Robert E S Bowen1, Wendy J Grant1, Kimberly D Schenarts2. 1. Department of Surgery, University of Nebraska Medical Center, 983280 Nebraska Medical Center, Omaha, NE 68198-3280, USA. 2. Department of Surgery, University of Nebraska Medical Center, 983280 Nebraska Medical Center, Omaha, NE 68198-3280, USA. Electronic address: kim.schenarts@unmc.edu.
Abstract
BACKGROUND: This study examines grading component distributions to determine whether alterations in clinical grade determination reduce skew and improve predictive capability of the clinical evaluation. METHODS: Rotation evaluations, examination scores, and final grades were collected for third-year medical students over a 2-year period. Conditional logistic regression and ordinary least squares regression models were run using SAS 9.3. RESULTS: Conditional logistic regression demonstrated significant association between global clinical score and final grade and between average clinical evaluation score and final grade. Inclusion of shelf score into either model demonstrated increase in overall final grade. CONCLUSIONS: Regressions using global and average clinical evaluation score indicate that average score is a better fit for a norm-based grading system. Arguably, the Shelf measures clinical knowledge more objectively than clinical evaluation, but both were significant. Clinical evaluation is prone to inflation because of its subjective nature; conceivably, inflation leads to the decreased correlation with shelf score.
BACKGROUND: This study examines grading component distributions to determine whether alterations in clinical grade determination reduce skew and improve predictive capability of the clinical evaluation. METHODS: Rotation evaluations, examination scores, and final grades were collected for third-year medical students over a 2-year period. Conditional logistic regression and ordinary least squares regression models were run using SAS 9.3. RESULTS: Conditional logistic regression demonstrated significant association between global clinical score and final grade and between average clinical evaluation score and final grade. Inclusion of shelf score into either model demonstrated increase in overall final grade. CONCLUSIONS: Regressions using global and average clinical evaluation score indicate that average score is a better fit for a norm-based grading system. Arguably, the Shelf measures clinical knowledge more objectively than clinical evaluation, but both were significant. Clinical evaluation is prone to inflation because of its subjective nature; conceivably, inflation leads to the decreased correlation with shelf score.