Shown in Table 4 is the model predicting alcohol use and volume across the 4 years of high school (Grades 9–12). The average level of alcohol volume across middle school was included as a control variable and was a significant predictor of the Grade 9 intercept. Findings for prediction to use versus nonuse (Part 1) indicated that the high school use intercept was predicted by the youths’ antisocial behavior and by peer alcohol use. The use slope in high school was negatively predicted by both the use intercept and by peer alcohol use at Grade 9, possibly indicating that because of the relatively high initial use levels associated with these predictors, there was less than average growth in likelihood of use across the period. Prediction from the change scores indicated that increases across time in peer use were associated with a greater likelihood of use versus nonuse.
