Many reports of successful school-based intervention programs suffer from an overestimation of the real effects, errors which many times arise from statistical technicalities. An illustration of such overestimations, as seen through a reanalysis of data on a successful program, is reported in this paper. The data were reevaluated using a variety of statistical techniques, such as multilevel analysis. The result of the analyses shows that the reported effects of the normative school-based drug prevention program could not be replicated. The subsequent search for moderator effects of the program, such as a lowering effect on the relationship between the pre- and post-test, or on the relationship between respondents' use and their friends' use, was not successful either. More successful was a search for individual characteristics that show significant relationships with respondents' alcohol use, such as school problems, low grades, and rebelliousness. (RJM)