Faculty Scoring of General Surgery Residency Interviewees: A Comparison of In-Person and Virtual Interview Formats

•Virtual interviews are more efficient and less costly than In-Person.•Overall interview ratings of general surgery applicants via In-Person and Virtual formats are similar.•However, virtual interviews may have more difficulty differentiating non-cognitive attributes and fit for the program. With new rules regarding social distancing and non-essential travel bans, we sought to determine if faculty scoring of general surgery applicants would differ between the in-person interview (IPI) and virtual interview (VI) platforms. A single institution, retrospective review comparing faculty evaluation scores of applicant interviewees in the 2019 and 2020 MATCH® application cycles (IPIs) and the 2021 and 2022 application cycle (VIs) was conducted. Faculty scored applicants using a 5-point Likert scale in 7 areas of assessment and assigned each student to 1 of 4 tiers (tier 1 highest). A composite score for the 7 assessments (maximum score 35) was calculated. Mean and composite scores and tiers were compared between VI and IPI cycles and adjusted for within-interviewer correlations. The variance of the 2 groups were also compared. Harbor-UCLA Medical Center, an academic, tertiary care hospital. General Surgery applicants for the 2019 to 2022 MATCH® application cycles. Four hundred forty-one faculty IPI ratings of General Surgery applicants were compared to 531VI ratings. No difference in mean composite scores, individual assessments, or tier ranking. Less variance was identified in the VI group for academic credentials (0.6 vs 0.6, p = 0.01), strength of letters (0.7 vs 0.4, p = 0.005), communication skills (0.4 vs 0.6, p = 0.01), personal qualities (0.2 vs 0.5, p = 0.02), overall sense of fit for program (0.6 vs 0.9, p = 0.01), and tier ranking (0.3 vs 0.4, p = 0.004). Faculty ratings of General Surgery applicants in the VI format appear to be similar to IPI. However, faculty ratings of VI applicants demonstrated less variability in scores in most assessments. This finding is potentially concerning, as it may suggest an inability of VI to detect subtle differences between applicants as comparted to IPI.