As aptly noted by other researchers recently examining developmental trajectory groups of personality (Johnson et al., 2007), mixture analysis is a methodology to allow for potentially useful categorizations of heterogeneous patterns of development rather than a method to “carve nature at its joints.” We explicitly reject common misconceptions of trajectory groups, such that individuals “belong” to a trajectory group, that the number of trajectory groups is absolute, and trajectories of group members follow the group-level trajectory in lock step fashion (see Nagin and Tremblay, 2005; for more details). Rather, as with Nagin and Tremblay (2005), we view trajectory groups as a useful statistical device to make appropriate comparisons on important outcomes among meaningful classifications of individuals. Despite the limitations of mixture modeling, the impulsivity trajectory solution estimated and discussed in this article has numerous strengths, including a rigorous and empirically driven model selection procedure, the “free” estimation of developmental patterns across time (i.e., changes in impulsivity were not forced to occur in “straight lines” as in typical linear growth mixture modeling), and empirical tests of theoretically meaningful associations between impulsivity trajectory groups and several indices of alcohol involvement.
