Teachers’ Race Linked To Students’ Scores

Both black and white children score higher on mathematics and reading tests when their teachers are the same race as they are, a study of 6,000 Tennessee schoolchildren suggests.

“Teachers, Race and Student Achievement in a Randomized Experiment,” is available from The National Bureau of Economic Research, Inc.

“I think most people in education have felt strongly that racial dynamics in the classroom are important,” said Thomas S. Dee, the study’s author. “But the few studies that have tried to find evidence of test-score effects haven’t found that.”

Mr. Dee’s report was published on the Internet last month by the National Bureau of Economic Research, a nonpartisan, scholarly group based in Cambridge, Mass. The project was paid for by the Smith Richardson Foundation of Westport, Conn.

The findings come at a time when minority teachers are at a premium in the nation’s schools. Minority children make up 40 percent of elementary and secondary enrollment nationwide, while minority teachers account for 13.5 percent of the teaching force, reports the National Center for Education Statistics. And that gap is expected to grow along with the minority population.

For his study, Mr. Dee reanalyzed data from a landmark Tennessee experiment begun in the 1980s that was designed to determine whether students learned more in smaller classes during their early-elementary years. That experiment, known as Project STAR (for Student Teacher Achievement Ratio), has attracted praise for its powerful design, which involved randomly assigning students to classes of different sizes. Some researchers consider random assignment a “gold standard” for quantitative education research.

In the STAR study, which followed students from kindergarten through 3rd grade, 94 percent of the white students and 45 percent of the black students were assigned to teachers of their own races.

“Because we have students and teachers who were randomly assigned to class sizes, we have students and teachers who are effectively randomly assigned to each other,” said Mr. Dee, an assistant professor of economics at Swarthmore College in Swarthmore, Pa.

What he found was that students who had a teacher of their own race for at least one of the four years of the study tended on average to score 3 to 4 percentile points higher on standardized tests of reading and math than peers who had teachers of different races. That kind of an edge is sizable, and only slightly smaller than the advantage that being in smaller classes gave the children who participated in the STAR study, according to Mr. Dee.

The race effects were particularly strong among poor children, children with inexperienced teachers, and children attending segregated schools—especially for African-American children. Whether or not a teacher had a graduate degree, however, seemed to have no influence on the size of the gains.

Cumulative Impact Found

The report concludes that the edge students got from having a teacher who shares their race seems to be cumulative, building for every year a student has a same-race teacher.

Any race-linked test-score differences disappeared, though, when students were assigned to smaller classes.

Ronald K. Ehrenberg, a Cornell University researcher who has also looked at the classroom impact of having same-race teachers, said that Mr. Dee’s methods seemed credible and that his work adds to a very limited body of research on a highly sensitive topic.

“You run the risk of getting attacked from all sides, so many researchers stay clear of these controversial issues. It’s good to have people like Dee doing this,” Mr. Ehrenberg said. “It would be very good to have other researchers look at this with more up-to-date databases.”

For his own studies, Mr. Ehrenberg drew on data collected from two national surveys of students. One was the National Education Longitudinal Study, a nationally representative sample of students followed since 1988. The other database was put together for the landmark Coleman Report, a congressionally mandated study by the late James S. Coleman in the 1960s. Both of Mr. Ehrenberg’s analyses were published in 1995.

In each case, he and his research partners found that having a teacher of the same race had little effect on test scores. One study did conclude, however, that teachers tended to evaluate students of their own races higher when asked in surveys for their opinions of students’ academic potential.

Policy Implications Unclear

Both researchers caution against making policy based on their studies.

“You have to look at the preponderance of evidence,” said Mr. Ehrenberg, a professor of industrial and labor relations and economics. "[Mr. Dee] is talking about one state with lower-grade students. One of our studies was middle school, and the other was elementary and high school. There are questions about generalizability in all these studies.”

Mr. Dee said he was particularly worried that his report might be used to justify school segregation, since the findings appeared to apply to both black and white children to varying degrees. One fear is that the parents of white, middle-class children might demand teachers who share their own race.

“All these findings suggest is that race in the classroom appears to matter,” he said. “But we still don’t understand the nature of that dynamic.”

Mr. Dee was unable to pinpoint, for example, whether students fared better simply because they saw their teachers as role models or because teachers, in some way, might treat students differently.

But the study does discount one other potential explanation for the test-score differential: the idea that predominantly black schools manage to attract and retain high-quality black teachers but only low-quality white teachers.

“If black teachers were simply better than white teachers on average,” the study concludes, “we would have instead expected to find that all students gained from exposure to them.”