Post-Pandemic (Dis)enrollment in New York City’s Choice-based School System

• In Event: Public School Student Disenrollment: Causes, Effects, and Potential Solutions

Abstract

The COVID-19 pandemic shifted school enrollment patterns and, in many contexts, exacerbated existing trends of disenrollment from traditional public schools. In most districts across the country, enrollment in traditional public schools has been in decline since the turn of the century—partly due to increased enrollment in alternative schooling options, such as homeschooling or charter and private schools. However, the literature that has examined these patterns has largely treated them descriptively, offering limited insight into the mechanisms driving families’ decisions. 

In this paper, we address this gap by leveraging longitudinal, student-level administrative enrollment data from New York City (NYC) between 2016 and 2025 to shed light on the mechanisms underlying enrollment decisions. NYC represents an ideal setting for this analysis: as the largest school choice system in the country, families face a broad and structured set of schooling options, making enrollment decisions more directly interpretable as constrained choice behavior rather than purely residential sorting. In addition, NYC also offers specialized schooling services to students with disabilities in District 75 (D75) schools, which enable us to examine whether access to differentiated supports shapes patterns of disenrollment. 

Using longitudinal, student-level administrative data linked to school characteristics and neighborhood contexts, we examine patterns of exit across public, charter, and homeschool sectors. Our preliminary analyses show that during this period, traditional public school enrollment declined while charter school, D75 school, and homeschool participation increased. Ongoing analyses incorporate multiple dimensions of schooling, including discipline incidents, academic performance, attendance, school climate, and teacher turnover, alongside student demographic and socioeconomic characteristics to investigate the mechanisms underlying such shifts.

Our empirical approach follows a funnel structure, moving from broad school-level patterns to student-level heterogeneity and cross-level interactions. First, we engage in school-level analyses to identify the characteristics of schools that experienced the greatest enrollment losses over time. Second, we fit student-level fixed-effects models to (a) estimate how prior school conditions predict subsequent enrollment transitions, with particular attention to differences between pre- and post-pandemic periods, and between transition and non-transition grades, and (b) examine interactions between school characteristics and student attributes to capture heterogeneous disenrollment patterns, introducing one moderator at a time. 

In this study, we test several potential mechanisms, including increased behavioral challenges, changes in academic quality, and differential access to alternative schooling options across socioeconomic groups. For example, prior literature (Francis & Goodman, 2025) suggests that disenrollment occurs due to negative peer behavior or academic performance. We can directly test this hypothesis by investigating whether schools with negative behavior climate as measured by infraction records and climate surveys experience the greatest attrition post-pandemic. By identifying which factors most strongly predict disenrollment, this research contributes to the growing evidence on how disruptions to schooling environments translate into system exit. The findings have direct implications for policymakers seeking to address enrollment declines, improve school conditions, and reduce inequities in access to educational options in large urban systems.

Authors

• Jo R. King, Boston University
  Non-Presenting Co-Author

• Berta Bartoli, New York University
  Presenting Author

• Jerray Chang, Boston University
  Non-Presenting Co-Author