tus, science GPA, and MCAT scores in physical sciences and biological sciences explained only 29.1 percent of the variation in USMLE Step 1 test scores (Basco et al., 2002). Most variation in standardized test scores is not predicted by either demographic information or traditional measures of academic performance; grades in the clinical years are least well predicted by either MCATs or GPAs.

Medical schools also consider potential variables that suggest that a student may encounter academic difficulty—traditionally defined as withdrawal from school, a nonmedical leave of absence, dismissal, or delay of graduation date. Most studies have found that an increased risk of encountering academic difficulty is associated with low MCAT scores (particularly biological sciences scores), low science GPAs, low selectivity of the undergraduate institution, female gender, being an URM member, or older age. The majority of students who experience academic difficulty, however, eventually graduate from medical school, and the risk and timing of these episodes has been found to vary among the different groups of students studied (Huff and Fang, 1999).

Standardized test results and grades are thus useful but not exceptional or unique predictors of medical school performance. They are designed and validated by their ability to predict future test scores. GPAs and MCAT scores are not useful in predicting clinical performance, even when adjusting for the students’ undergraduate institution (Silver and Hodgson, 1997). Standardized tests measure already developed skills but not the mastery of a particular curriculum or a student’s innate ability. Experiences that are closely tied to an individual’s racial and ethnic identity can lower the results of standardized tests independent of socioeconomic status. This outcome is more likely to occur for a minority or other student whose abilities are negatively stereotyped by society; this is particularly true for the student who is deeply invested in achieving good results on a test. This “stereotype threat” may interfere with test performance for any student (or group of students) for whom abilities are negatively stereotyped in the larger society (Steele and Aronson, 1995). Admissions committees that place the greatest weight on standardized test scores limit the opportunities of minority students to participate in the medical profession.

Personal Characteristics

Academic success during the first 2 years of medical school is not, in and of itself, predictive of success in meeting the clinical training and patient care requirements of the third and fourth years of medical school and future practice. To be successful in clinical settings, students must demonstrate an ability to apply what they have learned and communicate and interact effectively with patients, faculty and staff, peers, and others. Good



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement