A potentially helpful method to both characterize individual differences in personality-trait change as well as examine covariates of individual differences in change is longitudinal mixture modeling. Briefly, this framework allows for classes (or clusters) of individuals who exhibit similar patterns of personality stability and change to be identified and then compared on both time-invariant and time-varying covariates. For example, similar approaches have been used to distinguish individuals characterized by “adolescence-limited” endorsements of antisocial behavior from individuals that exhibit more chronic “life-course-persistent” patterns of externalizing behaviors (Mofitt, 1993). Similarly, longitudinal mixture modeling involving impulsivity may distinguish individuals that show relatively high but developmentally limited levels of impulsivity (i.e., individuals that undergo substantial decreases in impulsivity across time) from individuals characterized by relatively high and relatively stable levels of impulsivity. These distinct groups of individuals could then be compared on a number of outcomes, including differences in developmental changes in alcohol involvement. As recently discussed by researchers that have utilized similar methodology to study personality development (Johnson et al., 2007), the purpose of this statistical methodology is not to “carve nature at
