5 Differential Item Functioning and Item Bias

Abstract

Publisher Summary This chapter presents a description of many of the commonly employed methods in the detection of item bias. Because much of the statistical detection of item bias makes the use of differential item functioning (DIF) procedures, the majority of this chapter focuses on the description of statistical methods for the analysis of DIF. DIF detection procedures for dichotomous and polytomous items are presented in the chapter, along with the methods for the categorization of DIF effect in dichotomous items. It also presents several recent innovations in DIF detection, including Bayesian applications the detection of differential test functioning, and studies examining sources or explanations of DIF. While much of this chapter focuses on the statistical approaches to measuring DIF, conducting a comprehensive DIF analysis requires series of steps aimed at measuring DIF and ensuring that the obtained DIF statistics are interpreted appropriately. The chapter also outlines a set of six steps that practitioners can use in conducting DIF analyses. The steps are demonstrated using a real dataset.