Preceptor and student gender effects on clerkship assessment scores in medical education Authors: Elisabeth Kac, Isain Zapata, Amanda Brooks, TyRee Jenks, Terry L Hudgins Published: 2026 ABSTRACT AbstractBackgroundClerkship evaluations play a critical role in assessing medical student performance and informing residency selection. Prior studies suggest there are gender-related differences in narrative feedback and numerical evaluations of medical trainees; however, findings remain inconsistent across specialties and institutions. This study examined whether preceptor gender and student gender were associated with differences in clerkship evaluation scores.MethodsA retrospective analysis was conducted using 6,855 clerkship evaluations from students in the graduating classes of 2022–2024 at a U.S. Osteopathic medical school. Evaluations were completed by 2,425 preceptors across six required clerkships: Family Medicine, Internal Medicine, Obstetrics and Gynecology, Pediatrics, Psychiatry, and Surgery. Associations between preceptor gender, student gender, and evaluation scores were analyzed using generalized linear mixed models, adjusting for academic year and clerkship rotation.ResultsPreceptor gender was significantly associated with evaluation scores, with female preceptors assigning modestly lower Total Average and Clinical Competence scores than male preceptors (P < 0.0001). Student gender demonstrated smaller effects, with female students receiving slightly higher Total Average scores (P = 0.0260) and Professionalism scores (P = 0.0026). No significant interaction between preceptor and student gender was observed. Clerkship rotation was strongly associated with evaluation scores, with higher adjusted scores in Family Medicine and Pediatrics and lower in Surgery and Internal Medicine.ConclusionEvaluator characteristics and specialty-specific grading cultures appear to influence numerical clerkship assessments. Although findings may be limited by the single-institution design, these results highlight the potential value of increased standardization and target preceptor development to promote equity and consistency in clinical evaluation.Keywords: Clerkship evaluation, Gender bias, Medical students, Clinical assessment, Preceptor characteristics, Grading variability Introduction Clinical clerkship evaluations are a major component of medical student assessment and may influence residency applications, academic opportunities, and professional development. Given the subjective nature of clinical evaluations, concerns remain regarding the extent to which implicit bias may influence assessment and grading practices within medical education. Prior research has identified potential gender differences in narrative evaluations of medical trainees. A comprehensive review spanning several decades found that 62.5% of included studies identified evidence of gender bias in subjective evaluations, most commonly demonstrating higher quantitative ratings for male students as well as differences in the language used in narrative assessments [1]. In emergency medicine and surgical rotations, studies have shown that male trainees are more frequently described using assertive and leadership-oriented traits, whereas female trainees are more often characterized using interpersonal and supportive descriptors [2, 3]. In addition to disparities observed in narrative assessments, studies have examined whether similar gender-related patterns are present in students’ numerical grades. In one internal medicine clerkship, women outperformed men in five of eight evaluation categories but did not receive higher overall grades [4]. In contrast, during an obstetrics and gynecology rotation, women earned higher final grades and achieved higher scores on the standardized National Board of Medical Examiners (NBME) exam [5]. However, other studies have reported no gender-based differences in numerical grades, suggesting that additional factors may influence how final grades are assigned [6]. These mixed findings indicate that numerical grading patterns may vary based on underlying contributors, underscoring the need for closer examination of factors that shape assessment outcomes. The objective of this study is to evaluate gender bias in core clerkship grading among third-year medical students in a U.S. osteopathic medical school. We hypothesized that assessment scores would vary based on student and evaluator gender. Identifying where and how these disparities arise may inform efforts to develop evaluation systems that reduce implicit bias and more accurately reflect clinical performance and potential. Methods Medical student population This study included third-year medical students (OMS III) from the classes of 2022, 2023, and 2024 at Rocky Vista University College of Osteopathic Medicine. Students were enrolled at the Colorado and Utah campuses and completed required core clerkships at clinical training sites across the United States. Clerkship participation occurred during the 2021–2022, 2022–2023, and 2023–2024 academic years. The sample consisted of 495 male students (57%) and 374 female students (43%) with a total of 6,855 evaluations submitted. Gender classifications were derived from institutional records and publicly available