Determination of item bias according to expert opinion before application and after classical test theory

The aim of the study is to carry out a research on finding the item bias analyzed after the exams without taking the exam. The study was conducted with students at 8th grade in South East region. For the research, the previous researches were examined and the biased item topics were determined and the items related to these topics were prepared. After getting expert opinions, necessary permissions were obtained and a trial application was made for 245 people. After the trial application, the data were analyzed by EFA and items with a factor load below .30 were excluded from the test. With the final test, the main application was made to a group of 881 students. After the application, expert opinions were taken without analyzing which items were biased and at which level. After receiving expert opinions, item bias was examined with Mantel – Haenszel and Logistic Regression method, which is one of the Classical Test Theory bias determination methods. According to the results of the analysis, it was determined that there was a bias in five items. According to the results of the analysis, it was determined that there was a bias in five items. Experts have correctly estimated the gender element in the wrong items at the rate of 80%. As for the level, it can be said that it has been estimated at a high level, although there are disagreements.