Same race teachers do not necessarily raise academic achievement

Same race teachers do not necessarily raise academic achievement. Jeffrey Penney. Economics Letters, Volume 223, February 2023, 110993. https://doi.org/10.1016/j.econlet.2023.110993

Abstract: Numerous studies have found that students who are of the same race as their teacher experience increased academic achievement. In this paper, I attempt to explain when these benefits occur and which students are most likely to achieve the largest gains. Using exogenous variation in student–teacher matches and classroom composition from Tennessee’s Project STAR experiment, I find that below average achieving students benefit most from having a teacher of the same race, but the benefits from matching can be substantially reduced in smaller classes. Moreover, the effect is decreased in racially homogeneous classes where the teacher is the majority race.

Introduction

There is substantial evidence that students do better academically when matched with a teacher of the same race (e.g. Dee, 2004, Fairlie et al., 2014, Egalite et al., 2015, Penney, 2017, Delhommer, 2022).1 However, educational interventions thought to improve student achievement can potentially fail to achieve the desired results under some circumstances; for example, Gilraine (2020) finds that class size reductions only increase student achievement when experienced instructors are teaching the smaller classes.

In this paper, I investigate the heterogeneities of own-race teacher effects, considering their distribution and examining which scenarios they are most likely to be beneficial. I conduct an empirical analysis making use of data from Project STAR, a large-scale education experiment in Tennessee that was designed to investigate the effects of class size and teacher’s aides on student achievement. This is the first paper that focuses on investigating own-race teacher effects specifically with regards to student achievement and classroom composition using experimental data, filling a crucial gap in the literature.

The results of this analysis are as follows. I find that the benefit of student–teacher racial matching varies with academic ability, with lower-scoring students generally profiting the most whereas above-average achieving students see no distinguishable increase in test scores. An examination of the effects of classroom composition reveals that the magnitude of own-race teacher effects varies along this dimension: where the teacher shares the same race as only a few of their students, the own-race teacher effect on mathematics and reading scores is high; if instead almost all students in the class are the same race as their teacher, it is considerably lower.

This paper is organized as follows. Section 2 outlines the data from the Project STAR experiment and considers several threats to validity. The empirical analysis takes place in Section 3. The paper concludes with a discussion of the policy implications in Section 4.