collectivism rankings, 468–471 113 (table), 256
gender egalitarianism and, 365–366 (tables) gender egalitarianism and, 368, 369
humane orientation and, 573–574 (tables) individualism/collectivism and, 484–485,
power distance scores, 539–540 (tables) 484–485 (tables)
response bias scales, 740–749 (tables) relationship development and, 300
Smallest space analysis, 182 uncertainty avoidance and, 630, 632 (table)
Social constructionism, 79 Societal respect, 165
Social information processing, 80 Socioeconomic condition, 294, 298–299, 556,
Socialization process, 78, 80 557 (table)
sex differences and, 344–345 Socioemotional selectivity theory, 293, 298
societal-level future orientation and, 294 Sociotechnical orientation, 166–167
Social network theory, 79 South Africa, 12, 107, 109, 130, 180, 184, 188, 269,
cliques and, 80 304, 306, 326
cohesion perspective, 79, 80 assertiveness and, 410–411 (tables)
group boundaries and, 79 collectivism rankings, 468–471
international level and, 80 future orientation scores for, 310
socially constructed perceptions and, 79 gender egalitarianism and, 365–366 (tables)
structural equivalence perspective and, 79–80 humane orientation and, 573–574 (tables)
See also Organizational culture; Resource power distance scores, 539–540 (tables)
dependency theory response bias scales, 740–749 (tables)
Social protection systems, 568 uncertainty avoidance and, 622–623 (tables)
Social responsibility, 585 Southern Asia, xvi, 188–189
Social welfare, 242 assertiveness and, 423, 424 (table), 427,
Societal attitudes, 114, 115–116 (table) 427 (table)
achievement factors, 115 climatic effects and, 212–215, 213–214 (tables)
family/friends factors, 115 collectivism practices/values scores and,
gender equality factors, 117 476–481, 478 (table), 480 (table)
political ideology factors, 115–116 cultural dimension scores, 192–200, 193–194
religious factors, 116–117 (tables), 195–199 (figures)
See also Cultural values-practices relationship; cultural values-practices relationship, 32,
Regional clustering 35 (figure)
Societal clusters. See Regional clustering diversity, interactive coexistence and, 188–189
Societal culture, xiii, xiv, 3, 15, 37 future orientation and, 322–323, 322 (table)
cultural dimensions, values-practice correlations, gender egalitarianism and, 374, 376 (table)
733–736 humane orientation and, 580–581, 582 (table),
economic competitiveness and, 19 585, 587, 587 (table)
future orientation and, 284–286, 294–296 humane-oriented leadership and, 591–592,
health status and, 19 592 (table)
implicit theories of leadership and, 18 individual values/perceptions, cluster effect on,
leader behavior and, 1, 5, 6, 10, 17, 25 200–201, 202 (table)
masculinity/femininity dimension of, 344–348 leadership dimension scores, 42, 42–45 (figures),
operational definition of, 15–16 493, 495 (table)
organizational culture and, 18, 37 metaconfiguration, cultural clusters and, 200,
questionnaire on, 21–22, 23 (table) 201 (figure)
values-practices correlations, 733–736 performance orientation and, 261–264, 263
values-practices relationship, statistical analysis (table), 271 (table)
of, 30–32, 31 (table), 33–37 (figures) power distance and, 547–549, 548 (table),
See also Climatic clustering; Construct 550 (table)
validation; Cultural immersion theory; uncertainty avoidance and, 637 (table)
Cultural values-practices relationship; visionary leadership and, 327–328, 328 (table)
Cultural variation; GLOBE culture and See also Culturally endorsed implicit theories of
leadership scales; Organizational cultural leadership (CLTs); Regional clustering
practices; Organizational culture; Regional South Korea, 12, 130, 188, 269, 286, 304, 306,
clustering 326, 588
Subject-Ind.qxd 3/19/2004 5:51 PM Page 804
Robert J. House received his PhD degree in management from the Ohio State
University, June 1960. He was appointed the Joseph Frank Bernstein Professor
Endowed Chair of Organization Studies at the Wharton School of the University of
Pennsylvania in 1988. He has published 130 journal articles. In total, his articles have
been reprinted in approximately 50 anthologies of readings in management and organi-
zational behavior. He received the Award for Distinguished Scholarly Contribution to
Management, and four awards for outstanding publications. The awards were conferred
by the Academy of Management and the Canadian Association of Administrative
Sciences. He has also authored two papers that are Scientific Citations Classics.
He is a Fellow of the Academy of Management, American Psychological
Association, and Society for Industrial/Organizational Psychology. He has served as
chairperson of the Academy of Management Division of Organizational Behavior
(1972–1973) and President of the Administrative Science Association of Canada
(1985–1986).
He was the Principal Investigator of the Global Leadership and Organizational
Behavior Effectiveness Research Program (GLOBE) from 1993 through 2003. In this
capacity he visited universities in 38 countries. He has also been a visiting scholar or
visiting professor at 14 universities, most of which are in Europe or Asia.
