We report results from a randomized evaluation of a merit scholarship program in Kenya in which girls who scored well on academic exams at the end of 6th grade had their school fees paid and received a cash grant for school supplies over the next two years. In the sample as a whole, girls eligible for the scholarship showed substantial gains in academic exam scores, and teacher attendance also improved significantly in program schools. There was also evidence of positive externalities: girls with low pre-test scores, who were unlikely to win scholarships, showed test score gains in program schools. We cannot reject the hypothesis that test score gains were the same for girls with low and high pre-test scores. We see no evidence for weakened intrinsic motivation or gaming, and effects persist after incentives were removed. There is also evidence of heterogeneity in program effects, suggesting the impact of incentives is context dependent. In one of the two study districts, test score effects were large, there were positive spillovers to boys, and student attendance increased in program schools. In the other district, attrition complicates estimation, but we cannot reject the hypothesis that there was no program effect.