identifying information, when necessary, as the retrospective dataset did not include self-reported gender identity. Classifications were limited to male and female categories because no additional gender identity data were available within the dataset. Student performance data were obtained from preceptor-submitted evaluations. Records were de-identified prior to analysis. The study was reviewed and approved by the Rocky Vista University Institutional Review Board (IRB# 2019-0081). Clerkships and preceptors Evaluations were obtained from six required core clerkships: Family Medicine, Internal Medicine, Obstetrics and Gynecology, Pediatrics, Psychiatry, and Surgery. Preceptors practiced at clinical sites across Arizona, Colorado, Idaho, Kansas, Nevada, New York, South Dakota, Utah, and Wyoming. A total of 6,855 evaluation forms including 3,978 male student evaluations and 2,877 female student evaluations. These were submitted by 2,425 unique preceptors, including 1,993 male preceptors (82.2%) and 432 female preceptors (17.8%). Preceptor gender classifications were derived from institutional records and publicly available identifying information, when necessary, as the retrospective dataset did not include self-reported gender identity. Classifications were limited to male and female categories because no additional gender identity data were available within the dataset. Preceptors assessed students using standardized evaluation forms administered through the New Innovations system and mapped to the Association of American Medical Colleges Core Entrustable Professional Activities (EPAs). The evaluation included 18 items, consisting of 14 clinical competence items (score range 0–10) and 4 professionalism items (score range 0–16), along with qualitative narrative comments based on direct observation of clinical performance. A composite Total Average score (0–10.75) was calculated. Evaluations were completed at the conclusion of each clerkship rotation via automated email notification to preceptor (Fig. 1). Study population and evaluation dataset. Summary of student cohort, preceptors, and clerkships included in the analysis Statistical analysis Clerkship evaluation data was deidentified and compiled into a single dataset for analysis. The experimental unit for this study was the individual evaluations. Data was first assessed descriptively using means and standard deviations for continuous variables and frequencies and percentages for categorical variables. Then data was evaluated for associations using Generalized Mixed Models (GLMMs). For these models the Total average, Clinical Competence average and Professionalism average scores were defined as the dependent variables while year cohort (2022–2024), Evaluator’s gender, Student’s gender and the Clerkship rotation were defined as independent variables. These models were developed in two iterations, first a model evaluating the evaluator and student bias effects along with their interaction without considering the clerkship rotation. A second model followed focused on estimating the evaluator bias within each specific clerkship rotation, this was achieved though the evaluator’s gender by clerkship rotation interaction. This second model included the adjustments for the first model. All model estimates are presented along with their standard error of the mean (SEM). All descriptive and modeling assessments were performed using SAS v.9.4 (SAS Institute Inc., Cary, NC). Significance is declared at P ≤ 0.05. Results A total of 6,855 preceptor evaluations were included in the analysis. The study consisted of 3,978 evaluations for male students (58.03%) and 2,877 evaluations for female students (41.97%). Evaluations were submitted by 2,425 unique preceptors, including 1,993 male preceptors (82.2%) and 432 female preceptors (17.8%). Male preceptors submitted 5,029 evaluations (73.4%) and female preceptors submitted a total of 1,826 evaluations (26.6%). Preceptor-student gender pairings included 3,044 evaluations of male students by male preceptors, 1,985 evaluations of female students by male preceptors, 934 evaluations of male students by female preceptors, and 892 evaluations of female students by female preceptors. A summary of the evaluations used in this study are presented in Table 1. Table 1Evaluations frequency descriptive of the cohort assessed in this study. These evaluations are presented stratified by evaluator’s genderVariableTotalFemale EvaluatorMales EvaluatorFrequencyPercentFrequencyPercentFrequencyPercentYear2022242135.3263734.88178435.472023230833.6762234.06168633.532024212631.0156731.05155931Evaluator GenderFemale182626.641826100--Male502973.36--5029100Student GenderFemale287741.9789248.85198539.47Male397858.0393451.15304460.53RotationPsychiatry78311.4224213.2554110.76Family Medicine108815.8728115.3980716.05Internal Medicine165624.1630216.54135426.92Pediatrics85812.5227014.7958811.69Surgery162523.7132517.8130025.85Obstetrics & Gynecology84512.3340622.234398.73 Evaluations frequency descriptive of the cohort assessed in this study. These evaluations are presented stratified by evaluator’s gender Associations between preceptor gender, student gender, and evaluation scores were examined through linear modeling. Across all evaluations, the mean Total