Displaying faculty grade averages during 3(rd) year medical student clerkship evaluations: effects upon grade inflation.

摘要

BACKGROUND: Medical educators recognize grade inflation as a significant problem with multi-factorial etiologies beginning as early as junior medical student clerkship rotations. Colletti et al reported faculty educators' inability or unwillingness to provide medical student trainees with unfavorable face-to-face evaluations as a source of grade inflation. The effects of un-countered grade inflation are significant, rendering academic deans and advisors less able to provide useful comparative evaluations and recommendations for matriculating students, and in one study (Speer et al), leaving nearly half of the faculty in an internal medicine clerkship concerned about "social-promotion" of failing students due to uncontrolled grade inflation. Little literature specifically addresses techniques to minimize grade inflation. METHODS: Using our previously reported 360-degree web-based competency evaluation tool for 3(rd) year medical student clerkships, evaluation data were collected assessing nine evaluation items based upon the six ACGME "core competencies". During part one of the study, evaluations were completed with faculty evaluators blinded to their averaged evaluation data. During part two of the study, real-time faculty mean and standard error evaluation data were prominently displayed on the evaluation instrument during faculty evaluations. We hypothesized that passive faculty notification of evaluation data would alter evaluations toward a more centrist mean, based upon changes in clinician ordering patterns observed when cost data were displayed. RESULTS: Prior to and during Part One of the study, mean evaluation scores were significantly greater than the anticipated mean (2.5) for all items evaluated. 2043 individual evaluation scores were acquired, equally allocated between pre- and post-notification periods of the study. Mean evaluation scores were significantly greater than the anticipated mean (2.5) for all items evaluated. Data were evaluated using parametric (t-test) and nonparametric(Wilcoxon signed ranks test) statistical analyses comparing blinded and passive faculty notification evaluation data. Mean evaluation scores were noted to increase slightly for each of the nine evaluation items after commencement of passive notification, with one item achieving and three items approaching statistical significance. CONCLUSIONS: These findings appear to contradict our hypothesis that passive notification of faculty evaluation averages decrease evaluation means or improve grade inflation. Increasingly large datasets, obtained using on-line evaluation tools, will likely achieve statistical significance demonstrating that faculty should not be shown real-time grade averages during evaluation. We believe that these data support the need to develop alternate methods of managing grade inflation, including faculty education or parametric grade re-calculation as tools to counter grade inflation.