An Investigation of Item Bias in Turkish Sub Tests in Level Determination Exam

Abstract

The purpose of this study is to determine whether 2009 Level Determination Exam (LDE) 6th, 7th and 8th year Turkish sub tests indicate item bias in terms of the gender of the students and school types. In the light of this purpose, Mantel-Haenszel (MH) and Logistic Regression (LR) methods were used in order to determine the differential item functioning of the items in the tests. The study was carried out with 22656 students in total, 6922 of whom were 6th year students (3620 males and 3302 females), 6360 of whom were 7th year students (3290 males and 3070 females) and 9374 of whom were 8th year students (4290 males and 5084 females). In the study, in terms of the gender variable, a) DIF was not found in any of the items at B and C level in the 6th year, (b) DIF was found in 2 items in the 7th year according to the MH technique and in 1 item according to LR technique, (c) DIF was also found in 5 items in the 8th year according to MH technique and in 3 items according to LR technique. In terms of school type variable, 7 items in the 6th year had DIF according to MH technique and 5 items had DIF according to LR technique; 2 items in the 7th year according to MH technique and 1 item according to LR technique had DIF; and 2 items in the 8th year had DIF according to MH technique. In the light of the experts’ views, it was found that only one of the items that include DIF caused item bias and other items did not cause any item bias according to the gender and school types; and the difference between the groups was caused by the item effect.