Average score was 9.54 ± 1.80. Academic year was significantly associated with Total Average scores (P = 0.0435), with higher scores in 2022 compared with 2023 (P = 0.0132). No significant differences were observed between 2022 and 2024 or between 2023 and 2024. Preceptor gender was also significantly associated with Total Average scores, with female preceptors assigning lower average ratings than male preceptors (0.309 ± 0.049, P < 0.0001). Student gender was likewise significant, with female students receiving higher ratings than male students (0.110 ± 0.049, P = 0.0260). The interaction between preceptor gender and student gender was not significant (P = 0.7136), indicating that preceptor-student gender combinations were not associated with differences in overall scores. Model estimates for Total Average scores are presented in Table 2. Table 2Evaluator and student gender bias estimates obtained through generalized linear models. Each dependent variable (Total average, Clinical competence, and Professionalism) was evaluated separately. Total Average scores ranged from 0–10.75, Clinical Competence scores ranged from 0–10, and Professionalism scores ranged from 0–16. Errors presented are standard errors of the mean (SEM)VariableLevelTotal AverageClinical CompetenceProfessionalismEstimateP-valueEstimateP-valueEstimateP-valueYear20229.54 ± 0.040.04357.98 ± 0.040.015114.00 ± 0.050.596920239.41 ± 0.047.82 ± 0.0413.94 ± 0.0520249.46 ± 0.047.88 ± 0.0413.95 ± 0.05Evaluator GenderFemale9.32 ± 0.04< 0.00017.70 ± 0.04< 0.000113.90 ± 0.050.0727Male9.63 ± 0.038.09 ± 0.0314.02 ± 0.03Student GenderFemale9.53 ± 0.040.02607.93 ± 0.040.121014.05 ± 0.050.0026Male9.42 ± 0.037.85 ± 0.0313.87 ± 0.04Evaluator Gender x Student GenderFemale x Female9.36 ± 0.060.71367.72 ± 0.060.648514.01 ± 0.080.7821Female x Male9.27 ± 0.067.67 ± 0.0613.80 ± 0.07Male x Female9.69 ± 0.048.14 ± 0.0414.10 ± 0.05Male x Male9.56 ± 0.038.04 ± 0.0313.93 ± 0.04 Evaluator and student gender bias estimates obtained through generalized linear models. Each dependent variable (Total average, Clinical competence, and Professionalism) was evaluated separately. Total Average scores ranged from 0–10.75, Clinical Competence scores ranged from 0–10, and Professionalism scores ranged from 0–16. Errors presented are standard errors of the mean (SEM) The Clinical Competence average score was 7.98 ± 1.86. Preceptor gender remained a significant factor, with female preceptors assigning lower Clinical Competence scores than male preceptors (0.394 ± 0.051, P < 0.0001). Academic year was also significantly associated with Clinical Competence scores, with higher scores observed in 2022 compared with 2023 (P = 0.0041), while no significant differences were identified between the other years. Student gender was not significantly associated with Clinical Competence scores (P = 0.1210), and no interaction between preceptor gender and student gender was observed (P = 0.6485) (Table 2). The Professionalism average score was 13.97 ± 2.26. Student gender was significantly associated with Professionalism scores, with female students receiving higher ratings than male students (0.187 ± 0.062, p = 0.0026). Academic year was not significantly associated with Professionalism scores (P = 0.6969). Preceptor gender (P = 0.0727) and the preceptor-student gender interaction were also not significant (P = 0.7821) (Table 2). Clerkship rotations were included in all models to account for differences in grading practices across clinical specialties and were significantly associated with evaluation scores. For Total Average scores, rotation was a significant predictor (P < 0.001), with adjusted mean scores differing across clerkships. The highest adjusted Total Average scores were observed in Pediatrics and Family Medicine whereas lower scores were observed in Internal Medicine and Surgery. Similarly, rotation was significantly associated with Clinical Competence scores (P < 0.001). The highest adjusted Clinical Competence scores were observed in Pediatrics and Family Medicine, while lower scores were observed in Surgery and Internal Medicine. Rotation also demonstrated a strong association with Professionalism scores (P < 0.001). Professionalism scores varied across clerkships, with higher adjusted scores observed in Family Medicine, Pediatrics, and Women’s Health and lower scores in Surgery and Psychiatry. Across all evaluation domains, Family Medicine and Pediatrics consistently demonstrated among the highest model-adjusted scores, whereas Surgery consistently demonstrated among the lowest. Preceptor gender by rotation interactions were observed, with most consistent differences seen in Family Medicine, Internal Medicine, and Obstetrics and Gynecology, where female preceptors assigned lower scores than male preceptors. More limited or domain-specific differences were observed in Psychiatry and Pediatrics, while no significant differences were identified in Surgery. No interaction between preceptor gender and student gender was identified across rotations (Table 3). Table 3Clerkship rotation bias estimates obtained through generalized linear models. Each dependent variable (Total average, Clinical competence and Professionalism) was evaluated separately. Total Average scores ranged from 