It is no surprise to educators and prevention specialists that for more than two decades, the majority of evaluation studies of individual psychosocial prevention strategies not only in substance-abuse prevention but in all of social science research have failed to find sustained positive outcomes (Feldman, 1983; Kreft & Brown, 1998). Many explanations have been proffered: lack of program intensity (many prevention activities run one hour per week for eight to 16 weeks), methodological limitations (such as one-time measurements), lack of long-term follow-up, and the use of myriad data analysis approaches with conflicting results. Findings have been difficult to generalize from program to program. Small numbers, different program strategies, target populations, outcome measures, intensities, implementations, and research designs have been research obstacles (Tobler 1986, p. 538). This has created confusion for prevention planners and policymakers who design and implement research-based programs. If evaluation and research are to be effective, systematic compilations of program effects are essential. For these reasons, Tobler conducted the first of three meta-analyses of alcohol, tobacco, and other drug-abuse prevention programs (hereafter referred to as drug prevention). Meta-analysis is a quantitative statistical procedure that synthesizes findings across many studies, overcoming the problems of small samples and diverse outcomes and programs. According to Tobler, the computation of the effect size is not dependent on statistically significant results which are seldom found in drug studies. Instead of discounting the studies whose results do not reach statistical significance, as would be the case in a literature review, the quantitative results of each study are converted into a common metric [effect size], thereby allowing comparison of results across studies (1996, p. 539). Meta-analysis, then, can provide answers to the most important questions in program planning and prevention: What works? How does it work? This article summarizes three of Tobler's meta-analyses that focused, respectively, on program content; program content and process; and program content, process, and size.