3b) throughout the observed period.9 In the present study, we decided to focus on those who improved versus those who did not, primarily to increase power and improve accuracy of parameter estimates in detecting predictors and moderators in subsequent logistic regression analyses. For instance, cross-tabulating incident seriousness with the four classes resulted in a few cells with a limited number of observations, especially for the smallest class (class 1). Previous studies in the literature have combined empirically identified groups into a smaller number of groups based on other practical and conceptual considerations (e.g., Bongers, Koot, van der Ende, & Verhulst, 2004), or rejected alternate solutions from consideration based on additional criteria (e.g., minimum cluster size [5% or more], distinctively shaped trajectories, or large bivariate residuals). Therefore, in all subsequent analyses, classes 1 through 3 were combined into a single non-improved group (n = 162, 46.6%). The remaining group was labeled as an improved group (n = 186, 53.4%).10
