Multiple Group Item Response Theory Applications Using Stata irt Package

The purpose of this paper is to briefly introduce two most common applications of multiple group item response theory (IRT) models, namely detecting differential item functioning (DIF) analysis and nonequivalent group score linking with a simultaneous calibration. We illustrate how to conduct those analyses using the Stata item response theory (irt) package. To demonstrate the applications, we use a publicly available item response data set from Programme of International Student Assessment 2000. After reviewing the basics of multiple group IRT, we demonstrate how to identify DIF items using likelihood ratio tests. The second example is included to demonstrate how a test-form linking can be achieved by utilizing the multiple group IRT analysis in the presence of anchor items.