Preferences and Heterogeneous Treatment Effects in a Public School Choice Lottery

This paper combines a model of parental school choice with randomized school lotteries in order to understand the effects of being assigned to a first-choice school on academic outcomes. We outline a simple framework in which those who place the highest weight on academics when choosing a school benefit the most academically when admitted. Although the average student does not improve academically when winning a school lottery, this average impact conceals a range of impacts for identifiable subgroups of students. Children of parents whose choices revealed a strong preference for academic quality experienced significant gains in test scores as a result of attending their chosen school, while children whose parents weighted academic characteristics less heavily experienced academic losses. This differential effect is largest for children of parents who forfeit the most in terms of utility gains from proximity and racial match to choose a school with stronger academics. Depending on one's own race and neighborhood, a preference for academic quality can either conflict with or be reinforced by other objectives, such as a desire for proximity and same-race peers.