Evaluating Performance of Mantel Test and Generalized Mantel-Haenszel With Model Misspecification in Causal Estimation

• In Event: AERA Roundtable Session Sunday 11:40 am Four Seasons Ballroom 1
  In Roundtable Session: Advances in Differential Item Functioning (DIF) Detection: Methods, Applications, and Implications (Table 12)

Abstract

The application of propensity score matching (PSM) to the Mantel test and Generalized Mantel-Haenszel (GMH) can reduce confounding impacts on polytomous differential item functioning (DIF), enhancing interpretability and expert agreement. However, correct model specification in PSM is challenging, leading to confusion about its appropriate use in polytomous DIF detection. This study investigates the performance of the Mantel test and GMH with correct and misspecified PSM models through a simulation study. Results indicate promising performance of GMH DIF with and without correct model specification. In sum, this study demonstrates the feasibility of using PSM for polytomous DIF, improving DIF interpretability with causal-based evidence.

Authors

• Ziying Li, Independent Researcher

• Corinne Huggins-Manley, University of Florida

• Walter L. Leite, University of Florida

• M. David Miller, University of Florida

• Nikolay A. Bliznyuk, University of Florida