Program Description

For this study, faculty evaluations were used from all courses offered through Mayo Clinic Center for Clinical and Translational Science (CCaTS) track within the Mayo Clinic Graduate School of Biomedical Sciences (MCGSBS) for the 2017-2018 and 2018-2019 academic years. The program includes approximately ~52 unique courses delivered once or twice a year, or every other year, with each course having between 1 and 3 course directors. Coursework includes content across the clinical translational spectrum, from basic laboratory and clinical research to population and policy-focused work.

Student Evaluation Data

This study utilized data gathered as part of the standard CCaTS post-course evaluation process. A standard evaluation form is sent to students at the end of each quarter for each course in which they are enrolled. The evaluation is anonymous and optional, though with high completion rates. The primary item of interest for this study was “Please assign a grade for the co-director, [name here], and comment on what you liked or can be improved.” This item included both quantitative data (assigning faculty a grade) and qualitative data (commenting on areas for improvement in an open text box). Students could assign faculty a letter grade (A, B, C, D, F), which was assigned a corresponding number on a Likert scale (4, 3, 2, 1, and 0, respectively). Data were used in aggregate (average for each course delivery) for each faculty member; that is, each faculty member received a mean grade for each course, yielding faculty*course units.

Faculty Information

For each of the courses included, faculty degree, rank, gender, and team composition were gathered and incorporated into student evaluation feedback. Several faculty members teach multiple courses; this was taken into account during data cleaning and analysis. Once data were merged for the quantitative and qualitative data sets, faculty names and course numbers were removed.

Data Analysis

Quantitative data were analyzed to generate descriptive statistics, and to test for differences between males and females using chi-squared/fisher’s exact and/or one-way and two-way ANOVA tests, depending on the variable and data type.

For qualitative data, two reviewers separately coded each as having gendered language (Y/N); containing positive, negative, or mixed feedback; and key words (adjectives, adverbs specifically) were noted.

There were 57 unique faculty members who taught at least one course during the 2017-2018 and 2018-2019 academic years; a total of 88 course offerings were delivered within the CCaTS program during those two years (Table 1). As most courses are team-taught by 2 or 3 faculty members, this yielded 144 faculty*course units. There were statistically significant differences in terms of highest degree attained by gender (p=0.02). Of these, 24 (42%) were female and 33 (58%) were male. Among the 32 faculty with a PhD, 19 (59.4%) were female and 13 (40.6%) were male; among the 17 faculty with MDs, 4 (23.5%) were female and 13 (76.5%) were male. There were no female faculty with dual MD/PhD degrees; however, there were 3 male faculty members who held both.

There were also statistically significant differences in rank by gender (p=0.018; Table 1). Of the 10 who held the rank of Assistant Professor, 8 (80%) were female; 7 of the 14 (50%) who held the rank of Associate Professor were female; and only 6 of 24 (25%) of individuals who held the rank of full Professor were women.  There were not enough data to analyze fields of rank for each faculty member; however, more women held rank in epidemiology and health services researchers than men, and more men held rank in medicine. Notably, biostatistics was fairly evenly split (5 [45.5%) female and 6 [54.5%] male).

Overall, there was a borderline statistically significant difference (p=0.051) in faculty grades between female and male faculty members, with female faculty members scoring slightly higher (Table 2). No statistically significant difference was found when adding highest degree earned to the analysis (p = 0.647; Table 2). However, when examining the composition of the teaching team for each class, significant differences were seen in grades, with male course directors teaching solo earning the lowest grades, and mixed gender course director teams earning the highest grades (p=0.033).

Review of qualitative comments revealed that there were few explicitly gendered comments made by students in providing faculty feedback. Our team noted that comments about female faculty tended to focus on personality and perceived authority, while comments about male faculty addressed expertise and included more superlatives. For example, students suggested that some female faculty “take more ownership,” appreciated that they were “passionate,” and were “approachable” and “nice.” Male faculty members were more frequently described as “dismissive,” “thoughtful,” “knowledgeable,” and “excellent.”

Our findings are in stark contrast to those reported previously for faculty in medical schools, in which males receive more favorable grades for their teaching than women (Morgan et al., 2016). In this study, we found that women in the graduate research program received higher grades on average for teaching than their male counterparts. This is of particular interest as women with advanced degrees in biomedical science are increasingly leaving academia for industry and government due to personal, interpersonal, and professional factors (the proverbial “leaky pipeline”) (Sarseke, 2018; Ysseldyk et al., 2019). Female faculty in the biomedical research space may face a catch-22 in that, while they are viewed more favorably by their students, they have a steeper climb for promotion and tenure in the academic world than their male counterparts (Carr et al., 2015). Despite the higher ratings quantitatively, subtle differences in language in student comments reinforce that gender patterns and biases still exist; this has been established previously in several other fields of academia (Basow et al., 2006).

Furthermore, these results support the fact that teaching teams involving women in the biomedical/translational sciences are rated more highly by students than male or female faculty teaching alone. This is in line with studies that have shown that teams involving women are more successful than those without women, as measured by collective intelligence, and that diversity is associated with creativity and productivity (Woolley and Malone, 2011; Rock and Grant, 2016; Kubik-Huch et al., 2019). Therefore, seeing this reflected in our education data shows that having women as part of the team and team success translates to graduate biomedical research education.

This is the first study that our team is aware of to explore gender differences in student evaluations of graduate biomedical research faculty. A strength of this study is that the CCaTS faculty is diverse in their areas of expertise which is reflected in the variety of courses offered through the center; therefore, this study encompasses faculty who teach basic, clinical, population health, and policy-focused courses. Limitations of this study include those associated with small sample sizes, such as limitations in the type of analyses we were able to conduct. Additionally, while the content taught by faculty spans the entire translational spectrum, these findings are only representative of one track within MCGSBS; data were not available school-wide, as CCaTS has led the way with implementing these evaluation procedures. Finally, we did not have data as to which students completed the evaluation surveys, so could not determine whether there were differences on faculty rating based on whether the students were male or female. This factor has been shown in other studies to influence the association between gender and faculty teaching evaluations (Santhanam and Hicks, 2002; Sprague and Massoni, 2005). Future work should emphasize faculty in more “basic” science settings and include data to explore whether faculty ratings depend on student gender as well as the gender of the faculty, and should be expanded to include non-binary gender categories (MacNell et al., 2015).