Data-driven Decision-making in School Matching
Abstract
School choice mechanisms determine how students are assigned to public schools, striving to balance individual preferences, school priorities, and broader goals of fairness. While simple, well-studied algorithms have strong theoretical guarantees, real-world applications face practical constraints and policy goals that complicate implementation. We focus on three challenges in modern school choice systems: determining the optimal length of student preference lists, designing the allocation of reserved seats, and targeting scholarship allocations to support disadvantaged students. For these problems, we explore how data can be used to come up with relevant models, and how solutions to the latter lead to policy recommendations. Based on works with Swati Gupta (MIT), Aapeli Vuorinen (Columbia) and Xuan Zhang (Meta).