A new report by Seth Gershenson sparks fresh ideas about new directions for the literature on student-teacher matching along demographic characteristics. While previous work has shown teacher-student race/ethnicity matching has a detectable impact on test scores, academic perceptions and attitudes, attendance and suspensions, gifted and talented referrals, and educational attainment, this new work offers a fresh perspective by examining differences in exposure and impact associated with assignment to same race/ethnicity teachers between the traditional public school and public charter sectors.
Here’s what he finds: The North Carolina educator workforce in both school sectors—charter schools and traditional public schools—is dramatically unrepresentative of the student body it serves, which is 54 percent white, 26 percent Black, 13 percent Hispanic, and 7 percent other races. The report focuses on black teachers only, noting that charter schools fare slightly better than the state's traditional public schools as they have more black teachers (about 3.5 percentage points’ difference).
The report then compares the magnitude of matching effects by school sector. Teacher-student race/ethnicity matching has no effect on student achievement in English language arts and a small, positive effect on math scores of 0.018 of a standard deviation (SD). Further examining the math effect by sector, we see a larger effect of race/ethnicity matching in charter schools (0.029 SD) compared to traditional public schools (0.016 SD). What should we make of all this? I see three key takeaways.
First, we need to be pushing harder on the question of what is different about the charter school environment that seems to magnify the potential benefits of race-matching. Gershenson proposes a couple of ideas that at least explain why minority teacher representation is higher in the charter sector, such as North Carolina’s approach to teacher certification in the charter sector: “North Carolina requires that only half the teachers in charter schools be certified,” the report explains. “Given racial gaps in passage rates on teacher certification tests, this relaxed rule might contribute to sector differences in the racial representation of the teaching force.” But we've only scratched the surface of potential explanations for why the race/ethnicity matching impact is larger in the charter sector. It will take further research to test plausible hypotheses. Perhaps greater principal autonomy in hiring the teachers they are most excited about plays a role. Perhaps flexibility over curriculum choices, textbooks, and scheduling is important (a longer school day would create more opportunities for teachers and students to form a bond, after all). Perhaps it’s not a race/ethnicity matching effect at all, but a culture match resulting from charter schools finding it easier to hire uncertified teachers to work in the neighborhoods they call home. In a nutshell, what are the structural barriers keeping traditional public schools from diversifying their teacher workforce, and how might they learn from the charter sector?
Second, we should be paying greater attention to the intersection between the race/ethnicity matching literature and school choice research, more broadly. The charter effects presented here are tantalizing, raising questions about how this phenomenon also plays out in the context of private school choice, where schools have even greater autonomy, often a religious orientation, and very different organizational structures. Students arriving with a state-sponsored voucher or tax credit scholarship may be positively impacted by the presence of teachers of color in their private school of choice. Are the race/ethnicity matching effects even larger there? If so, why?
Third, and finally, when sample sizes are reasonably large, I would love to see greater attention to the Hispanic population in studies of this nature. I found it insightful and helpful that the report describes the disparity between the percentage of black teachers (14 percent of all charter school teachers are black) and black students (27 percent of all charter school students are black) before introducing the impact estimates. Understandably, there might be concerns about highlighting potentially underpowered impact estimates for Hispanic teacher/student matching (the appendix reports there are null race-matching effects for this group), but let’s at least examine the descriptive statistics for this group so we can get a sense of how often Hispanic students—which make up 5 percent of all charter school students—encounter a teacher that looks like them.
Reflecting on this report also got me thinking hard about the state of the race-matching literature. The sector differences uncovered here are small but meaningful, and they should prompt us to think about where this research should go next. What outcomes should be examined, for whom, and during what time periods?
When choosing which outcomes to examine, let’s pay more attention to differences in the magnitude of race-matching impacts across the various outcomes that have been examined to date. Numerous theoretical frameworks have been proposed in the race-matching literature, none of which make the case for effects showing up most strongly in students’ academic achievement. Indeed, the test score impacts observed here and elsewhere are relatively small, so let’s return to the theories about why this potentially matters before choosing which outcomes future research should examine. In my own work with Brian Kisida, we have found that students assigned to a demographically similar teacher report feeling more cared for, more motivated to do their best work in class, and experience higher quality student-teacher communication. Related work has found that teachers’ expectations play a role in future success. These findings paint a picture of the mechanisms involved and suggest that demographically similar teachers are not merely playing a passive symbolic role for students. They are actively forging improved relationships, setting higher expectations, and potentially serving as cultural translators.
When choosing which student subgroups to examine, let’s do our best to understand impacts for subgroups like Hispanics, which represent a small (10 percent) but rapidly growing segment of North Carolina’s population. This report finds null impacts for this group. Is that because the analysis was underpowered or because no such effect exists? It’s hard to tell but important to think about moving forward. We can think of race/ethnicity matching as providing social-emotional support, comfort, and reassurance to students of color, especially during stressful times, but these effects might play out differently for different student groups. In general, a more diverse educator workforce perhaps reinforces a positive self-identity for students of color, temporarily relieving the pressure to code-switch, and allowing them to let down their guard. Furthermore, a more diverse faculty allows students of color to feel seen and understood, knowing the adults in their lives share common experiences with them. But the potentially varying experiences of different student subgroups deserve our attention, especially if the effects are not consistent across outcomes.
Returning to the underlying theory also gives us more insight into the time periods we should be examining. For example, research could test if race/ethnicity matching has differential impacts during key transition periods, such as the first weeks of a new school year, particularly if a student is new to that school, or during structural transitions, such as the jump from middle to high school.
The state of educator diversity research can be pushed forward by thoughtful work like this, but let’s not lose sight of the underlying theories as we develop the next generation of research questions.