Social influence is central to models of adolescent substance use. Nonetheless, researchers fail to delineate the various forms of social influence. A framework that distinguishes between active (explicit drug offers) and passive (social modeling and overestimation of friends' use) social pressure was tested. The effect of these processes on alcohol and cigarette use was examined with 526 seventh graders taking part in an alcohol prevention program. Hierarchical regression analyses demonstrated that pretest measures of alcohol use, offers, modeling, and overestimation each accounted for unique variance in posttest alcohol use. Similar results were obtained for cigarette smoking. The general model was not significantly different for boys and girls, or for prior users and prior nonusers, supporting the generalizability of the framework. Implications for intervention programs are discussed.