The social development model describes "social development" as a "process in which the most important units of socialization -- families, schools, and peers -- influence behavior sequentially, both directly and indirectly" (Hawkins and Weis, 1985, p 79). As an extension of classic social development theory, the study also examined a concurrent effects model, which posits that ethnic identification will significantly impact the relative effects of parent-child relations, peer relations, school bonding, and adolescent drug use. This model posits that developmental stage and ethnic group will significantly impact the association between psychosocial factors and drug use, such that discontinuities may be accounted for by differences in culturally normative patterns. For the current study, a non-probability sample of 446 participants was recruited primarily from Boy's Club recreation centers in Ventura County, CA. At least one parent or legal guardian of each child was also recruited as a collateral informant. A total of 392 parent-child pairs completed both initial and 12-month follow-up surveys. Interviews included questions about participants' interactions with parents and peers, academic and extracurricular activities, and specific antisocial behavior, including drug use. Three sets of analyses focused on the effects of age, ethnicity, and gender on subjects' reported patterns of drug use over time. The study then examined the consistency between patterns of adolescent, parental, and peer drug use and psychosocial influences at initial interview as a function of age within ethnic groups. Structural equation models found that patterns and predictors of drug use were distinct for each subgroup by age and ethnicity. Later drug use was primarily associated with prior use, exposure to parental and peer models of drug use, and family disruption. These findings are partially consistent with social development theory, in that strong links to conventional social institutions and behavior apparently protected children and adolescents from significant investment in drug-using behaviors; however, these effects were not found to be necessarily influenced by association with peers or peer groups. 2 tables, 4 figures, and 58 references