Based on an evaluation of the results of 119 In-Training tests taken by 60 residents from 1975 to 1981, guidelines were developed for using the test in surgery training programs. Regression analysis revealed a high correlation (r = 0.79, p less than 0.001) between scores on the In-Training test and the American Board of Surgery Qualifying Examinations. At the 95 percent confidence level, a PGY V resident scoring higher than the 25th percentile on the final In-Training Examination will pass the Qualifying Examination. Residents who do not continue the general surgery residency (dropouts, those with a subspeciality year on general surgery, those not selected for further training) score significantly lower (p less than 0.002) than general surgery residents. This selection process tends to lower (2.5 percentile points per year) the percentile scores of general surgery residents in later years of training. Participation in key teaching conferences has a positive correlation, although it is disappointingly low, with In-Training results. Objective guidelines can assist the teaching staff when they consider residents for selection or retention in the training program.