A study of 1,093 adolescents from six different public high schools showed that the same theoretical model of control and social learning theories fit the data on alcohol use, heavy alcohol consumption, and use of marijuana. The same model did not explain the use of amphetamines, barbiturates, cocaine or opiates. The various forms of drug use studied are different expressions of problem behavior and are the result of independent processes. One possible explanation of the results is that alcohol and even marijuana use are more acceptable among most adolescents than are the other forms of drug use and, therefore, different influences are relevant. The study also found that religiosity was a significant influence only on certain forms of drug use, whereas the feedback effect of drug use on religiosity was significant for all forms studied. It appears that youth reduce their commitment to religion to reduce the cognitive dissonance arising from the discrepancy between internalized moral proscriptions and drug-using behavior. Figure, table, notes, references