The reliability of subjects' reports on stressful life events inventories: a longitudinal study.

Previous studies of the reliability of stressful life events inventories employing test-retest, informant report, and fall-off designs have all suffered from significant methodological limitations. In order to provide a more rigorous test of the reliability of subjects' reports on life events inventories, the present study employed a longitudinal design with multiple, frequent assessments. Fifty-three subjects completed life events questionnaires every 5 weeks for 5 months. At the end of this time, subjects were asked to report all events occurring during the entire study period. The intraclass correlation between the number of events reported in the four 5-week assessments and the final, 5-month assessment was .63. There was a 26% decline in the number of events reported between the 5-week assessments and the 20-week assessment. Agreement on specific inventory items was only 52%. These data are consistent with previous, less methodologically rigorous studies in raising serious questions about the reliability of stressful life events questionnaires.