The Development of a Multidimensional Graded Testlet Response Model

• In Event: AERA Roundtable Session Friday 8:00 am Four Seasons Ballroom 4
  In Roundtable Session: Advancing Psychometric Modeling: Innovative Techniques for Educational Assessment (Table 8)

Abstract

This study develops the multidimensional graded testlet response model (MG-TRM), enhancing the multidimensional graded response model (MGRM) framework for polytomous scoring in target multidimensional abilities and multidimensional testlets within items. MG-TRM’s parameter estimation accuracy is evaluated via Monte Carlo simulation, considering local item dependence (LID) and test length. Results show MG-TRM significantly outperforms MGRM in estimating item and person parameters, with accuracy declining with higher LID but improving with longer tests. MG-TRM, thus, offers superior accuracy in analyzing tests with multidimensional abilities and testlet effects, broadening model options for test analysis with practical implications.

Authors

• Lingling Xu, Hangzhou Normal University

• Zhehan Jiang, Peking University

• Jinying Ouyang, Peking University