There has been relatively little research on the relationship between the clinical performance of physicians and the criteria used by medical school admissions and promotions committees. The studies which have been done primarily relate clinical performance to cognitive criteria. They have usually found only negligible relationships. By considering both cognitive and noncognitive variables, several investigators recently have improved the prediction of clinical performance in medical school. The present authors attempt to determine the increase in predictive efficiency attained by adding noncognitive variables to cognitive variables in predicting clinical performance of residents. The results of this investigation indicate that a combination of cognitive and noncognitive predictor variables functions much better than any individual variable or even any specific class of variables in predicting the postgraduate clinical performance of physicians.