Relying on historical research, a longitudinal data set, and multivariate analyses, the drug-violence relationship is scrutinized. A proposed model is tested and supported, indicating that attitudes toward violence, gender, neighborhood problems, minor delinquency, and victimization were persistent correlates that must be considered within the drug-violence relationship. Parental attachment and importance, exposure to delinquent peers, and drug dealing were also important. Both licit and illicit drug use were significant within the models, although the relationship changed from year to year. In year 1, youth who used drugs reported more violence. In year 2, youth who were not using drugs reported more violence. Association with delinquent peers and initial involvement in drug dealing were likely explanations for this transition. Findings offer support for prevention efforts that disrupt drug markets and target male youth who are involved in crime and drugs, repeatedly victimized, associating with delinquent peers, and developing attitudes favorable toward the use of violence.