Despite an intense health promotion campaign in past decades, many youths living in high-poverty neighborhoods have been seemingly immune to interventions designed to prevent or change unhealthy behaviors. One explanation for this failure is that the decision making styles of these youths, influenced by their developmental status, interact with the social and environmental factors associated with poverty to make them susceptible to risk behaviors. We explore this explanation using data collected from adolescents living in impoverished inner-city neighborhoods. The basic model guiding our research considers three multidimensional constructs as determinants of substance use among low-income adolescents: (a) developmental status (e.g., mental age; self concept;); (b) environment (e.g., peer and family relations; neighborhood problems); and (c) decision making processes, including both systematic cognitive processes (e.g., generative thinking) and heuristic processes (i.e., automated, unreflective processing). This model is tested using a longitudinal design in which we measure these constructs at three annual time points. Data are collected from 500 10- to 14-year old adolescents living in high-poverty neighborhoods in Mobile, AL; the longitudinal design allows us to examine developmental trajectories in the measured constructs and to infer causality. In addition to the data collected specifically for the proposed study, we will use data from the Mobile Youth Survey (MYS), a multiple cohort longitudinal study of adolescent risk behavior (based on approximately 300 questions) which we have conducted since 1998 and which will continue for at least five more years. All adolescents recruited for the proposed study will have at least one prior wave of MYS data, and they will participate in the MYS throughout the course of the proposed study.