This study addresses replication in candidate gene × environment interaction (cG×E) research by investigating if the key findings from Brody, Beach, Philibert, Chen, and Murry (2009) can be detected using data (N = 1,809) from the PROSPER substance use preventive intervention delivery system. Parallel to Brody et al., this study tested the hypotheses that substance misuse initiation would increase faster from age 11 to age 14 and be higher at age 14 among: (a) 5‐HTTLPR short carrier adolescents versus long homozygotes, (b) control versus intervention adolescents, and (c) 5‐HTTLPR short carriers in the control condition versus all other participants. The hypotheses were generally supported and results were consistent with Brody et al.'s cG×I finding. Results are discussed in light of replication issues in cG×E research and implications for intervention.