The present study develops a rating scale method for evaluating operative skills, assesses the predictive utility of neuropsychologic tests of nonverbal cognitive and psychomotor abilities in accounting for individual differences in surgical skills, and compares the efficiency of these measures with those of traditional residency selection criteria. According to a multifactorial design, 120 general surgery residents were tested with a neuropsychologic test battery and then rated by attending surgeons on surgical skills exhibited during the course of 1445 surgical procedures. Analysis of the neuropsychologic battery resulted in three factors (complex visuo-spatial organization, stress tolerance, psychomotor abilities) that were statistically unrelated to traditional measures such as Medical College Admission Test and National Board scores. Multiple regression analyses indicated that academic predictors, taken alone, either do not correlate (National Board scores) or correlate negatively (Medical College Admission Test scores) with the surgery ratings. Conversely, neuropsychologic test scores show significant positive correlation (r = 0.68) with the ratings. When both sets of predictor variables are combined, a multiple regression coefficient of 0.80 is found with the ratings, with more than two thirds of the predictive power attributable to the neuropsychologic test scores. These tests may provide a useful addition to traditional methods of predicting operative skills.