0–10.75, Clinical Competence scores ranged from 0–10, and Professionalism scores ranged from 0–16. Errors presented are standard errors of the mean (SEM). The difference corresponds to the male minus female evaluator scoresVariableRotationAll evaluators estimateFemale evaluator estimateMale evaluator estimateFemale vs. Male evaluator differenceFemale vs. Male evaluator P-valueTotal averagePsychiatry9.49 ± 0.079.37 ± 0.119.61 ± 0.08−0.23 ± 0.140.0930Family Medicine9.59 ± 0.079.22 ± 0.119.96 ± 0.06−0.74 ± 0.12< 0.0001Internal Medicine9.37 ± 0.069.14 ± 0.109.60 ± 0.05−0.46 ± 0.11< 0.0001Pediatrics9.76 ± 0.079.63 ± 0.119.89 ± 0.07−0.26 ± 0.130.0521Surgery9.34 ± 0.069.41 ± 0.109.26 ± 0.050.15 ± 0.110.1861Obstetrics & Gynecology9.54 ± 0.069.21 ± 0.099.86 ± 0.09−0.65 ± 0.12< 0.0001Clinical CompetencePsychiatry7.91 ± 0.077.82 ± 0.127.99 ± 0.08−0.17 ± 0.140.2227Family Medicine8.01 ± 0.067.54 ± 0.118.49 ± 0.07−0.95 ± 0.13< 0.0001Internal Medicine7.73 ± 0.067.39 ± 0.118.06 ± 0.05−0.68 ± 0.12< 0.0001Pediatrics8.19 ± 0.078.05 ± 0.118.32 ± 0.08−0.27 ± 0.140.0492Surgery7.83 ± 0.067.92 ± 0.107.75 ± 0.050.17 ± 0.110.1378Obstetrics & Gynecology7.91 ± 0.067.55 ± 0.098.26 ± 0.09−0.72 ± 0.13< 0.0001ProfessionalismPsychiatry13.77 ± 0.0913.54 ± 0.1414.01 ± 0.10−0.47 ± 0.170.0068Family Medicine14.34 ± 0.0814.17 ± 0.1314.51 ± 0.08−0.33 ± 0.160.0310Internal Medicine14.03 ± 0.0714.10 ± 0.1313.97 ± 0.060.13 ± 0.140.3488Pediatrics14.10 ± 0.0813.97 ± 0.1414.23 ± 0.09−0.26 ± 0.160.1205Surgery13.68 ± 0.0713.81 ± 0.1213.55 ± 0.060.25 ± 0.140.0660Obstetrics & Gynecology14.11 ± 0.0813.83 ± 0.1114.39 ± 0.11−0.56 ± 0.150.0003 Clerkship rotation bias estimates obtained through generalized linear models. Each dependent variable (Total average, Clinical competence and Professionalism) was evaluated separately. Total Average scores ranged from 0–10.75, Clinical Competence scores ranged from 0–10, and Professionalism scores ranged from 0–16. Errors presented are standard errors of the mean (SEM). The difference corresponds to the male minus female evaluator scores Discussion The objective of this study was to examine whether preceptor and student gender influence the numerical grading of third-year medical students across required core clerkships. Building on prior work demonstrating that preceptor gender can influence both narrative feedback and numerical assessment, this study examined these patterns across all required core clerkships using a large, multi-site dataset. Preceptor gender demonstrated a stronger association with quantitative grading outcomes than student gender, with female preceptors assigning lower overall and clinical competence scores than male preceptors. In contrast, student gender demonstrated only modest effects, and no interaction between student and evaluator gender was identified, suggesting that preceptor gender did not differentially influence how male and female students were graded. These findings help clarify mixed results reported in earlier studies and underscore the importance of considering evaluator characteristics when interpreting clerkship performance data. Although the observed differences in evaluation scores were relatively small, even modest systematic differences may become meaningful in high-stakes assessment environments where cumulative evaluations contribute to clerkship grades, academic standing, and residency selection. These findings support continued efforts to improve consistency and equity in clinical evaluation practices. Prior studies examining evaluator gender have produced inconsistent findings. One study found that female preceptors tended to assign slightly higher grades to female students, whereas male preceptors demonstrated minimal differences in grading based on student gender [7]. In contrast, another study reported that male preceptors assigned higher grades to female students, while female preceptors assigned lower grades to male students [8]. A separate study using natural language processing to analyze narrative feedback found no significant differences in tone or word use between male and female students; however, students were more likely to receive honors when evaluated by a preceptor of the same gender [6]. The findings further support the role of evaluator characteristics in quantitative grading outcomes across multiple clerkships. Clerkship rotation also demonstrated a strong independent association with evaluation scores across all domains, highlighting systematic differences in grading cultures across specialties. Prior studies have shown substantial variability in clerkship grading both across and within institutions, with wide differences in grade distribution observed between rotations even within the same medical school [9]. More recent work has similarly demonstrated variability in grading patterns across core clerkships such as surgery and internal medicine, further supporting the presence of inconsistent evaluation practices across specialties [10]. A similar pattern was observed in the present study, in which Family Medicine and Pediatrics demonstrated higher adjusted scores, while Surgery and Internal Medicine demonstrated lower scores across evaluation domains. These patterns may, in part, reflect institutional or structural factors rather than individual