Cross-Classified Testlet Modeling of Complex Item Clustering in Large-Scale Alternate Assessment

• In Event: Advances in Item Response Theory Models

Abstract

Previous studies have mainly focused on investigating one source of local item dependence (LID). However, in some cases, LID might be caused by two possible sources simultaneously, which is referred to as dual local item dependence (DLID) in the literature. This study proposes analyzing the large scale alternate assessment data from a cross-classified modeling perspective in order to account for the DLID issue caused by complex item clustering. A cross-classified model and another three competitive models are estimated in SAS software using the PROC GLIMMIX macro package. Model fit and model selection are evaluated based on multiple information-based fit indexes.

Authors

• Chao Xie, American Institutes for Research

• Ming Lei, American Institutes for Research