Student Ratings of Professors Are Not Gender Blind
Susan A. Basow, Lafayette College, Easton, PA 18042 (written 1/94)
Copyright 1994 by Susan A. Basow.
Student ratings of professors may be biased against women in subtle but significant ways.
This is not the result reported by most field research, however, as Peter Seldin has noted ("The Use and Abuse of Student Ratings of Professors," Chronicle of Higher Education, July 2, 1993). The typical study finds that the average rating of all male instructors does not differ significantly from the average of all female instructors at most colleges. This is a very reassuring finding; it is also deceptive because most studies ignore the gender of the students doing the evaluations, the disciplines involved, and the fact that female professors are often judged on a double standard.
Researchers who consider the gender of the rater find a more complex pattern. The ratings of male professors are unaffected by student gender, but female professors frequently receive lower ratings from their male students and higher ratings from their female students. Female professors also appear to be evaluated according to a heavier set of expectations than are male professors and these expectations affect student ratings.
Research conducted at Lafayette College, a small liberal arts institution in Pennsylvania, demonstrates these complexities. In one study (Basow and Silberg, 1987), 16 female professors were matched with a male professor in the same division, at the same rank, and with the same number of years at the college. More than 1,000 students in classes taught by these 32 professors filled out two questionnaires. One was a standard student rating form consisting of 26 questions, summarized into five factor scores (scholarship, organization/clarity, instructor-group interaction, instructor-student interaction, and dynamism/enthusiasm) and an overall rating. The second (the Bem Sex Role Inventory) asked students to rate their professor on two sets of personality traits: instrumental (such as assertive or dominant), often viewed as "masculine," and expressive (such as warm or nurturant), often considered "feminine."
The results revealed a consistent pattern. On all five factor scores and the overall rating, male students rated female professors more negatively than they rated male professors--and generally more negatively than did female students in the same class. This type of interaction between the gender of the student and the gender of the professor has been found in laboratory research, but less frequently in field studies, which typically neglect to ask the gender of the student rater or fail to match professors on important variables like rank and discipline.
More recent research conducted at Lafayette College confirms the effect of gender variables on these evaluations. A study of student ratings of all professors in all classes over four years reveals that male faculty were evaluated similarly by their female and male students on virtually all questions, but female faculty were evaluated differently by their male and female students--especially female faculty in the humanities and social sciences, and particularly on certain questions.
In general, female faculty received higher ratings on questions addressing interactions with students (for example, "treats students with respect"), but female students rated female faculty even higher than did male students. On questions tapping teaching style (such as, "speaks in an appropriate manner"), female faculty tended to be rated higher than their male counterparts by their female students but lower by their male students.
Few college promotion and tenure committees look at responses to specific questions, however. They usually just review the overall rating, which appears on the surface to be unaffected by teacher gender. However, when we examined student gender and course division (humanities, social sciences, natural sciences), we found that teacher gender affected student responses regarding overall effectiveness. Male students tended to rate female faculty lower, while female students tended to rate female faculty higher on this question, especially in the humanities and social sciences. Thus, certain women professors may indeed receive ratings affected by their gender.
How should we interpret these findings? First we must understand that in college teaching, males are the norm. Men are professors, women are women professors. Thus women are marked for gender in ways men are not. Indeed, students appear to respond to male professors in a uniform manner, regardless of their own gender. Students respond differently to female professors, however, perhaps because women faculty are still a minority (less than 30% of all full-time faculty are women).
Furthermore, stereotyped expectations of women (for example, to be nurturant and warm) overlap very little with expectations of professors (for example, to be knowledgeable and competent). However, stereotyped expectations of men coincide with what we expect of professors.
Research documents that people who violate expectations generally are rated more negatively than people who behave as expected. To receive good evaluations, male professors simply must demonstrate their competence and knowledge; that is, they need to fulfill their stereotypical gender role expectations. But female professors bear a double burden: they must fulfill both their gender role by being nurturant and warm, as well as their professional role by being competent and knowledgeable.
For example, separate studies led by Sheila Bennett and Anne Statham found that women professors are judged more negatively than males if they are not more interested in and available to students than male professors. But even when women professors are more available and more helpful, their overall ratings are no higher. In order to receive comparable ratings, female professors need to do more than their male counterparts. Thus, findings of no difference between male and female professors in overall ratings may mask the fact that different standards are being used to judge male and female faculty.
