Poster #34 - Detecting Differential Item Functioning Items Using Bayesian Multilevel Item Response Theory With Covariate Model

• In Event: 21.063 - MTCC Poster Session Three
  In Poster Session: 21.063-3 - Differential Item Functioning and Measurement Invariance

Abstract

Recently, the IRT with covariate (IRT-C) approach was developed for DIF detection (Tay, Huang, & Vermunt, 2016). However, many large-scale educational assessment data sets commonly have multilevel structures. Because ignoring the multilevel structure in data analyses may compromise the parameter estimation, applying the single-level IRT-C to large-scale educational assessments with multilevel structures may result in erroneous DIF detection. Therefore, the purpose of this study is to expand the IRT-C model in DIF detection to account for the multilevel structure of Programme for International Student Assessment (PISA) 2006 Math data using a Bayesian approach. The results of the study demonstrated the utility of the multilevel IRT-C models to detect DIF items for large-scale assessment data, accounting for cross-county differences.

Authors

• Hwanggyu Lim, University of Massachusetts

• Yooyoung Park, University of Massachusetts - Amherst