Although research has documented the positive effects of family-based prevention programs, the field lacks specific information regarding why these programs are effective. The current study summarized the effects of family-based programs on adolescent substance use using a component-based approach to meta-analysis in which we decomposed programs into a set of key topics or components that were specifically addressed by program curricula (e.g., parental monitoring/behavior management,problem solving, positive family relations, etc.). Components were coded according to the amount of time spent on program services that targeted youth, parents, and the whole family; we also coded effect sizes across studies for each substance-related outcome. Given the nested nature of the data, we used hierarchical linear modeling to link program components (Level 2) with effect sizes (Level 1). The overall effect size across programs was .31, which did not differ by type of substance. Youth-focused components designed to encourage more positive family relationships and a positive orientation toward the future emerged as key factors predicting larger than average effect sizes. Our results suggest that, within the universe of family-based prevention, where components such as parental monitoring/behavior management are almost universal, adding or expanding certain youth-focused components may be able to enhance program efficacy.