Heterogeneous Beliefs, School Choice Systems and Mechanism Design Solutions

Mechanism Design in Practice in New Haven, CT

New Haven School Choice Project

Continuing on a long path of collaboration with the New Haven Public Schools (NHPS), this project aims to better understand how parents and students make choices about which schools to attend. How their choice processes interact with the rules that districts use to assignt students to schools, and how these interactions affect family welfare and academic achievement.

The implementation was enacted over a period of almost 4 years, and incorporated two household surveys in which families and students were surveyed. The survey included questions regarding beliefs about admissions, preferences about schools, and educational plans for both the family and the student.

This project makes three technical contributions. The first is to integrate survey data on students' elicited preferences over schools and subjective expectations about the school choice process into an estimatable empirical model of school choice in the context of a non-strategy proof school choice mechanism. The survey data allow to separate students' beliefs about admissions probabilities from their preferences over schools under relatively weak assumptions about student behavior. Second, comparing counterfactual predictions of behavior based on model estimates to those observed under randomized informational interventions or policy changes will provide a strong test of the accuracy of model predictions. Using these tests of accuracy, additional simulations will provide valuable evidence on the effects of future interventions in other contexts. Third, using administrative data on academic achievement in addition to school assignment, this project will link the choices districts make in school choice mechanism design to outcomes of direct policy interest such as test scores and high school graduation. Project website Academic Paper Online Appendix