Validity of the MCAT in predicting performance in the first two years of medical school.

In this paper, the authors present the first systematic summary of predictive validity research on the new Medical College Admission Test (MCAT) since its introduction in 1977. Data are drawn primarily from the MCAT Interpretive Studies Program, a cooperative effort between the Association of American Medical Colleges and 30 of its member schools to conduct research that will both facilitate local use of the test scores and contribute to a national perspective on their value in medical school admissions. The results show that MCAT scores by themselves have significant predictive validity with respect to first- and second-year medical school course grades and National Board of Medical Examiners Part I examination scores and that they complement the predictive validity of undergraduate college grades. The MCAT Science Knowledge areas of assessment, particularly Biology and Chemistry, and the Science Problems subtest tend to have higher correlations than the Skills Analysis subtests with initial performance in medical school; however, the Skills Analysis: Reading subtest may retain its predictive value best over time. Correlation values are discussed in terms of methodological factors which constrain their size. They are also compared with those found for other professional and graduate school admission tests. Further directions for MCAT validity research are described.