We provide experimental evidence on the relationships between education, HIV/AIDS education, risky behavior and early fertility in Kenya. We exploit randomly assigned variation in the cost of schooling and in exposure to the national HIV/AIDS prevention curriculum for a cohort of over 19,000 adolescents in Western Kenya, originally aged 13.5 on average. We collected data on the schooling, marriage, and fertility out-comes of these students over 7 years, and tested them for HIV and Herpes (HSV2) after 7 years. We find that subsidizing education at the upper primary level reduces the dropout rate by about 18 percent. For girls, the education subsidy also leads to a significant reduction in teen pregnancy and teen marriage, but does not reduce the risk of sexually transmitted infection (STI). In contrast, bundling the education subsidy with the delivery of the HIV curriculum (with its abstinence-until-marriage message) leads to a lower STI risk for girls, but to a smaller decrease in early fertility than the subsidy alone. Finally, the HIV curriculum by itself has no impact on STI rates or on early pregnancy, although it reduces the number of unwed teenage pregnancies. These results are consistent with a model of sexual behavior and schooling decisions where girls choose whether to have casual or committed relationships, and teenage pregnancy may b e a comparably desirable outcome for girls who cannot continue their education.