Gender Disparity in Teaching Evaluations of Pediatric Faculty by Residents Free

Participants included pediatric faculty who served as teaching attending physicians on inpatient pediatrics services (general pediatric inpatient services, well newborn service, pediatric subspecialty inpatient services, PICU, NICU, pediatric emergency department [ED]), and pediatric faculty who supervised outpatient continuity clinic and subspecialty outpatient clinics at a single Midwestern academic tertiary care children’s hospital between January 1, 2015, and March 9, 2020. Inpatient teams were variably composed of medical students, interns (pediatric and/or off-service), residents (pediatric and/or off-service), fellows, and an attending. In the ED, residents worked directly with a faculty member or fellow–faculty dyad. Trainees rotated monthly and faculty rotated every half month on general pediatric services and weekly on inpatient subspecialty services, PICU, and NICU. Outpatient continuity clinics occurred at 1 of 9 university-affiliated, community-based pediatric ambulatory care clinics. Residents averaged 1 half day per week in continuity clinic, working with the same 1 or 2 faculty members over the 3 years of training. Subspecialty outpatient clinics occurred either at the main children’s hospital or at a university-affiliated, community-based pediatric subspecialty clinic. Residents averaged 20 half days per year for up to 2 years of residency.

Inpatient general pediatrics and subspecialty faculty were evaluated by residents at the end of a weeklong or half-month rotation, whereas faculty supervising primary care continuity clinics were evaluated by trainees twice annually. Outpatient subspecialty faculty were evaluated at the end of a half-month rotation. Completion of evaluations is mandatory; however, residents have the option to select out of completing an evaluation should they have “insufficient contact to evaluate.” Evaluations were distributed and collected using MedHub (Minneapolis, MN).

The faculty evaluation tool was developed internally by our residency program graduate medical education committee in 2006. The form utilizes a 5-point Likert scale to answer 17 prompts, including a global assessment of overall teaching ability (Fig 1). The faculty evaluation tool is used for all faculty evaluations across settings and pediatric divisions. Faculty evaluation scores are reviewed during each faculty’s annual review and are included in our institutional promotion packets.

The results of trainee evaluations of faculty members were compiled by pediatric residency program administrative staff. The faculty evaluations were deidentified before analysis and only included variables of interest including date of evaluation, rotation for which the evaluation was completed, gender of the trainee who completed the evaluation (evaluator) and of the faculty member (evaluatee), and evaluatee program (general pediatrics versus subspecialty). Overall pediatric faculty demographics and academic rank were collected from a departmental database; these were not linked to individual evaluation scores.

Our primary outcomes of interest were the mean rating of the faculty’s overall teaching ability and the mean rating of each prompt on the evaluation form. Secondary outcomes included association between evaluator–evaluatee gender concordance, teaching setting (inpatient versus outpatient), and faculty specialization status (general pediatrics versus subspecialist). Mean ratings were compared by faculty gender.

Univariate statistics were used to determine frequencies and to check the distributions of evaluation ratings scores. These faculty ratings were all left skewed, with faculty receiving a larger proportion of high ratings compared with lower ratings. The Wilcoxon Mann-Whitney test was used to look for rating differences between 2 groups. Given the large number of statistical comparisons, a more restrictive P value of .01 was used to determine statistical significance to limit those differences that could have resulted from chance. All analyses were completed using SAS 9.4 software and Microsoft Excel.

The study was determined to be exempt by the University of Michigan institutional review board.

In total, 16 845 faculty teaching evaluations were completed by trainees. One-thousand seven-hundred and three evaluations were excluded (1345 [8%] had incomplete data and 358 [2%] were of nonpediatric faculty). Of the final 15 142 evaluations included, 5091 evaluations were of male faculty (33.6%) and 10 051 were of female faculty (66.4%) (Fig 2). Most evaluations were completed by female trainees (n = 11 552, 76.3%). Two-thirds of faculty were subspecialists (67.3%). There was more even distribution among subspecialist faculty (female = 57.8% versus male = 42.4%) compared with general pediatrics, which was predominantly female (female = 86.6% versus male = 13.4%). A higher percentage of male compared with female faculty held higher academic ranks: 26.3% of males held the rank of associate professor versus only 15.3% of females, and 23.2% of male faculty held the rank of professor versus only 11.4% of female faculty (Table 1).

Faculty of both genders were rated among trainees as having excellent clinical performance and teaching ability. Female faculty were rated as having higher overall teaching ability compared with their male colleagues (female = 4.67 versus male = 4.65; P = .004) and were rated significantly higher than male faculty in 11 of the remaining 16 question prompts (Fig 3). When examining the role of gender concordance between the evaluator and evaluatee, we found no statistically significant differences for any of the evaluation tool prompts (Supplemental Table 2).

