Psychometrically improved, abbreviated versions of three classic measures of impulsivity and self-control

Psychol Assess. 2014 Sep;26(3):1003-20. doi: 10.1037/pas0000003. Epub 2014 Jun 2.

Abstract

Self-reported impulsivity confers risk factor for substance abuse. However, the psychometric properties of many self-report impulsivity measures have been questioned, thereby undermining the interpretability of study findings using these measures. To better understand these measurement limitations and to suggest a path to assessing self-reported impulsivity with greater psychometric stability, we conducted a comprehensive psychometric evaluation of the Barratt Impulsiveness Scale-11 (BIS-11), the Behavioral Inhibition and Activation Scales (BIS/BAS), and the Brief Self-Control Scale (BSCS) using data from 1,449 individuals who participated in substance use research. For each measure, we evaluated (a) latent factor structure, (b) measurement invariance, (c) test-criterion relationships between the measures, and (d) test-criterion relations with drinking and smoking outcomes. Notably, we could not replicate the originally published latent structure for the BIS, BIS/BAS, or BSCS or any previously published alternative factor structure (English language). Using exploratory and confirmatory factor analysis, we identified psychometrically improved, abbreviated versions of each measure: 8-item, 2-factor BIS-11 (root-mean-square error of approximation [RMSEA] = .06, comparative fit index [CFI] = .95); 13-item, 4-factor BIS/BAS (RMSEA = .04, CFI = .96); and 7-item, 2-factor BSCS (RMSEA = .05, CFI = .96). These versions evidenced (a) stable, replicable factor structures, (b) scalar measurement invariance, ensuring our ability to make statistically interpretable comparisons across subgroups of interest (e.g., sex, race, drinking/smoking status), and (c) test-criterion relationships with each other and with drinking/smoking. This study provides strong support for using these psychometrically improved impulsivity measures, which improve data quality directly through better scale properties and indirectly through reducing response burden.

Publication types

  • Research Support, N.I.H., Extramural

MeSH terms

  • Adult
  • Alcohol-Related Disorders / epidemiology
  • Alcohol-Related Disorders / psychology
  • Binge Drinking / epidemiology*
  • Binge Drinking / psychology
  • Factor Analysis, Statistical
  • Female
  • Humans
  • Impulsive Behavior*
  • Inhibition, Psychological*
  • Male
  • Middle Aged
  • Personality
  • Personality Inventory
  • Psychometrics
  • Reproducibility of Results
  • Risk Factors
  • Self Report*
  • Smoking / epidemiology*
  • Smoking / psychology
  • Young Adult