Teachers, race and student achievement revisited

Abstract

Within the education literature, a controversy exists with respect to the issue of matching student and teacher race in an effort to improve student performance. Ehrenberg et al. (1995) finds very little support for this issue, while more recently Dee (2004) finds that there are significant educational gains when students are assigned to an own-race teacher. Dee's result is found after confirming that there was no association between assignment of an own-race teacher and student characteristics, i.e., sorting of students did not transpire. We extend Dee's work by including the effects of student innate ability and teacher gender on student achievement. Our findings indicate that once these two variables are taken into consideration, sorting of students does transpire, and matching students and teachers of similar race has no statistically significant affect on student achievement.

Notes

¹ Data for the variables described in the texts were collected for third-grade students in individual elementary schools in the state of Kentucky for the school year ending in 1990. Even with omissions due to missing data, 25 871 students enrolled in 1192 classes in 120 schools were employed in the analysis. The Division of Evaluation, Kentucky Department of Education, provided the primary data.

² Teacher gender and student innate ability were added to the math and reading achievement models separately before the models were estimated with both variables. The results suggested that adding teacher gender did not change the significance of the own-race teacher variable. The inclusion of only the student innate ability variable did result in the change in significance of the own-race teacher variable. Both variables were included in the final model to address Dee's concern about omitting gender and Hanushek's concern about omitting a measure of innate ability in achievement equations.