The effect of the subject's residence on differential item functioning in the national math test of the experimental group of mathematics and technology 2017

Abstract

Due to social, economic, cultural and linguistic differences in different provinces of the country, the university entrance exam items may not be valid for all provinces, which makes comparisons unfair. The purpose of this study is to investigate the effect of the province of residence of the subject as a source of differential item functioning(DIF). Approximately a sample of 1000 people from each province was selected. multilevel factor mixture analysis placed the provinces in 3 homogeneous classes, which shows that the subjects do not have the same response pattern in the provinces, and therefore the subject's residence is one of the sources of the DIF. Post hoc comparison of homogeneous classes showed that age groups are also one of the factors affecting DIF in the provinces. In this study, the ratio of men and women in the latent classes is the same, so the subject gender cannot be reasons of DIF. The statement that subjects in a group (for example, men) answer a DIF item differently leads to a stereotypical view of the item, which benefits all members of the group (men). Therefore, if latent classes are formed first and then the differences in the characteristics of the subjects in the classes are examined, it leads to a more informative analysis of DIF.

Keywords

• educational justice
• differential item functioning
• multilevel factor mixture modeling