UCSF

Purpose  Prior studies have revealed grading discrepancies in evaluation of personal statements and letters of recommendation based on candidate's race and gender. Fatigue and the end-of-day phenomenon can negatively impact task performance but have not been studied in the residency selection process. Our primary objective is to determine whether factors related to interview time and day as well as candidate's and interviewer's gender have a significant effect on residency interview scores. Methods  Seven years of ophthalmology residency candidate evaluation scores from 2013 to 2019 were collected at a single academic institution, standardized by interviewer into relative percentiles (0-100 point grading scale), and grouped into the following categories for comparisons: different interview days (Day 1 vs. Day 2), morning versus afternoon (AM vs. PM), interview session (Day 1 AM/PM vs. Day 2 AM/PM), before and after breaks (morning break, lunch break, and afternoon break), residency candidate's gender, and interviewer's gender. Results  Candidates in the morning sessions were found to have higher scores than afternoon sessions (52.75 vs. 49.28, p  < 0.001). Interview scores in the early morning, late morning, and early afternoon were higher than late afternoon scores (54.47, 53.01, 52.15 vs. 46.74, p  < 0.001). Across all interview years, there were no differences in scores received before and after morning breaks (51.71 vs. 52.83, p  = 0.49), lunch breaks (53.01 vs. 52.15, p  = 0.58), and afternoon breaks (50.35 vs. 48.30, p  = 0.21). No differences were found in scores received by female versus male candidates (51.55 vs. 50.49, p  = 0.21) or scores given by female versus male interviewers (51.31 vs. 50.84, p  = 0.58). Conclusion  Afternoon residency candidate interview scores, especially late afternoon, were significantly lower than morning scores, suggesting the need to further study the effects of interviewer's fatigue in the residency interview process. The interview day, presence of break times, candidate's gender, and interviewer's gender had no significant effects on interview score.