Predictive validity of the American Board of Family Practice In-Training Examination

Most research into the validity of graduate medical education in-training examinations has focused on construct validity and concurrent validity issues. This study examined the predictive validity of the American Board of Family Practice In-Training Examination using multivariate analysis of all U.S. family practice residents who took the certification examination in either 1987 or 1988. Results of these two certification examinations were analyzed for the cohorts of physicians who had taken in-training examinations as first-, second-, and third-year residents. Multiple regression analysis showed that the composite score, and all but one part-score, of each in-training examination were independently predictive of performance on the certification examination. This study also found that the older residents did less well on the certification exam, and that men and women each did better on selected portions of the examination. The implications of these findings for in-training examinations in other specialties and for program directors and residents are discussed.