In an experiment in non-formal schools in Indian slums, an incentive for attending a target number of school days increased average attendance when the incentive was in place, but had heterogeneous effects after it was removed. Among students with high baseline attendance, the post-incentive...