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

Whether early school stratification is conducive or detrimental to scholastic performance has been subject to controversial debates in educational policy and science across many countries. We exploit a reform in Lower Saxony, Germany, where performance based tracking was preponed from grade 7 to grade 5 in 2004, i.e. with the completion of primary school. In particular, we measure the long-run effects of this reform on PISA achievement test scores based on a difference-in-differences setup. In order to disentangle average from distributional achievement effects, we complementarily rely on a changes-in-changes framework. Our results indicate that preponed school tracking increases test scores in the upper tail of the skill distribution and lowers test scores in the lower tail of the skill distribution, compensating each other on average.