Objectives: Most studies of risk factors for human immunodeficiency virus (HIV) rely heavily on retrospective self-reports. The degree to which these reports provide reliable information has received little research attention. The purpose of this study was to assess the extent to which the reliability of retrospective self-report data is affected over time.
Methods: Data were examined from a longitudinal study of risk behaviors among injection drug users. Structured interviews were administered to 366 injection drug users who were asked to recall behaviors that they had reported 6, 12, or 18 months earlier.
Results: Kappa coefficients showed moderate initial memory loss for injection and sexual risk behaviors. After 6 months, time had little effect on the ability of injection drug users to recall injection and sexual risk behaviors. For ordinal measures of risk, subjects who gave different reports over time were most likely to disagree by one response category.
Conclusions: These findings support the use of retrospective self-reports in assessing HIV risk behaviors among injection drug users. Retrospective inquiry is likely to remain an important method for collecting data over time.