The Effect of Desegregation on Academic Performance: Evidence from Huntsville City Schools

Abstract

This paper analyzes the effect of the rezoning of attendance zones in Huntsville City Schools in the fall of 2015 on academic performance. More specifically, using district-level data published by the Stanford Education Data Archive and the Alabama State Department of Education, I employ the synthetic control method, which constructs a counterfactual using a weighted combination of other school districts in Alabama, to evaluate the effect of the rezoning on math and reading standardized test scores for students in grades 3–8 in Huntsville City Schools. This paper contributes to the literature in two ways. First, I use machine learning to harmonize data from the two sources and create a unified data set that contains a consistent measure of academic performance for all school districts in Alabama from the 2008–09 school year to the 2018–19 school year, inclusive. Second, using the synthetic control method, I analyze the effect of the rezoning on standardized test scores in Huntsville City Schools and find suggestive evidence that the policy did not have an effect on academic performance in the district. This result holds across different subgroups and is validated by a number of robustness checks that include modifying the predictors used in the analyses, using nested optimization to construct the synthetic controls, and running placebo tests.