We tested the effectiveness of a social resistance/competence enhancement approach to smoking prevention among predominantly Hispanic seventh graders (N = 3,153) from 47 New York City schools. After blocking on school type (public and parochial) and ethnic composition (percent Hispanic), schools were randomly assigned either to receive the 15-session prevention program or to serve as no-contact controls. Using the school as the unit of analysis, significant program effects were found for cigarette smoking, normative expectations concerning peer and adult smoking, smoking prevalence knowledge, social acceptability knowledge, and knowledge of smoking consequences. Using structural modeling techniques, a significant relation was found between the normative expectation and knowledge variables affected by the intervention and posttest smoking, suggesting that changes on these variables mediated the impact of the intervention on cigarette smoking. This study extends the results of previous prevention research and demonstrates the generalizability of this approach to predominantly Hispanic urban minority students.