Correlations of family medicine clerkship evaluations with scores on standard measures of academic achievement.

Abstract

BACKGROUND: The family medicine clerkship at the University of Arizona uses three methods to evaluate students. The purpose of this study is to 1) look for possible sources of bias in each method, 2) validate the evaluation methods by correlating with other measures of academic achievement, and 3) measure the amount of correlation between the three evaluation methods. METHODS: The three clerkship evaluation methods include a final exam, clinical evaluations, and problem-solving scores. The clerkship evaluations received by 482 students were correlated with undergraduate grade point averages, MCAT scores, National Board of Medical Examiners parts I and II subtests and total scores, gender, age, race, college major, and timing of the clerkship. RESULTS: Females and older students scored higher on problem solving and minority students scored lower on the final exam. Students in the second half of the year scored higher on the final exam. There were significant correlations found between the final exam score and all the scores on standard measures of academic performance except the MCAT quantitative subtest. Problem-solving scores correlated only with NBME part II. The clinical evaluations correlated with five NBME part I subtest scores, three NBME part II subtest scores, and both parts I and II total scores. All three evaluation methods correlated significantly with each other. CONCLUSIONS: These analyses highlight the need to examine all evaluation methods used in order to improve their validity and reliability and to find potential biases. In addition, more work is needed to document the predictive validity of clerkship evaluations by correlating them with future clinical performance.