Do successful drug prevention programs suppress the risk factors they were intended to modify? This paper addresses that issue for Project ALERT, a school-based program for seventh and eighth graders that has been shown to curb both cigarette and marijuana use. Evaluated with over 4,000 students in an experimental test that included 30 diverse California and Oregon schools, the curriculum seeks to help young people develop both the motivation to avoid drugs and the skills they need to resist pro-drug pressures. Using regression analyses, we examine the program's impact on the intervening (cognitive) variables hypothesized to affect actual use: adolescent beliefs in their ability to resist, perceived consequences of use, normative perceptions about peer use and tolerance of drugs, and expectations of future use. The analysis depicts program effects for perceptions linked to each target substance (alcohol, cigarettes, and marijuana), across all students and for those at different levels of risk for future use. Results show that the curriculum successfully dampened cognitive risk factors from each of the above categories for both cigarettes and marijuana, indicating that social influence programs can mitigate a broad range of beliefs associated with the propensity to use drugs. However, it had a limited impact on beliefs about alcohol, the most widely used and socially accepted of the three drugs. Implications for drug prevention programs and practitioners are discussed.