DETECTION OF GENDER BIASED USING DIF (DIFFERENTIAL ITEM FUNCTIONING) ANALYSIS ON ITEM TEST OF SCHOOL EXAMINATION YOGYAKARTA

Abstract

This study aim to detect an item that biased on gender (divided by male and female) by DIF (Differential Item Functioning) on  school examination test in Yogyakarta regency in islamic education subject for third grade of Senior High School based on item respons theory . This research is quantitative research and the subject for this study are 937 Senior High School students in Yogyakarta regency taken by their response sheets on 50 items test with of choices of islamic education subject in school examination at year 2016. Data analysis based on Rasch model with Winstep 3.73 software. The research found that the there are item that biased on certain gender, there are 26 items that have probabilitas value less than 0.05 and t-value outside the range of -1.96 ≤ t-value ≤ 1.96, based on plot from the result analysis there are 14 items (6, 9, 10, 13, 15, 16, 20, 25, 28, 30, 31, 32, 41, 48) tend to give advantage for Male, and there are 12 items (1, 11, 12, 14, 21, 22, 29, 38, 40, 43, 46, 50) tend to give advantage for Female.