Evaluating GPAs and MCAT scores as predictors of NBME I and clerkship performances based on students' data from one undergraduate institution

Acad Med. 1997 May;72(5):394-6. doi: 10.1097/00001888-199705000-00022.


Purpose: To examine the relationships among undergraduate student characteristics, standardized indicators of medical school academic performance, and clinical performances in an institution where the undergraduate grading system was not a confounding factor.

Method: In 1993 data were collected retrospectively for the classes of 1990-1993 in the University of California, Riverside (UCR)/University of California, Los Angeles, Biomedical Sciences Program, in which all the students completed their undergraduate studies at UCR. Data were collected on the students' demographic characteristics, undergraduate grade-point averages (UGPAs), Medical College Admission Test (MCAT) scores, National Board of Medical Examiners Part I (NBME I) scores, and clinical performances as measured by core clerkship grades. Regression analyses were used to evaluate (1) the relationship between clinical performances and data available at admission and (2) the relationship between NBME I scores and the same variables. The top and bottom 25% of each class were identified by UGPAs and compared by their NBME I scores and clinical performances to determine whether differences noted at admission continued to separate the students. Differences were tested using independent two-tailed t-tests.

Results: Of the 92 students, data were available for 88 (96%); 39% were female and 38% were foreign-born. The sample was fairly heterogeneous in terms of UGPAs, MCAT scores, clinical performance grades, and NBME I scores. Results of the first regression analysis indicated that mean clinical performance was not related to any of the undergraduate predictors of performance or to the students' demographics. The second regression analysis indicated that MCAT scores and cumulative science UGPAs were related to performance on the NBME I, as were country of origin and sex. The students in the top and bottom 25% differed significantly in their scores on the NBME I, but not in their clinical performances.

Conclusion: Although UGPAs and MCAT scores are good indicators of NBME I performance, they are still not useful in predicting clinical performance, even when the students' data are taken from the same undergraduate institution.

MeSH terms

  • California
  • Clinical Clerkship*
  • Educational Measurement*
  • Female
  • Humans
  • Male
  • Predictive Value of Tests
  • Retrospective Studies
  • Schools, Medical