Background: CARCS (computer-assisted resident candidate selection) is a database application developed in 1983 at the Department of Anesthesiology of the Medical University of South Carolina to deal with the greatly increased quantity of applicant information. This article relates a representative sample of CARCS data to the process of selecting residents in general.
Method and results: CARCS files were analyzed for 1985-86, 1986-87, 1990-91, and 1991-92, and data for each year were derived as simple averages and percentages for two groups: (1) the entire pool of residency applicants (approximately 200 per year) and (2) the eight residency candidates per year who actually matched with the program through the National Resident Matching Program, Analyses showed that the standardized test scores, grades, and class ranks of the matched candidates were not significantly higher than those of the applicants; however, the matched candidates' scores for letters of reference and for interviews were consistently higher than those for the applicant pool.
Conclusions: The results support the view of medical educators that the traditional academic criteria are not sufficiently predictive of clinical performance or interpersonal skills. Research relating residents' performances to personality, learning style, and other neuropsychological factors may provide needed alternatives to knowledge testing by developing combined cognitive-noncognitive profiles. The anesthesiology clerkship experience is now almost universal among applicants and could be structured to provide pertinent information about potential residents through direct observation as well as behavioral testing.