Gender and Racial Bias in Letters of Recommendation for Orthopedic Surgery Residency Positions

The primary purpose of this study was to retrospectively analyze letters of recommendation written for medical students applying to orthopedic residency for implicit race and gender bias. The secondary purpose was to determine if the presence of bias was influenced by the gender of the letter writer.

This was a retrospective institutional review board (IRB) approved study. All letters of recommendation received in the years 2016 to 2018 were deidentified and analyzed using the Linguistics Inquiry and Word Count (LIWC) 2015 software. Independent variables in our analysis were applicant gender and applicant race. Dependent study variables included the summary and characteristic variables of a letter of recommendation, which are word count, analytic, clout, authenticity, tone, and positive and negative emotion word categories. Separate analyses were completed by gender of the letter writer as well.

Institution: Temple University Hospital, Philadelphia, Pennsylvania.

Medical students applying to Temple University Hospital Orthopaedic Surgery Program from 2016 to 2018. A total of 2113 applicants were included in the study.

Female, Asian and underrepresented minority applicants' letters were more likely to have a higher word count. In our subset analysis by gender of letter writer, when the letter writer was male, Asian applicants' letters were more likely to convey analytical thinking and authenticity. When the letter writer was male, male applicants scored higher for authenticity. Letters written by female attendings demonstrated no significant difference for male or female applicants in terms of composite variables or word categories.

Our study shows that letters of recommendation for orthopedic surgery residency positions are likely to contain some degree of bias. Further studies are required to fully characterize the degree and magnitude of bias in letters of recommendation and whether the findings of our study are significant enough to contribute to the difference in socioeconomic demographics between orthopedic residents and society at large.