Testing Fairness

Numerous experts have underscored the need of fairness in National Entrance Examination items. This study examines whether examinees' performance on items of the National University Entrance Exam for Foreign Languages (NUEEFL), known as “Konkour,” varies based on background, specifically gender, school type, and ethnicity, rather than language proficiency, as the detection of differential item functioning (DIF) may enhance the fairness of high-stakes tests. The research employed a quantitative non-experimental, cross-sectional design. The participants included 200 male and female students, who were chosen randomly from students studying at Islamic Azad University, Science and Research branch in Tehran, Iran. The instruments consisted of a mock NUEEFL test and a researcher-made questionnaire. Upon taking the participants’ consent, the researcher took the mock version of NUEEFL. Next, the participants were asked to answer the questionnaire about their demographic information, including their gender, school type, and ethnicity. A three-phase DIF analysis was conducted to explore examinees' performance across these demographic variables. The results indicated that school type exhibited the most significant DIF, particularly in grammar and cloze assessments, whereas gender DIF was mostly seen in grammar and language function. Moreover, ethnically differential item functioning was significant in vocabulary and cloze assessments. Furthermore, reading comprehension was mostly impartial, with the exception of school type. The results underscore the need for test developers to consider demographic factors to ensure fairness and validity in high-stakes testing contexts.