It has been observed with other U.S. samples that the slope of the increase in frequency of alcohol use rises in the first year of high school, and thus there is discontinuity from middle to high school in growth rate. One strategy to address this issue has been to examine piecewise models of growth across these two developmental periods (Brown et al., 2005; Li et al., 2001). In the current study, because of model size and complexity, the issue was addressed by estimating models separately for middle school (Grades 6–8) and high school (Grades 9–12). To test changing influences over time (e.g., increases in peer alcohol use) more systematically, time-varying influences (i.e., change scores of the predictors from the prior time period) on both indicators of alcohol use were examined, in addition to prospective predictors from the initial years of middle and high school (Grades 6 and 9, respectively).
