Models examining the impact of AO, delay, and the covariates on the intercepts and slopes for heavy drinking and problems were structured similarly. Covariates (gender, ethnicity [4 dummy codes contrasting ethnic minority groups to Caucasians], impulsivity, and family history) were treated as exogenous predictors of AO, delay, and the intercepts and slopes for heavy drinking or problems. A directional path was included from AO to delay, permitting an examination of indirect effects of AO on drinking outcomes, operating through delay. Directional paths from AO and delay to the intercepts and slopes for heavy drinking and problems assessed direct effects of each. Finally, the AO by delay interaction was added to each model as a predictor of intercepts and slopes. Model-fit was evaluated using root mean square error of approximation (RMSEA) <0.08 (Browne and Cudek, 1993; MacCallum et al., 1996), comparative fit (CFI), and Tucker-Lewis (TLI) indices >0.90 (Bentler, 1990).