A total of 11 444 evaluations were completed in the inpatient setting (63.5% female versus 36.5% male faculty evaluated) and 3698 were completed in the outpatient setting (75.4% female versus 24.6% male faculty evaluated). In the inpatient setting, there was no statistical difference in overall teaching performance or in the mean ratings of male and female faculty for any of the evaluation tool prompts. In the outpatient setting, female faculty were rated as having higher overall teaching ability compared with their male colleagues (female = 4.79 versus male = 4.73; P = .005). They were also rated significantly higher than male faculty in 3 of the remaining 16 question prompts (Fig 4 and Supplemental Table 3).

A total of 5281 evaluations were completed of general pediatrics faculty (72% female versus 28% male faculty evaluated), which includes faculty who attended on the inpatient general pediatrics services and outpatient continuity clinic. A total of 10 409 evaluations were completed of subspecialty pediatrics faculty (63% female versus 37% male faculty evaluated), which includes faculty who attended on inpatient subspecialty pediatric services, PICU, NICU, ED, and outpatient subspecialty pediatric clinics. For general pediatric faculty, females received higher ratings than male faculty for overall teaching ability (female = 4.75 versus male = 4.70; P < .0001) and for 15 of the remaining 16 questions on the evaluation tool (Fig 5). By contrast, for subspecialty faculty, there was no statistically significant difference by gender in overall teaching ability or for any of the evaluation tool prompts (Supplemental Table 4).

This study adds to the growing body of literature looking at gender disparities in academic medicine, specifically in the female-dominant field of pediatrics. Our study found that Pediatric faculty were rated highly with similar ratings by gender. Unexpectedly, female faculty were rated slightly higher than male faculty in overall teaching ability, as well as in a majority of prompts on our evaluation tool. Although the statistically significant differences in evaluation scores were small in this large sample and may not be educationally significant, over time, statistically small effects related to bias can compound and have real-life downstream effects.^15,16 

Our findings contrast with previous studies in male-dominant fields of internal medicine and surgery where male faculty were ranked higher in overall teaching evaluations.^17,18  One possible explanation for our contrasting results is rooted in gender-based stereotypes. Social role theory is the idea that different descriptions of males and females arise from historical distribution of roles within society.¹⁹  For example, females may be more associated with caregiver roles and “communal” traits (ie, compassion, empathy) versus males associated with leadership roles and “agentic” traits (ie, decisiveness, assertiveness).^19,20  Research supports that known historical cultural stereotypes disadvantage women in agentic domains, such as medicine and leadership.^21–23  Additionally, gender-biased language continues to be found in faculty evaluations of trainees, trainee evaluations of faculty, and faculty letters of recommendation.^24–27  These findings are exacerbated in male-dominated and procedurally focused specialties,²⁸  where female physicians face misalignment between their professional role and preconceived gender-based normative behaviors.^13,29 

Our study in pediatrics differs from other studies because of the high representation of female faculty. Representation of females in a field may impact the extent of gender bias, and pediatrics’ high proportion of females may mitigate societal stereotypes explaining our findings of female faculty being scored slightly higher than male faculty. A previous study found that, across all specialties, gender bias was attenuated when there was a higher proportion of females in a field.³⁰  Our findings also align with a recent study in pediatrics that found no difference in the odds of female faculty receiving a score in the top quartile compared with male faculty.³¹  These findings have been replicated in other female-dominant health fields such as nursing, where gender bias favoring female over male nursing students have been found.³² 

Given that pediatrics is a specialty that cares for children and societal expectations of those who work with children to be warm and nurturing, there is potential alignment of expected sex- and role-expectations of pediatricians. As such, trainees may perceive communal traits to align with role expectations of a pediatrician, leading to higher evaluations seen in our female faculty. This is consistent with studies finding pediatric faculty are more likely to describe female residents with communal terms,³³  as well as patients identifying communal traits in female physicians.³⁴  Future investigation, such as qualitative assessment, is needed to explore this possible explanation further. Although our study was female predominant for both faculty and residents, gender congruence or discordance in the evaluator–evaluatee relationship was not found to statistically impact total teaching score or individual evaluation prompts. Other studies have found mixed results in terms of impact of gender dyad and evaluations in male predominant fields.^35,36 

Differences were found between teaching setting (inpatient versus outpatient) and role (general versus subspecialty faculty). Female faculty practicing in the outpatient setting and in general pediatrics had slightly higher educator ratings, but no significant difference was seen in the inpatient setting and for subspecialty faculty. Interestingly, the slightly higher evaluation of female faculty overall appears to be driven by outpatient and general pediatric faculty evaluations despite only accounting for one-fourth and one-third, respectively, of total evaluations. In both the outpatient setting and general pediatrics, there were significantly more evaluations for female faculty compared with the inpatient setting and subspecialty pediatrics (Table 1). It is known that the percentage of males and females in a field influences the perception of whether gender-based stereotypes are required for success in a particular occupation.³⁷  Given this, it is possible that the differences in gender representation may have influenced our findings to both increase resident perception of teaching in predominantly female work areas and level ratings in subspecialty areas.

