The same measures of behavioral self-report for an overlapping period of 4-5 years were collected at each interview and were analyzed within a test-retest design. Results revealed reasonably good agreement between measured obtained at the two recall points, except for low-frequency behaviors and for measures of income and monetary costs. The absolute difference levels of socially undesirable activities showed systematic discrepancies, increasing with proximity to the interview, a trend most likely due to concealment. On the other hand, all variables may have been affected by memory loss and inadequate memory recall strategies. The correlation coefficients of the intervariable relationships measured at each interview increased as the constituent item reliabilities increased. In addition, urinalyses conducted at each interview generally validated the reports of recent narcotics use. Findings suggest that quality information can be obtained from identified heroin abusers under appropriate conditions. Tables, figures, and 18 references