A Comparison of Item Parameter Standard Error Estimation Procedures for Unidimensional and Multidimensional Item Response Theory Modeling. (February 2014)

• Record Type: Journal Article
• Authors: Paek, Insu; Cai, Li
• Is Part Of: Educational and psychological measurement. Volume 74:Number 1(2014:Feb.)
• Page Start: 58
• Page End: 76
• Publication Date: 2014-02
• Subjects: item response theory (IRT) -- standard error estimation -- supplemented expectation maximization -- multilevel and multiple group item response theory -- bifactor model; Educational tests and measurements -- Periodicals; Psychological tests -- Periodicals; 151.205
• DOI: 10.1177/0013164413500277
• Languages: English
• ISSNs: 0013-1644

Abstract

The present study was motivated by the recognition that standard errors ( SE s) of item response theory (IRT) model parameters are often of immediate interest to practitioners and that there is currently a lack of comparative research on different SE (or error variance–covariance matrix) estimation procedures. The present study investigated item parameter SE s based on three error variance–covariance matrix estimation procedures for unidimensional and multidimensional IRT models: Fisher information, empirical cross-product, and supplemented expectation maximization. This study centers on the direct comparisons of SE s from different procedures and complements a recent study by Tian, Cai, Thissen, and Xin by providing insights and suggestions on the nature of the differences and similarities as well as on practical matters such as the computational cost. The simulation results show that all three procedures produced similar results with respect to bias in the SE estimates for most conditions. When the number of items is large and sample size is small, empirical cross-product, which was the most computationally efficient procedure, appeared to be affected most, producing slight upward bias.