Residency selection: should interviewers be given applicants' board scores?

The aim of this study was to determine the influence of advance knowledge of board scores on interviewers' assessments of residency applicants.

During a 2-year period we prospectively evaluated our residency selection process. In year 1 interviewers were provided with each candidate's entire application, whereas in year 2 the United States Medical Licensing Examination scores were not included. In each year interviewers were asked to provide numerical assessments of the applicants solely on the basis of their own impressions of the interviews. Analysis was performed only for evaluations by interviewers who participated during both study periods under review. Interview scores were compared with United States Medical Licensing Examination part I scores within each year by means of a scatter plot and correlation coefficients.

Applicant demographic characteristics were similar during years 1 and 2. Interview scores did not differ between year 1 (4.2 +/- 0.1) and year 2 (4.3 +/- 0.1; P > .05). During year 1 interview and board scores were significantly correlated (correlation coefficient, 0.64; slope of best-fit line, 13.9), whereas there was a negative correlation in year 2 (correlation coefficient, -0.06; slope, -1.3).

When they are available to interviewers, markers of academic achievement such as United States Medical Licensing Examination scores may bias the interview evaluation. The interview process when conducted in this manner may simply be a validation process for candidates already judged on the basis of the application alone. Knowledge of United States Medical Licensing Examination scores by the interviewers may therefore negate the interview as an independent means of evaluating candidates.