ADAPTIVE EVALUATION STRATEGIES FOR ESTIMATING EFFECTS OF COMMUNITY-BASED DRUG-ABUSE PREVENTION PROGRAMS

Most evaluation methods applied to estimating effects of prevention pr ograms assume that a unit of experimental assignment or observation is static in the absence of intervention. This paper presents common pro blems encountered in evaluating the effectiveness of community-based p revention programs and evaluation strategies for addressing each probl em, assuming that a community is dynamic. The evaluation strategies ar e adaptive to dynamic changes that can be expected to occur in a commu nity at the levels of person, situation, and environment, as a result of normative development, historical events, and secular trends. In tu rn, these changes can affect design, sampling, program implementation, and outcomes. Illustrations of each type of change and how it can be dealt with in analyses are provided, based on experiences from the Mid western Prevention Project, a multi-community-based trial for drug abu se prevention. Illustrations include modeling the nested effects of so cial units within a community; adapting evaluation to changing units o f experimental assignment and analysis; evaluating program effects on stages of drug use transition; and estimating the ''true'' interventio n effect as a function of assignment and program implementation. Resea rchers and program evaluators are encouraged to rely on convergence of findings from multiple samples and analysis methods to interpret effe ctiveness of prevention programs.