Investigating Differential Item Functioning of Ankara University Examination for Foreign Students by Recursive Partitioning Analysis in the Rasch Model

Abstract

This study aims to determine whether items in the Ankara University Examination for Foreign Students Basic Learning Skills Test function differently according to country and gender using the Recursive Partitioning Analysis in the Rasch Model. The variables used in the recursive partitioning of the data are country and gender. The population of the study is composed of 2476 individuals. Since the study includes comparisons across countries, the country is accepted as a criterion in determining the sample group. Thus, the sample of the study consists of 615 individuals selected from Azerbaijan, Bulgaria, and Syria. To investigate differential item functioning (DIF) of the items of the test, the Rasch tree method was used. As a result of the analysis, DIF has been detected in 16 items at the 0.001 significance level. However, these items have been identified to have similar difficulty parameters in all countries. Finally, items have not shown DIF according to gender.This study aims to determine whether items in the Ankara University Examination for Foreign Students Basic Learning Skills Test function differently according to country and gender using the Recursive Partitioning Analysis in the Rasch Model. The variables used in the recursive partitioning of the data are country and gender. The population of the study is composed of 2476 individuals. Since the study includes comparisons across countries, the country is accepted as a criterion in determining the sample group. Thus, the sample of the study consists of 615 individuals selected from Azerbaijan, Bulgaria, and Syria. To investigate differential item functioning (DIF) of the items of the test, the Rasch tree method was used. As a result of the analysis, DIF has been detected in 16 items at the 0.001 significance level. However, these items have been identified to have similar difficulty parameters in all countries. Finally, items have not shown DIF according to gender.