His major research interests are the role of personality traits and motives as they
relate to effective leadership and organizational performance, power, and personality in
organizations, leadership, and the implications of cross-cultural variation for effective
leadership and organizational performance.
Paul J. Hanges is an Associate Professor of Industrial–Organizational Psychology and
Chair of the I/O Area at the University of Maryland. He received his PhD in indus-
trial–organizational psychology from the University of Akron and has been a faculty
member at the University of Maryland since 1986. His research interests center around
topics in social cognition, cross-cultural leadership, research methodology, and person-
nel selection. The majority of his work has focused on understanding factors affecting
social perceptions (e.g., leadership) and the factors (e.g., societal culture, gender stereo-
types, personality) that cause these perceptions to stabilize or change over time. He has
developed a dynamic measure of perception and explored the utility of various mathe-
matical models (e.g., catastrophe analysis, neural network analysis) to model changes
in leadership ratings over time. His methodology work has dealt with such topics as
multilevel statistical modeling, range restriction, and test banding methodology. He
joined the GLOBE project in 1993 and became a coprincipal investigator. His work has
appeared in such journals as Applied Psychological Measurement, Applied Psychology:
809
ATE-House.qxd 3/19/2004 5:55 PM Page 810
813
ATC-House.qxd 3/19/2004 5:57 PM Page 814
He has received many awards and honors including the Distinguished Scholar
Award, Management Department, College of Business Administration (2000), the Best
Paper Award from the International Division of the Academy of Management (1996),
the Distinguished Service Award from the East West Center (1989), and the Lum Yip
Kee Outstanding MBA Student Award from the College of Business Administration,
University of Hawaii (1990). He is a Founding Fellow of the International Academy of
Intercultural Research.
Muzaffer Bodur is a Professor of Marketing at the Management Department of
Bogaziçi University in Istanbul, Turkey. She received her doctor of business adminis-
tration from Indiana University in 1977 and acted as a Visiting Assistant Professor at
George Mason University upon graduation. In 1979 she joined Bogaziçi University
faculty where she teaches a global marketing management course to MBA students and
research methods courses to doctoral students. She has organized training programs and
seminars for executives and has served as the department head.
She is a member of the Academy of International Business and the Consortium for
International Marketing Research . Currently, she is the editor of Bogaziçi Journal:
Review of Social, Economic and Administrative Studies and serves on the editorial
board of Journal of International Marketing. She has visited Uppsala University of
Sweden and Odense University of Denmark to teach international marketing courses
and conducted cross-cultural research on the implications of business culture for inter-
nationalization of firms. Her publications focus on marketing strategies of multina-
tional firms in emerging markets; export marketing management; expatriate managers;
and consumer satisfaction, dissatisfaction, and complaining behavior with services and
intangible product.
Felix C. Brodbeck holds a doctoral degree from the University of Giessen, and a Post-
Doctorate (Doctor Habilitated) from the University of Munich. He is a Professor of
Organizational and Social Psychology and Head of the Department of Work and
Organisational Psychology at Aston Business School, Aston University, Birmingham,
England. He is a member of the coordination team of the GLOBE Program. He has
written or edited six books and authored more than 70 scholarly articles in the fields of
leadership, cross-cultural psychology, social psychology, group performance and group
decision making, team climate and effectiveness, innovation, human–computer inter-
action, and applied research methods in organizations. His repertoire of experience and
practice comprises experimental, applied, and field research; development of theory
and practical tools; as well as executive training and business consulting in the above
areas and human resource management in general.
Dale Carl is an Associate Professor in the School of Business Management at Ryerson
University in Toronto, Canada. He has a BA degree in history, an MBA in international
business, and a PhD in international business and strategy from the University of
Calgary. He was a diplomat in the Canadian Foreign Service, with postings in Pakistan,
Norway, the United States, and Iraq. Subsequently, he lived in Dubai, working in
the private sector, with a territory that included the Middle East and East Africa.
His research interest is comparative cross-cultural leadership. He has coauthored
papers titled “Motivational Consequences of Charismatic Leadership: An Empirical
Investigation,” “East Meets West: Exploring Charismatic Leadership Among Canadian
and Iranian Executives,” and “Universality of Charismatic Leadership: A Multi-County
Study.”
Deanne N. Den Hartog holds a master of science and a doctoral degree in organiza-
tional psychology from the Free University in Amsterdam, the Netherlands. She is
ATC-House.qxd 3/19/2004 5:57 PM Page 815
Bogaziçi Journal: Review of Social, Economic and Administrative Studies and on the
editorial boards of several international and national journals focusing on management
and organizational studies.
Melinda S. Mayfield is a Compensation Analyst in Human Resources Strategy at
Cerner Corporation in Kansas City, Missouri, and was a PhD candidate in organiza-
tional behavior at the Krannert Graduate School of Management at Purdue University
at the time this research was conducted. She earned an MBA in human resources
management from the University of Missouri-Columbia, and a BS in economics and
business administration from William Jewell College in Liberty, Missouri.