Dutch children with similar IQ are twice as likely to attain higher education when their parents are highly educated compared to when parents are lower educated. The magnitude of such unequal opportunities is thought to depend on characteristics of the school environment, such as school quality and the way children are tracked into different educational levels. However, testing these claims is problematic with conventional sociological methods, because it is difficult to measure to what extent educational outcomes depend on “family background” rather than “ability”. This project takes a different approach and applies twin methods to administrative data available through Statistics Netherlands (CBS) to investigate the interplay between genes, family background, and school characteristics. Twin methods allow to disentangle genetic from environmental influences and investigate the interaction between them. The CBS administrative data have notable advantages over twin samples because of their size, representativeness and coverage. This design will therefore greatly improve our understanding of the causal mechanisms underlying educational inequality. We assess the potential effectiveness of proposed Dutch educational reforms by measuring the school characteristics targeted by these reforms (i.e., school quality, standardization of tracking). Results will directly feed into policymakers because of a close collaboration with the Inspectorate of Education.