The distributional effects of early school stratification: Non-parametric evidence from Germany

The effects of early school stratification on scholastic performance have been subject to controversial debates in educational policy and science. We exploit a unique variation in Lower Saxony, Germany, where performance based tracking was preponed from grade 7 to grade 5 in 2004. We measure the long-run effects of early school stratification on in- dividual PISA test scores along the entire skill distriubution using the changes-in-changes estimator. Our results indicate that preponed school tracking increased test scores at the upper tail and lowered test scores at the lower tail of the skill distribution, compensating each other on average.