Education Research: Linguistic Analysis of Racial and Gender Biases in Letters of Recommendation of Neurology Residency Applicants
Year: 2026
Type: article
Abstract: Background and Objectives: Letters of recommendation (LORs) are a key component of the residency selection process, particularly as objective metrics become less available. Previous studies across specialties have identified implicit gender and racial biases in narrative LORs (NLOR), potentially influencing perceptions of applicants. This study assessed whether such biases are present in NLORs for Neurology residency applicants. Methods: We conducted a retrospective analysis of NLORs written for all American (US) medical school graduates applying to a single academic Neurology residency program during the 2021-2022 application cycle. Letters were analyzed using Linguistic Inquiry and Word Count (LIWC) software to evaluate structural, cognitive, and emotional language features. Custom dictionaries assessed applicant-related language (e.g., ability, doubt-raising, and gendered terms). Multivariable analyses were performed to examine variations in language use by applicant gender (man vs woman) and race (majority vs underrepresented in medicine [URiM]) across predefined linguistic categories. Analyses were further stratified by letter-writer gender, academic rank, and geographic region. Results: = 0.002). Secondary subgroup and regional analyses did not reveal any significant findings. Discussion: In this linguistic analysis of Neurology NLORs, we did not identify broad or systematic bias in letters written by adult neurologists for Neurology applicants. Our findings suggest that the continued use of NLORs for Neurology residency applicants is appropriate. Larger, multi-institutional studies are warranted to validate these findings and clarify the potential biasing effect of gendered language in LORs.
Source: Neurology Education
Language: English
Subfield: Gender Studies
Field: Social Sciences
Domain: Social Sciences
SDG: Reduced inequalities
Open Access status: hybrid
OpenAlex ID: W7154305163