In the Basow and Silberg study, female faculty who received low ratings on instructor-student interaction also got low marks on gendered personality traits, suggesting that they may not have fulfilled the double set of expectations placed upon them. In the more recent research, the higher ratings of female faculty on several interpersonal questions did not always translate into better overall ratings, especially when the rater was male. Again it seems that in order to receive comparable overall ratings, women professors must be better interpersonally when compared to their male counterparts.
Why do male students tend to rate certain female faculty more poorly than male faculty? Male students may be more influenced by gender stereotypes than are female students. Research documents that men, compared to women, hold more traditional attitudes toward gender roles and demonstrate more bias against gender-role violators. In the Basow and Silberg study, males majoring in business and economics or in engineering rated female faculty most negatively. We found that those students have the most traditional attitudes toward women and gender roles.
Ratings may also reflect gender differences in preferences regarding teaching style. Female students may appreciate a warmer and more involved interpersonal style, as perhaps found in their female teachers, whereas male students may appreciate a cooler interpersonal style but more displays of scholarship, as perhaps found in their male teachers.
The effects of gender on student ratings of professors are complex but real, and should not be dismissed with a general statement of "no effect." Although small on average, these effects may be quite marked for specific teachers. For instance, a female teacher whose direct teaching style lacks marked warmth or friendliness may find the cards stacked against her when teaching male students in a field where women are a rarity.
Studies at a wider range of institutions will help scholars better understand how and when gender factors affect student ratings, particularly teacher gender in interaction with both student gender and divisional affiliation. Until then, those of us who evaluate female faculty must be alert to the various and subtle ways in which gender bias can affect perceptions and evaluations.
Basow, Susan A. (1994). Student evaluations of college professors: When gender matters. Manuscript in preparation.
Basow, Susan A., and Silberg, Nancy T. (1987). Student evaluation of college professors: Are female and male professors rated differently? Journal of Educational Psychology, 79 (3), 308-314.
Bennett, Sheila. (1982). Student perceptions of and expectations for male and female instructors. Journal of Educational Psychology, 74, 170-179.
Statham, Anne, Richardson, Laurel, and Cook, Judith. (1991). Gender and University Teaching: A Negotiated Difference. SUNY Press.
People are welcome to make use of this essay as long as they cite author and title.
Susan A. Basow, Psychology Dept.
Lafayette College, Easton, PA 18042-1781
Re: Susan Basow's remarks about sex bias.
From: Michael Scriven
Good to keep an eye on this possibility, but two cautions:
- Finding small differences here can't be taken into account unless there's comparable research on all other types of bias, e.g., with respect to age (a serious one), race (ditto), good looks (allegedly but probably not serious). For example, it's a common feeling among elderly male profs that they are at an unfair disadvantage in humanities courses against young female profs (partly but not only because the latter tend to have more female students). Otherwise, it seems to me that if one applied an across the board correction, women would be receiving differentially favorable treatment, though the logic for that conclusion needs careful examination. But, in very special cases (woman engineer teaching first year courses is the usual case), this possibility should be addressed.
- Small differences in overall ratings should absolutely never be given any importance, partly because small differences may be due to bias of one kind or another. In the first place, evaluation of teaching should not be done solely on the basis of student ratings. It should include knowledge of content, out of class contributions to curriculum committees, technical quality of tests and marking procedures (blind grading, etc.), grading distribution (in extreme cases only), etc., etc.
- None of this research is worth much unless it is done with respect to a defensible rating form. There are few of those, for the reasons I've alluded to here, e.g., because most of them include references to teaching style (enthusiasm, small groups, advance organizers, etc.) and that makes them almost totally invalid, because of contamination by that question (cf. interviews where just one personal-life question is raised, whether or not it is answered).
So I think research on bias should continue, should be extended to other serious types of bias, and that the first line of protection is requiring big differences to justify any action or award. But until it's done on good forms, it's extremely hard to judge whether it calls for a change in policy e.g., a correction factor.
Last modified 2005-01-19 01:17 PM