Multiple Group Cognitive Diagnosis Models and Their Applications in Detecting Differential Item Functioning

• In Event: 18.020 - Research on Cognitive Diagnostic Modeling

Abstract

Cognitive diagnosis models (CDMs) have gained increasing popularity recently. However, most existing CDMs assume that students are all from the same population. In this study, a multiple group cognitive diagnostic model is developed based on a general CDM framework – the generalized DINA model. The multiple group models enable researchers to model different populations simultaneously, showing a potential usefulness in many areas (e.g., differential item functioning (DIF), matrix-sampling designed assessment). Based on the proposed models, a DIF detection procedure using the likelihood ratio test is developed as well. Preliminary results using simulation studies show that the multiple group CDMs can be calibrated accurately and that the DIF detection procedure is promising in terms of the type I error and power rates.

Authors

• Wenchao Ma, Rutgers University - New Brunswick/Piscataway

• Ragip Terzi, Rutgers University - New Brunswick/Piscataway

• Soo Youn Lee, American Institutes for Research

• Jimmy de la Torre, The University of Hong Kong