Sahin F, Clauser J. — American Board of Internal Medicine
Presented: National Council on Measurement in Education Annual Meeting, April 2016
Abstract: Key validation procedures typically rely on professional judgement to identify potentially problematic items. Unfortunately, lack of standardized flagging criteria can introduce bias in examinee scores. This study demonstrates the use of logistic regression to mimic expert judgment and automatically flag problematic items. The final model properly identified 96% of items.
For more information about this presentation, please contact Research@abim.org.