Differential Item Functioning: Implications for Test Validation

This paper attempts to recapitulate the concept of validity, namely construct validity (i.e., its definition and its approaches and role in language testing and assessment). Validation process is then elaborated on and proved to be integral enterprise in the process of making tests, namely English language proficiency tests. Then come the related concept of test fairness and test bias and its sources (e.g., gender, field of study, age, nationality and L1, background knowledge, etc) and contributions and threads to the validity of tests in general and in high-stakes tests of English language proficiency in particular. Moreover, in the present study, different approaches to investigate the validity of tests will be reviewed. Differential Item Functioning (DIF), among the other methods to investigate the validity of tests is also explained along with the description and explanation of its different detection methods and approaches mentioning their advantages and disadvantages to conclude that logistic regression (LR) is among the best methods till now.