Plenary - 01: ASSESSING GENDER BIAS IN QUALITATIVE EVALUATIONS OF SURGICAL RESIDENTS
Katherine Gerull, BA, Maren Loe, BS, Kristen Seiler, MD, Jared McAllister, MD, Arghavan Salles, MD, PhD; Washington University School of Medicine in St. Louis

Introduction: There are many well-documented disparities in the training, recruitment, promotion, and evaluation of men and women in surgery. The qualitative assessment of surgical residents may be implicitly gender biased.

Methods: We obtained 404 resident performance evaluations about 108 unique residents within 8 surgical subspecialties at one institution between 2010 and 2011. The comments were de-identified and independently coded by two individuals who identified 34 unique themes, of which 29 fell into the 6 ACGME core competencies. Differences between codes were resolved by consensus (84% agreement). Comments were also coded for valence, using a -2 to 2 scale. Finally, we examined the incidence of agentic terms (associated with leadership) and communal terms (associated with feminine traits). The data were analyzed using independent t-tests and chi-squared tests.

Results: Men received more comments about overall performance than women (5.61 for men vs 3.45 for women, p<0.001). Men also received more positive comments than women (p<0.05) in leadership (2.21 male, 0.6 female), teamwork (1.75 male, 0.0 female), and communication skills (4.76 male, 2.61 female). No categories had significantly more comments for women than men (ps>0.15). Communal words were significantly more likely to be used in evaluating female residents (p<0.05); specifically this is true for words such as cheerful/happy and enthusiastic (83%, and 59% of instances were in women’s evaluations, respectively). Agentic words were frequently seen in male evaluations, including words such as bright/intelligent/smart, well-liked, and leader (88%, 91%, and 89% of instances were in men’s evaluations, respectively).

Discussion: Evaluations of men were likely to have more comments regarding their overall performance than those of women, and men also received more comments on leadership, teamwork, and communication skills than did women. In addition, communal terms were more often used to describe women while agentic terms were almost exclusively used to describe men. This may handicap women if they are seen as less likely to perform as leaders based on these evaluations. Our analysis shows that stereotypical expectations of men and women are recapitulated in their evaluations, suggesting that implicit bias may play a role in the qualitative evaluation of surgical residents.