in adolescence but heavy use among friends in adulthood. This difference could serve as an alternative explanation for the developmental differences in findings regarding the moderating role of DRD4 genotype. It will be important to replicate the present results with identical measures of peer alcohol use over time. Self-reports of friends' alcohol use also represent a limitation, as they tend to be confounded by own levels of alcohol use; future studies should replicate these findings with different methodologies (e.g., experimental manipulation of peer drinking, friends' self-reports of own alcohol use). The sample size used in this study is relatively small (Duncan and Keller, 2011) and may not have been sufficient to detect smaller G × E effects or to expand the model by including additional moderators of the studied G × E interactions (e.g., sex, other genetic polymorphisms, or other personal or environmental characteristics). Finally, although the distributions of allele and genotype frequencies were similar to those of other Caucasian and European samples, ancestry informative markers were not collected and would be of benefit in future studies. Also of benefit would be to conduct genotyping in duplicate and to include concordance genotyping to increase the quality of the genotyping data.
