The main objective of the proposed project is to test a developmental model of older adolescent alcohol use and abuse that extends current psychosocial interaction models of usage. The model incorporates adolescent individuation, a key aspect of development, in the context of individual, family, and peer factors, perceived stress, and externalizing behaviors to further understand alcohol use during later adolescence and with heavy drinkers. The proposed model hypothesizes causal sequences between variables that effect individuation and subsequently usage, both directly and indirectly. The model will be explored with three ethnic groups, African American, Mexican American and non-Hispanic White adolescents, with the expectation that ethnic differences in individuation and family dynamics will have a unique bearing on the role of individuation in substance use. A sample of 3500 adolescents, starting with 9th graders, will be surveyed 7 times over 4 years concerning substance use and these psychosocial factors. In addition, assessment of school grades and behavior will be collected from schools. The proposed project will examine five specific aims. First, the investigators propose to examine individuation in the context of family factors, such as conflict, cohesion, communication and parenting; individual factors, such as impulsivity and affectivity; stress; and externalizing behaviors that are influential for alcohol use and abuse. Second, they will examine the influence of peer relations on alcohol use in the context of individual, family, stress, and developmental factors. Third, they will investigate the moderating role of individuation on the relationships between family factors, individual factors, stress, and peer factors and alcohol use. Fourth, they propose that over time there are reciprocal and bi-directional effects of individual family, peer, stress, externalizing behaviors, and individuation factors on subsequent drinking, and that drinking then impacts these psychosocial factors. Fifth, it is expected that the influence of developmental, individual, family, and peer factors will vary based on ethnicity and gender of the adolescents. Multivariate longitudinal methods including structural equation modeling and hierarchical linear modeling will be used to examine these aims.
