Moving targets: Modeling developmental trajectories of adolescent alcohol misuse, individual and peer risk factors, and intervention effects

From a developmental perspective, a central goal of prevention is to alter behaviors that are already changing—that is, to redirect potentially risky trajectories. This multiwave panel study examines the longer-term effects of a prevention program on trajectories of alcohol misuse and related risk factors (susceptibility to peer pressure to misbehave and exposure to peer alcohol use). We conceptually and empirically compare Hierarchical Linear Modeling (HLM) and Latent Growth Curve Modeling (LCM), working within the tradition of each technique and noting similarities and differences in approaches, hypothesis operationalizations, and findings. Five waves of data (Grades 6 to 10, n = 675) from the Alcohol Misuse Prevention Study, a randomized treatment-control group study designed to examine the effects of a school-based prevention program that focused on increasing social (especially refusal) skills, were analyzed. HLM and LCM analyses showed susceptibility, exposure, and alcohol misuse increased mostly linearly across adolescence and covaried positively within and across time, arguing that they exist in a mutually reinforcing web of influence. Both techniques revealed small treatment effects, with the prevention reducing the normative increase in alcohol misuse during adolescence. There were some minor, but important, differences between HLM and LCM in conclusions about intervention effects. Substantive and developmental methodological implications are discussed.