Many questions about the associations between risk factors and drug use remain unanswered. Data originally obtained from seventh graders participating in a school-based prevention program were examined in terms of a theoretical model. A series of multiple regression analyses were then conducted to select those risk factors which contributed unique variances in predicting individually five types of drug use: alcohol, cigarettes, marijuana, cocaine, and a composite of hard drug use. A final measurement model indicated the psychometric soundness of the relationship among the measured and latent constructs. Conceptually, a risk factor methodology supports the notion of a multiple pathway model for predicting adolescent drug use. The concept of multiple pathway model suggests a need for different prevention strategies tailored for specific high-risk groups. Educational intervention strategies that target multiple risk factors, aimed at specific high risk groups, can help to inoculate individuals from experiencing the deleterious effects associated with a drug using/abusing lifestyle. Differentiating patterns of risk for drug use is critical toward the development and implementation of sound primary prevention strategies. Five tables and two figures are included. (ABL)