Extensions of Mantel-Haenszel for Multilevel DIF (Differential Item Functioning) Detection

• In Event: Explorations in Differential Item Functioning

Abstract

This simulation study examines the performance of uniform differential item functioning (DIF) detection techniques in the presence of a multilevel structure that often underlies testing data. Analyses were conducted using Mantel-Haenszel (MH) procedure, a popular, flexible, and effective tool for DIF detection. Data were simulated in a multilevel framework, corresponding to examinees nested in schools, for example. Analysis employed the standard MH to DIF detection and a few multilevel extensions of MH that accounted for the multilevel data. Results highlight the differences in DIF detection when the analytic strategy matches the data structure.

Authors

• Brian F. French, Washington State University

• William Holmes Finch, Ball State University