Detection of Differential Item Functioning with Missing Values in the Grouping Variable

• In Event: Missing Data

Abstract

Licensure/certification exams that allow examinees to report their demographics voluntarily face missing values in the variables for manifest groups in DIF study. This research explores various approaches that can tackle this problem by comparing their performance in DIF detection in simulation. These methods are applied to a medical exam dataset.

Authors

• Isaac Li, American Nurses Credentialing Center

• chingwei David shin, American Board of Psychiatry and Neurology