Researchers must first consider extrascientific factors -- costs, ethics, and confidentiality -- when choosing a measurement strategy for a particular drug prevention study. Methodological issues refer to problems of outcome, definition, measurement error, and measurement bias. Prevention researchers have often conceptualized their interventions as changing the transition from nonuser to user, which usually occurs in adolescence. The standard statistical model for estimating the dependence of incidence rates on experimental variables is the proportional hazards model. The most frequently used outcomes in the evaluations of drug abuse prevention interventions include continuous dimensions, categories, ordered categories, stages, and indices. When evaluating program outcomes, researchers must distinguish between alternative models of drug use: the separate drug model, polydrug model, and index model. Other methodological issues that affect outcome measurements include quantity and frequency of use, multiple indicator models of measurement error, biochemical indicators, and conditional versus unconditional models of change in longitudinal studies. 16 references