The participants completed questionnaires in their classrooms in 1990 and individual interviews in 1995. Logistic regression analysis estimated increases in the risk of behaviors or attitudes in late adolescence associated with more frequent marijuana use in early adolescence. The participants had an average age of 19 years in 1995. Early adolescent marijuana use increased the risk in late adolescence of not graduating from high school, juvenile delinquency, having multiple sexual partners, and not always using condoms. Early adolescent marijuana use also increased the risk of perceiving drugs as not harmful; having problems with cigarettes, alcohol, and marijuana; and having more friends who exhibited deviant behavior. These relationships existed with controls for age, sex, ethnicity, and, when available, earlier psychosocial measures. Findings indicated that marijuana use in early adolescence relates to later adolescent problems that limit the acquisition of skills necessary for employment and that heighten the risks of contracting HIV and abusing legal and illegal drugs. Hence, clinical practice needs to incorporate assessments of and treatment for adolescent marijuana use. Tables and 47 references