The Rasch measurement approach to differential item functioning (DIF) analysis in language assessment research

Efficient use of resources often entails the use of the same test across different groups. For example, it is quite common to use one test to measure the language proficiency of students across different age groups or on groups of students with different academic backgrounds. A potential problem of this situation could be test results that are misleading, as some items inadvertently favor one group over another. Differential item functioning (DIF) is the term used when such problems exists in a test at the item (question/prompt) level. This chapter provides readers with an introduction to a Rasch measurement approach to detecting DIF. It also provides readers with an overview of DIF detection methods, a summary of DIF studies in language testing, and an example of a five-stage method to conduct a Rasch measurement approach to detect DIF in a diagnostic English listening test using Winsteps software.