Mapping Partisanship in U.S. School Districts Over Time

• In Event: Politics, Governance, and Policymaking in Education

Abstract

For much of U.S. History, American education was governed primarily at the local level through institutions deliberately insulated from partisan politics (Tyack 1974; Henig 2013). Progressive Era reforms produced a system of locally elected, nonpartisan school boards intended to protect education from the hostility of partisan electoral competition and national party conflict. This form of institutional “exceptionalism” distinguished education from most other domestic policy domains (Henig 2013). However, in recent years this exceptionalism has begun to erode (Henig 2013; Manna 2006; Manna 2011), and education has become increasingly nationalized and polarized, with partisan identities now shaping public attitudes toward education policy (Houston 2024), national political dynamics shaping local school board elections (Henig et al. 2019), and even the erosion of public support for local control of schools itself (Arnzen & Houston 2023). As a result, local school districts are now sites of unprecedented political conflict, shaped by partisan politics even in formally nonpartisan contexts (Henig et al. 2019; Shah et al. 2024; Gensterblum et al. 2026; Arnzen & Jacobsen 2025).

Despite these changes, scholars of education politics and local politics face a fundamental empirical limitation: there is no systematic, longitudinal measure of partisanship at the school district level. This challenge has plagued national analyses of American education politics because school district boundaries rarely align with conventional political geographies such as counties or municipalities. And further, school district elections often do not align with municipal or state elections chronologically, preventing linking of many election records and data. At the local, school district level, these issues have made standard electoral measures unusable in many analyses. As a result, partisanship is often proxied imprecisely, constructed for individual states only, or excluded entirely from empirical models of local education politics. However, this omission presents a number of empirical and substantive consequences, as education we are in a moment when partisan politics are increasingly central to understanding education policy and governance. 

This project addresses this challenge by developing a comprehensive, longitudinal dataset measuring the partisan composition for the roughly 13,000 U.S. public school districts for election years from 2016 to 2024. Using Geographic Information Systems (GIS) and dasymmetric interpolation via satellite images to estimate the locations of precinct-level voters to school district geographies (Eicher & Brewer 2001), we check voter locations against areal interpolation which assigns voters to school districts using weighted area where overlaps occur. We calculate partisan composition of school districts over time using both methods.  

We will develop a validated measure of school-district-level partisanship that can be used by scholars of education policy, political behavior, federalism, and local governance. With aims of making this dataset openly available, we hope to ease a long-standing challenge in the study of education politics and policy. Additionally, we use this measure of school district partisanship, linked to district-level data from the National Center for Education Statistics Common Core of Data, to examine how change in partisan composition relates to various educational outcomes, such as enrollment, local revenue, and school choice policies.

Authors

• Joshua Almes, Brown University
  Presenting Author

• Cameron Arnzen, North Central College
  Non-Presenting Co-Author