The purpose of this study was to determine the validity of the recently modified patient generated index (PGI) when used as a self-completed postal questionnaire by older people with arthritis. Two postal surveys were used, followed by interviews with a theoretically chosen sub-sample of respondents to both. Data from the self-completed modified PGI and the arthritis impact measurement scales (AIMS) and the interviews were assessed against 10 pre-defined validity criteria. A total number of 1027 people over 65 years old and with arthritis were surveyed and 53 were selected for interview. The response rates to the two-part survey were 78 and 83%. Complete data for the validity analysis were available from 47% of the respondents. The modified PGI met four and failed to meet six validity criteria. Pre-defined correlations were found with the AIMS and differentiation was possible between respondents with mild, moderate and severe arthritis and between those who had recently sought medical attention and those who had not. The areas that had been chosen for inclusion in the modified PGI were affected by arthritis but may not have been the 'most important' areas. The measurement method could not detect changes in health nor differentiate between respondents taking and not taking analgesics. There were several problems in the respondents' interpretation of the instructions, which were only revealed during the interviews. These data should help inform further modification of the PGI.