Study of Bias in 2012-Placement Test through Rasch Model in Terms of Gender Variable

Abstract

Validity and reliability are among the most crucial characteristics of a test. One of the steps to make sure that a test is valid and reliable is to examine the bias in test items. The purpose of this study was to examine the bias in 2012 Placement Test items in terms of gender variable using Rasch Model in Turkey. The sample of this study was 216363 participants that randomly selected out of 1075546 students who took Placement Exam (SBS) in 2012.  A stratified sampling was used for this study. Winsteps 3.8.1 package program was used in data analysis. Through statistical procedures it has been established that two items in Maths subtest and one item in Social Sciences subtest, namely three items in total, have Differential Item Functioning (DIF). According to expert opinion regarding the bias in these three items no bias was available in the items. Consequently, it had been established that there was no biased question in Turkish language, Mathematics, Science and Technology and Social Sciences subtests within the 2012 Placement Test and they were valid and reliable.

Keywords: Placement test, Item Bias, Differential Item Functioning, Rasch Model