Using data from the National Educational Longitudinal Study of 1988 (NELS), the authors find that the match between teachers' race, gender, and ethnicity and those of their students had little association with how much the students learned, but in several instances it seems to have been a significant determinant of teachers' subjective evaluations of their students. For example, test scores of white female students in mathematics and science did not increase more rapidly when the teacher was a white woman than when the teacher was a white man, but white female teachers evaluated their white female students more highly than did white male teachers.