Recent studies have suggested that a prevention program that addresses the social influences that encourage smoking can be effective in deterring cigarette use by adolescents. This study presents 1-, 2-, and 3-year follow-up results from two studies which evaluated three variations of the social influences curriculum and compared them to a health consequences program and a usual-care comparison group. These results suggest that a peer-led, social influences program can restrain smoking among both baseline nonsmokers and baseline experimental smokers at 2 years postintervention. Analyses of attrition data suggest no evidence to threaten the internal validity of these findings, although their generalizability to baseline smokers may be limited.