predicted variable was measured. Because some variables were not normally distributed, the MLR estimator which is robust to violations of normality was used. The role of DRD4 genotype as a moderator of the relationships between friends' and own alcohol use was tested with multigroup modeling. Specifically, we compared the fit of a model where all paths and covariances were constrained to be equal for those with low and high risk genotypes to a model where the paths and covariances were free to vary across the two genotype groups. Significant results were followed up with multigroup testing of each single path and covariance, comparing a fully constrained model to a model where single path or covariance could vary by genotype. Due to the large number of comparisons made for each model, the Bonferroni correction was used with a critical cutoff of p=.003 (05/16).
