Admission information was obtained on youths who entered the Youth Advocacy Program (YAP), a drug-abuse prevention program in southeast Austin, Tex., between March 1981 and December 1985. Information was obtained on variables related to sociodemography, family, school, legal status, and drug use. Follow-up information contained many of the same variables. The results provided mixed support for the hypotheses studied. With respect to the hypothesis that the quality of the adolescent's relationship with parents had an influence on drug use, only limited support was found. Correlations of the family relations variables with the outcome criteria were not significant, but parental control at admission was a significant predictor in the multiple regression analyses of the Crime Index. The influence of peers on outcome criteria was found to be the most consistent. Factors such as peer deviancy, peer drug problems, and peer activity were shown to have significant relationships with deviant behaviors. There was partial support for the hypothesis that involved attitudes toward school, in that females reported a negative relationship between satisfaction with school and drug-use severity. There was also limited support for the hypothesis that intrapersonal influences and psychological adjustment, such as self-esteem or satisfaction with one's environment, would be related to subsequent drug use. In this study, self-esteem did not have a significant relationship with any of the drug-use variables but was significantly related to alcohol use. The results of regression analyses suggested that by using the family, friends, and self scales, a statistically significant amount of variance could be accounted for in predicting alcohol problems, use of other drugs, and criminality. The most significant predictors in the regression models were the peer-related scales. 1 table and 16 references