gender differential item functioning analysis of the university of tehran english proficiency test

the university of tehran english proficiency test (utept) is a high-stakes entrance examination taken by more than 10,000 master’s degree holders annually. the examinees’ scores have a significant influence on the final decisions concerning admission to the university of tehran ph.d. programs. as a test validation investigation, the present study, which is a bias detection research in nature, utilized multistep logistic regression (lr) procedure to examine the presence of gender differential item functioning (dif) in the utept with a sample of 6,555 examinees who took the test in november 2006. specifically, the lr dif two-degree of freedom chi-squared test of significance was employed to test the significance of dif. following what has been recently suggested in the literature, the test of significance for dif was accompanied by a measure of magnitude, namely the r-squared effect size for lr dif, for the interpretation of which the two widely accepted classification schemes were used comparatively. the results reveal that 39 of the 100 items in the test display significant gender differences. however, these group differences are viewed as “negligible” based on both of the schemes. accordingly, it could be argued that the utept is a gender dif-free test, though the reading comprehension section of the test remains in need of further analysis as it seems that the general trend of dif indices at the item level may hint at an inclination towards males.