This component of the Treatment Research Center's competitive renewal is designed to both use and develop Latent Class Growth Models (LCGM) for analyzing the change over time in complex patients' behavior. The overall goal of this component is to apply this methodology to better characterize complex patients' drug using behaviors and to better understand the underlying factors that account for both the similarities and heterogeneity in those patterns. This work is designed to help answer questions such as; Are study participants who relapse once different from those who relapse two or three times? What baseline factors influence which latent trajectory a smoker will most likely take? Does mood or prior alcohol use influence the path or shape of the trajectory taken? The work applies to multiple Center components which focus on nicotine but is not limited to one drug. Thus, it is aimed at helping us compare and contrast trajectories of behavior across treatments and drugs of abuse. The work is organized around meeting four aims; (1) To apply LCGM to previously collected data to compliment and extend the findings of those studies and to examine the applicability, interpretability, and usefulness of LCGM in drug abuse research. (2) To use computer simulations to study the effects of various aspects of sample size on the ability of LCGM to capture the latent trajectories using studies based on computer simulation. (3) To uncover the latent trajectories of change in smoking and alcohol consumption in the two proposed TRC clinical trials of smoking cessation. (4) To determine optimal methods of interpretation and data display.