The authors study the impacts of two alternative merit-based scholarship designs for 5th to 8th graders through a field experiment in Malawi. For those in the Standard scholarship program, top performers on a final exam were awarded scholarships, regardless of baseline test score. For those in the Relative scholarship program, students were grouped by similar baseline score, and scholarships were awarded to the top performers in each group. In addition, feedback on students’ midline exam scores was randomly provided to a subset of students. It was found that the Standard program significantly decreased test scores and motivation to study, especially for those least likely to win the scholarship. By contrast, no evidence is found for test score impacts among those in the Relative scholarship program. The authors also find that feedback on ranking may improve test scores for high performers, particularly among those in the standard scholarship program.