Research on drug use has identified an additional factor that should be considered in analyzing differential reinforcement of behavior learned in social interactions (social learning theory), i.e., drug use as a method for solving a problem. The four hypotheses tested in this study pertain to patterns of drug use and the users perception regarding whether the drug use is solving a problem. The data used in the study were obtained from the first three waves of the National Youth Survey, a longitudinal survey of delinquency and drug use in a sample of households in the United States in 1976. For the current study, 7,998 households were selected, resulting in a sample of 1,725 youth ages 11 to 17. The participants were interviewed in 1-year intervals from 1977 to 1982, with one parent per household interviewed in the first year. The sample was judged to be reasonably representative of all youths ages 11 to 17. The key independent variables measured social learning constructs. Reinforcement was a key concept in the analysis. The two dependent variables were continuation of drug use and frequency of drug use. Chi-square analysis found that social learning variables, including problem solving function and effectiveness, distinguished between those who continued drug use and those who did not. Among the users, problem solving as a measure of reinforcement was a significant predictor of high-frequency drug users. Generally, the findings suggest that adolescents tended to view drug use as a way of solving perceived problems. Implications of these findings for drug policy, theory, and research are discussed. 7 tables and 57 references