preceptor bias alone. Female preceptors may apply more stringent performance standards due to differences in evaluative style, institutional expectations, or broader cultural pressures within academic medicine. These factors may contribute to more conservative grading practices independent of student gender. Although clerkship directors often emphasize feedback based on observable behavior, substantial variability remains in how preceptors interpret student performance [11]. Even in standardized assessments such as Objective Structured Clinical Examinations (OSCEs), grader variability persists, highlighting the influence of subjective judgment in clinical evaluations [12]. Prior studies have also reported that female preceptors consistently assign lower grades regardless of student gender, which is consistent with the present findings [13]. Additionally, the observation that female students received higher professionalism scores parallels earlier literature demonstrating that female trainees are more frequently described as “compassionate” or “enthusiastic” traits often associated with professionalism domains [14]. Similar patterns have also been identified in surgical clerkship evaluations, where women were more frequently described using terms emphasizing effort and interpersonal characteristics, while men were commonly praised for innate ability and leadership qualities [14]. Such distinctions in assessment may influence residency selection processes and future professional trajectories. Continued faculty development and greater standardization of evaluation criteria may help reduce variability and support evaluations based on consistent, observable measures of clinical performance. Limitations This study has several limitations. It was conducted at a single osteopathic medical school, which may limit the generalizability of the findings to other training environments. Additionally, female preceptors represented a smaller proportion of the evaluator cohort, which may have influenced comparisons between evaluator groups. Gender classifications were investigator-assigned rather than based on self-identified gender data, and the dataset did not include nonbinary or gender-diverse categories, which may limit the inclusiveness and accuracy of the analyses. This analysis also did not account for factors such as preceptor training, years of experience, or differences in clerkship culture, all of which may influence grading practices. In addition, the study evaluated only numerical scores and did not include narrative comments, which often provide important contextual information. These limitations highlight several areas for future investigation, including whether preceptor training could reduce grading differences and whether domains such as professionalism are more susceptible to subjective interpretation. Conclusion This study found that preceptor and student gender both influenced clerkship evaluations. Female preceptors assigned lower ratings than male preceptors in overall scores and clinical competence. Female students received slightly higher overall scores and higher professionalism ratings. No interaction effects were observed in the dataset, indicating that male and female students were graded similarly regardless of preceptor gender. These findings highlight how both preceptor and student characteristics can shape outcomes in the evaluation process. Strengthening preceptor training and simplifying grading criteria may help reduce variability and promote evaluations that more accurately reflect clinical performance. Future studies incorporating narrative comments alongside numerical ratings may provide a more comprehensive understanding of where bias occurs and how it can be addressed. Authors’ contributions EK: Conceptualization, Methodology, Investigation, Writing – Original Draft, Writing – Review & Editing. IZ: Methodology, Formal analysis, Data Curation, Visualization, Writing – Review & Editing. AB: Investigation, Supervision, Writing – Original Draft, Writing – Review & Editing. TJ: Investigation, Writing – Original Draft. TH: Conceptualization, Investigation, Resources, Supervision, Writing – Original Draft, Writing – Review & Editing. Funding This study did not receive special funding. Data availability The datasets generated and/or analyzed during the current study are not publicly available due to institutional and IRB-related privacy restrictions. De-identified data may be available from the corresponding author upon reasonable request. Declarations Ethics and consent to participate This study was reviewed and approved by the Rocky Vista University Institutional Review Board (IRB# 2019-0081). All methods were carried out in accordance with relevant guidelines and regulations, including the Declaration of Helsinki. As this study involved a retrospective analysis of de-identified educational data, the requirement for informed consent was waived by the Rocky Vista University Institutional Review Board. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The datasets generated and/or analyzed during the current study are not publicly available due to institutional and IRB-related privacy restrictions. De-identified data may be available from the corresponding author upon reasonable request.