Predictive validity of the MCAT for students with two sets of scores

Addressed in the study reported here is the question of which set of scores for those students who retake the Medical College Admission Test (MCAT) yields a better predictive validity. The sample was comprised of 304 students who retook the MCAT prior to entering Jefferson Medical College between 1978 and 1981. Five sets of MCAT scores were considered as predictors in the study: earlier, later, higher, and lower sets of MCAT scores and the average of the earlier and later scores for each MCAT subtest. Twenty-five criteria were used, including grades earned in the freshman and sophomore years and scores on the subtests of Part I and Part II of the examinations of the National Board of Medical Examiners. Correlational techniques, such as bivariate and multiple correlation analyses and canonical correlation followed by redundancy analysis were utilized. The magnitude of redundancy indices indicated that the set of MCAT scores in which the earlier and later scores were averaged was the best predictor, followed by the earlier, lower, higher, and later sets of MCAT scores. The implications of these findings for the admissions process and for validity studies are discussed.