Abstract
Recent literature identifies many difficulties in validly measuring sex and drug use behavior associated with HIV risk, particularly among adolescents. In evaluating HIV risk reduction interventions, this literature emphasizes that, in addition to the more obvious determinants of self-report bias (e.g., response style, interviewer characteristics, social desirability, the interview setting), other factors require careful consideration (e.g., whether the person is assigned to a treatment or comparison group; whether the interview occurred at in-take, in treatment, or post-discharge; and the severily of the respondent's drug use). It also emphasizes the importance of including toxicologic STD and drug use measures in treatment outcome research to validate respondents' self-reports of risk behavior and adjust them for under-reporting. Guided by the literature, this article suggests procedures for more validly measuring HIV risk reduction, particularly among drug abusing adolescents. These include medical record abstraction, qualitative interviews, and extensive data validity checks to detect illogical or inconsistent responding.