Can admissions interviews predict performance in residency?

OBJECTIVE The authors aim to determine whether admission interviews predict performance in residency. METHODS The authors determined whether interview and other admission data were correlated with performance during postgraduate years 2-4 and with remaining in the residency in 544 residents enrolled in a single psychiatry residency program between 1963 and 2004. RESULTS Considered together, admissions data predicted 13% of the variance in performance ratings in postgraduate year 2 (PGY-2) and 5% in PGY-4. Interview scores were moderately correlated with performance ratings in PGY-2, modestly correlated with performance in PGY-3, and not correlated with performance ratings in PGY-4. Letters of reference were moderately correlated with performance ratings in PGY-2 and modestly correlated with performance in PGY-3 and PGY-4. In PGY-2, interview scores differentiated between the top quartile of performance and the other three quartiles, while letters of reference differentiated performance in the top and bottom quartiles from the middle quartiles. Numerical differences among groups were not great enough to be practically useful, and no variables predicted which residents would leave the program before completing it. CONCLUSION As they are currently conducted, application interviews do not have sufficient power to predict performance during residency. Letters of reference may be useful to the extent that they reflect personal experience with the applicant, but differences in ratings of these letters are not great enough to base admission decisions on them. As it is currently performed, the interview process may be more useful as a means of interesting applicants in the program than of evaluating their potential for success in the residency.