To test Hypothesis 3, regarding within-individual changes in the associations between positive and negative reinforcement and alcohol consumption as a function of AD, we fit a multilevel growth model with alcohol consumption as the outcome variable. In this model, we disentangled between-person and within-person variations following the approaches described by Curran and Bauer (2011). Specifically, we created person-level means of positive and negative reinforcement (to represent between-person variance) and included them as timeinvariant, between-person covariates. We created person-mean centered variables for positive and negative reinforcement (to reflect within-person variance) and included them as time-varying, within-person predictors, along with AD, and the 2-way interactions between them. Because AD was coded binary, parameters for the 2-way interaction effects represent changes in the associations between positive/negative reinforcement and alcohol consumption as a function of the presence of AD. As in the previous models, linear and quadratic effects of age were included as time-varying covariates, and sex and self-reported race were included as timeinvariant covariates. Random effects of the intercept, linear and quadratic slopes of age, and all time-varying predictors were examined, with non-significant
