Adolescent alcohol use and adult alcohol disorders: a two-part random-effects model with diagnostic outcomes.

Alcohol use is often analyzed by treating the behavior as a single dimension, such as focusing on frequency of use. Based on data from a longitudinal study, this report considers two distinct aspects of semi-continuous alcohol use data. A two-part random-effects model was used to evaluate change in the log-odds and frequency of use from about age 13 to about age 18 years. Change features were then related to the log-odds of later alcohol disorders. Results suggested differences in the two aspects of use over time and their relationships with later disorders. Most important for the purposes of this study, different methods of analyzing antecedents and consequences of alcohol use trajectories were shown to generate both similar and disparate findings.