Evaluation of Gender Bias in Psychiatry Residency Letters of Recommendation

Applying to psychiatry residency has become increasingly competitive over the last decade. As such, increased weight is placed on each application element, including letters of recommendation (LOR). LOR has been shown to be sources of gender bias for applicants in other specialties. No studies currently exist that analyze gender bias in LOR for psychiatry applicants. This study aimed to evaluate the presence of gender bias in LOR for students applying to psychiatry.

LOR for residency applicants to a small to mid-sized university-based psychiatry program from the 2023-2024 residency match cycle was reviewed using linguistic text analysis. The software identified words in 96 thematic categories and word count. Text mining techniques were applied to determine word frequency for male and female letters. Logistic regression analyses were performed.

A total of 975 traditional LOR for male applicants and 764 LOR for female applicants was analyzed. Only the "female reference" and "male reference" categories demonstrated statistical significance. The word frequency analysis indicated no difference in the top ten words written in LOR for male and female applicants. There was no statistical difference in word count for male or female letters, and logistic regression with least absolute shrinkage and selection operator (LASSO) method failed to identify any categories that would be more likely to predict a female letter vs. a male letter.

A linguistic text analysis of LOR for those applying to a single psychiatry residency program in the U.S. found no meaningful evidence of gender bias.