Gendered Language in Letters of Recommendation for Applicants to Pulmonary Critical Care Fellowships.

ABSTRACT

Background: Previous work has demonstrated letters of recommendation for women in academic medicine are shorter and emphasize communal traits over grindstone or agentic traits. Objective: To determine if there are sex-based differences in letters of recommendation written for applicants applying to pulmonary critical care medicine fellowships and if the sex of the letter writer impacts these differences. Methods: All fellowship applications submitted to a pulmonary critical care medicine fellowship program in 2020 were included in this study. The applicant demographics and self-reported accomplishments were extracted from their application. The sex of letter writers was identified through public online searches. Word count and language differences in the letters of recommendation were analyzed for each applicant using the Linguistic Inquiry and Word Count (LIWC2015) program. Multivariable linear regressions were performed controlling for applicant characteristics to identify if applicant sex was associated with total word counts and total agentic word counts. Results: Of the 529 complete applications, 2,024 letters of recommendation were reviewed. A majority of the applicants (70%, n = 370/530) and letter writers (75%, n = 1,515/2,024) were male. When adjusting for applicant demographic and accomplishments, female applicants had longer letters of recommendation (30.91 words longer, 95% confidence interval [CI], 1.53-60.29; P = 0.04) and more supportive letters (3.27 words longer, 95% CI, 1.59-4.95; P < 0.01) as compared with male applicants. Female letter writers wrote longer and more supportive letters than male letter writers, and this difference was greatest for female applicants. Conclusion: Female applicants received longer and more supportive letters of recommendation. Further work is needed to understand if this finding is the beginning of a change in the letters of recommendation for female applicants.