Another explanation for the leveling of evaluations in the inpatient setting and for subspecialty faculty lies in social role theory. We hypothesize that the discordance between expected gender roles for female physicians and the clinical requirements of the inpatient setting, where assertiveness, decisiveness, and urgency are necessary, and of many pediatric subspecialties that remain procedurally heavy, may contradict gender-based expectations of female faculty, leading to male and female faculty being rated similarly. Previous studies have described female faculty intentionally working to navigate the more “masculine” expectations of inpatient medicine,³⁸  and showed general internal medicine female faculty were rated lower by trainees in the inpatient setting.¹⁷  Conversely, congruence between preconceived gender stereotypes and role expectations in the outpatient setting where communal behaviors, such as compassion and communication, are favored is a possible explanation for female faculty being rated slightly higher in this clinical setting. It is also possible that male faculty in the outpatient setting face incongruence between expected gender norms and role expectations, leading to lower evaluations in this setting.

It is not known if those practicing in general pediatrics or in the outpatient setting had more experience teaching trainees. Similarly, given fellowship training would not be required, it is possible those practicing in general pediatrics were more junior faculty, near peers, or able to relate to trainees in a way subspecialists were not. In addition to being early in career and more time spent teaching compared with conducting research, an area subspecialists may be more engaged in, could impact resident evaluations of faculty.³⁹  The teaching environment and time available for teaching also vary in the inpatient and outpatient setting. Each of these unmeasured factors may have influenced our findings.

Overall, these findings highlight the complex relationship between gender stereotypes and the impact of practice setting and subspecialty field on faculty teaching evaluations. Although evaluation prompts included in the institutional evaluation tool provide some information on factors residents considered when giving numeric scores to faculty members, more work is needed to understand reasons behind residents’ evaluations. For example, although the same overall trend by gender was seen for outpatient and general pediatricians, they differed significantly in responses to the 16 other evaluation questions (Figs 4 and 5). Further work, such as qualitative analysis of narrative evaluations, is needed to better understand how factors such as gender impact resident evaluations of faculty.

Despite female faculty at our institution being rated similarly to males on faculty evaluations, there continue to be lower percentages of female faculty members at the ranks of professor and associate professor, suggesting other factors contribute to the “leaky pipeline.”⁴  Future studies should seek to evaluate other factors impacting faculty promotion and advancement leading to the disparate advancement of females, like academic productivity, leadership roles, and delegation.

Our study had several limitations. First, this is the experience of a single institution with an internally created faculty evaluation tool, and thus may not be generalizable to other pediatric programs or other specialties. Evaluations at baseline in medical education are subjective and are dependent on many factors that were not measured. Overall, there is a lack of validity and reliability in evaluation tools used for faculty evaluations. Although a single evaluation tool across all faculty decreases variability in questions, some questions may be more or less applicable in different settings or interpreted differently on the basis of the clinical context. In addition, there was no rigorous evaluation of the language of the evaluation form, which could be playing a role in these gender differences. Although the sample size from our single site is quite large, this is also a limitation given that it increases the likelihood that small differences are statistically significant without true meaningful educational differences. Similarly, overall, the majority of faculty received evaluation scores on the high end, making it difficult to determine what is a meaningful difference in teaching ability. Additionally, many other variables that may impact a resident’s evaluation of a faculty member, such as length of clinical relationship and crossover between inpatient and outpatient exposures, were not controlled for. Similarly, we do not know how many faculty members were evaluated in both the inpatient and outpatient setting, or if a resident evaluated a faculty member more than once. Lastly, our study examined only quantitative and not qualitative feedback, so information is limited regarding why a trainee gave a particular response.

In conclusion, our findings show that, in the female-predominant field of pediatrics, at a single institution in the Midwest United States, female faculty are rated slightly higher by trainees than male faculty overall, in the outpatient setting, and for faculty practicing general pediatrics, although it remains unknown if these small findings are educationally significant. This work illustrates the influence of gender balance (and gender predominance in a field), practice setting, and subspecialty on faculty teaching evaluations. Culture-based gender stereotypes could potentially affect faculty evaluations; however, the role of faculty evaluations in promotion and advancement should be investigated further. Future work should continue to explore the extent of these findings across other institutions and specialties, and use qualitative methodology to better understand the potential reasons behind these findings, as well as develop evidence-based methods for mitigating gender bias.: