We evaluate the long-term effect of a “girl-friendly” primary school program in Burkina Faso, using a regression discontinuity design. The intervention consisted of upgrading existing three-classroom schools to six-classroom schools to accommodate more grades. After six years, the program...