BACKGROUND The continuity of undergraduate to postgraduate training suggests that performance in medical school should predict performance later in residency. PURPOSE The goal is to determine whether undergraduate performance is predictive of postgraduate performance. METHODS Residency program directors assessed the performance of medical school graduates (Classes 2004-2006) at the end of the 1st postgraduate year. Measures of undergraduate performance were retrieved including grade point averages, clerkship in-training evaluation reports, and the total score on the Medical Council of Canada Part 1 exam. RESULTS Complete data were available for 242 (81.5%) graduates. Postgraduate performance consisted of two reliable factors (clinical acumen and human sensitivity) that explained 78% of the variance. Correlations between undergraduate and the two postgraduate measures were low (.03-.31). CONCLUSIONS Measures of undergraduate performance appear to be poor predictors of performance in residency that consisted of two primary dimensions (clinical acumen and human sensitivity).