We evaluate the causal effects of a program that constructed high quality "girl-friendly" primary schools in Burkina Faso, using a regression discontinuity design 2.5 years after the program started. We find that the program increased enrollment of all children between the ages of 5 and 12 by 20...