There was an interaction between alc-PRS and friend support in predicting alcohol use (Model #). As illustrated in Figure 1, simple slope analysis indicated that alc-PRS was more strongly associated with alcohol use when friend support was low (−1 SD; B = .17, SE = .03, β = .14, p < .001) than when friend support was high (+1 SD; B = .07, SE = .03, β = .06, p = .011). The interaction effect between alc-PRS and friend support in relation to alcohol use remained statistically significant in follow up analysis where alc-PRS by covariate (age, sex, educational attainment, household income) and friend support by covariate interaction terms were included in the regression model (p = .013, see Supplemental Table 1), suggesting robustness of the interaction effect. There was no significant interaction between alc-PRS and family support in relation to alcohol use.
