Pediatric, Surgery, and Internal Medicine Program Director Interpretations of Letters of Recommendation

Abstract

Purpose: Literature describing program director (PD) perceptions of letters of recommendation (LORs) and "code" used by letter writers is limited. In 2016, a survey instrument was distributed nationally to pediatric PDs asking them to rate their interpretations of components of LORs. The results confirmed that letter phrases convey code, but these results were not known to be generalizable outside of pediatrics. The purpose of this study was to expand the survey to surgery and internal medicine (IM) PDs looking for areas of agreement or variation between the 3 specialties.

Method: The survey was sent nationally to surgery and IM PDs asking them to rate LORs in 3 areas on a 5-point Likert scale: 14 commonly used phrases, 13 letter features, and 10 applicant abilities. The LOR phrases were grouped using principal component analysis (PCA). Mean scores of components were analyzed with repeated-measures analysis of variance.

Results: Response rates: pediatrics 43% (486 of 1079), surgery 55% (151 of 277), and IM 42% (170 of 408). PCA generated groups of positive, neutral, and negative phrases with moderate to strong correlation with each other for all 3 specialties. There were significant differences between the mean Likert scores of the positive, neutral, and negative groups of phrases for all 3 specialties (all P < .001). "Showed improvement" was rated the most negative phrase by all 3 specialties.

Conclusions: Key elements of LORs include distinct phrases depicting different degrees of endorsement of candidates. Pediatric, surgery, and IM PDs interpret letter components differently.