Study participants were selected from cities in Texas, Ohio, Oregon, and Pennsylvania. The outpatient sample was 70 percent male, while the inpatient sample was 52 percent male. All participants were assessed using the Adolescent Drug Abuse Diagnosis instrument at admission and again at follow-up evaluation. The outcome criterion used was the degree of reduction in substance abuse. Of 33 predictor variables analyzed, 12 predicted inpatient outcome and 15 predicted outpatient outcome. Only two variables, school and not considering substance abuse to be harmful to health, predicted worse outcome for both treatment settings. Several characteristics found prognostic for adolescents were different than those reported earlier for adult substance abusers. For example, greater substance abuse at follow-up had been reported to be predicted by greater severity of alcohol and drug use at admission, for both inpatient and outpatient adults, but this was true only for outpatient adolescents in the current study. 18 references and 3 tables