By way of model building, our analyses followed three sequential steps. First, growth mixture modeling (GMM; Jung & Wickrama, 2007; Muthén & Shedden, 1999) was used to identify homogeneous subgroups of individuals manifesting distinct patterns of change in their externalizing problems from early adolescence through young adulthood (i.e., from age 12 to 22 years). Conventional growth curve modeling assumes a mean pattern of change in behavior within the population, with individual differences expressed in terms of normal variability around specified growth parameters (i.e., intercept and slope coefficients that define the level and shape of the change; see McArdle & Epstein, 1987). GMM is a widely used extension of this procedure that allows for the possibility of two or more discrete subgroups of individuals within a population, each having unique mean trajectories, as well as individual variability around the mean intercepts, linear slopes, and rates of acceleration (Muthén & Muthén, 2000). Because the identification and explanation of within-group differences was not a primary focus of the present analyses, the variances for all like growth parameters were constrained to be equal across
