This paper employs newly developed quantile regression techniques to investigate a policy that could differentially affect students' performance. The Colombian vouchers were assigned using lotteries, and were renewable as long as the students maintained satisfactory academic progress. This...