In the United States, there exist both overt and covert barriers limiting some students' access to high-quality schools, creating systematic inequities that affect labor market outcomes. Insurmountable challenges in dismantling these barriers have led policymakers to explore the capacity for school choice reforms as a means to improve student outcomes. The decentralized nature of public education in the United States has allowed for wide experimentation across states and districts, but the heterogeneous implementation of policies has also generated mixed results and many open questions. This dissertation contains three essays that aim to tackle some of these open questions.
In chapter 2, I outline a conceptual framework that provides some structure for subsequent chapters. The purpose of this chapter is present a benchmark model where the typical benefits that motivate contemporary school choice reforms materialize. The contributions of this chapter are to generate an empirically-oriented model at the cost of several simplifying assumptions. Other models in the literature take an opposite approach, but the primary contributions of this dissertation are empirical so I opt for an empirically-oriented theoretical framework. I link this model to the subsequent chapters in two ways. First, a key statistic that summarizes families' expected welfare gains from the program is an important covariate for the empirical analysis in Chapter \ref{sec:empiricalchapter}. Second, some of the simplifying assumptions regarding perfect information and preferences receive cursory treatment in Chapter \ref{sec:empiricalchapter} and are the focal point of a field experiment discussed in Chapter \ref{sec:experimentchapter}.
In chapter 3, co-authored work with Caitlin Kearns, I study how the demand and supply for school quality can evolve in response to a policy---the Zones of Choice (ZOC) program--that changed the underlying choice and competitive structure that families and schools face. The contributions of this chapter are to provide empirical evidence tackling the questions surrounding access, competition and student-school match quality. I find that the achievement among students enrolled in ZOC schools increases, with gains sufficiently large to eliminate eleventh-grade achievement gaps that existed in the years before the policy expansion. I distinguish between effects driven by parents choosing schools that best suit their children's needs and competitive effects induced by the policy, and find most of the student gains are driven by general improvements in school quality as measured by school value-added. The lack of improvements in student-school match quality are in part due to the homogeneous set of students within each zone, eliminating the scope for improvements in student-school match quality based on observable student characteristics. I find evidence that schools exposed to more competition (as captured by a statistic derived in Chapter \ref{sec:modelchapter}) improved more, providing suggestive evidence for a competitive effects story. I then provide suggestive evidence that parent's selected schools in a way that incentivized schools to compete on quality, a finding that contrasts a growing body of evidence suggesting otherwise. The highly segregated nature of each zone of choice can maybe explain differences with the past literature. In settings where parents can use race or income to proxy for school quality, they will do so, but in segregated neighborhoods parents will use other proxies that are correlated with school effectiveness. This presents a tradeoff present in the ZOC setting: short-run gains in terms of reductions in within-district inequality in exchange for potentially negative long-run effects through the entrenchment of school segregation in the district.
Information and preferences are the focal point of Chapter 4, where I report the findings from a field experiment I conducted in Fall 2019. My contribution is to study parental preferences through an information provision experiment, linking the information provision literature with another growing body of research studying parents' preferences for schools. By distributing information to students' households I address the information channel, and in addition provide settings where some parents are perfectly informed about both school and peer quality, allowing me to study relative preferences for peer and school quality. In addition, I explore the role that social interactions play in determining schooling decisions. I corroborate my findings from Chapter \ref{sec:empiricalchapter} by showing that parents not receiving any information tend to place higher weight on school quality than peer quality. Among parents receiving information, I first document evidence that their most-preferred schools changed in terms of quality and peer composition. I then provide evidence showing that the information campaign led to increases in relative preferences for school value-added, suggesting information campaigns can produce changes in choices that affect school quality. Through various avenues, I provide evidence of large and salient treatment spillovers, indicating substantial interactions between parents. The prevalent role of social interactions could introduce disadvantages to groups with less informed networks in this and other settings.
