academic discipline dif in an english language proficiency test

the purpose of this study was to detect differentially functioning items in the university of tehran english proficiency test (utept) which is a high stake test of english developed and administered by the language testing centre of the university of tehran. this paper is based on the answers of 400 test takers to the test. all participants earned a master degree either in humanities or science and engineering. to achieve the purpose of this study, the participants were divided into two equal groups. the results of generalized mantel-haenszel indicated that out of 100 items 12 items were displaying dif. logistic regression procedure also flagged 14 items as exhibiting dif. nevertheless, the associated test of effect size for logistic regression showed that none of these effect sizes were large according to the guidelines proposed by both zumbo and thomas (1997) and jodoin and gierl (2001). therefore, it was concluded that utept doesn't show significant academic discipline dif and is equally fair to both humanities and science and engineering groups.