We tested whether life events interacted with GABRA2 genotype to predict cross-lagged CBCL externalizing while simultaneously controlling for parental externalizing. The results are shown as a series of models in Tables 3 and 4. Model 1 includes the main effects of all variables. Model 2 includes the main effects and the interaction of life events × GABRA2. Model 3 includes the main effects, the life events × GABRA2 interaction effect, and additional statistical controls for all covariate × genotype and covariate × environment interactions. In terms of main effects, CBCL externalizing decreased over time. Adolescents who were missing a parental antisocial behavior report, or whose parents had higher parental antisocial behavior and alcohol problems had higher CBCL externalizing (see Model 1 in Tables 3 and 4). Sex, missing a parental alcohol problem report, life events, and GABRA2 genotype were not significantly associated with CBCL externalizing.
