Abstract

We use administrative data from the lottery-based open enrollment system in Beijing middle schools to obtain unbiased estimates of school fixed effects on student performance. To do this, we classify children in selection channels, with each channel representing a unique succession of lotteries through which a child was assigned to a school. Results show that school fixed effects are strong determinants of student performance. These fixed effects are shown to be highly correlated with teacher qualifications measured in particular by their official ranks. Teacher qualifications have about the same predictive power for student test scores as do school fixed effects.

pdf

Share