Using Rasch Trees Method to Detect Differential Item ‎Functioning in the Jordanian National Test in ‎Mathematics for Fourth Graders‎

Abstract

The present study aimed to use Rasch trees method to detect differential item functioning (DIF) in the Jordanian National Test to control educational quality in mathematics, for fourth graders, according to gender, school location, supervising authority, and test modality, as well as to investigate the effect of sample size on the method. The study population consisted of (31136) fourth graders’ responses on (25) multiple choice items fitted with the Rasch model. Rasch trees method was used to analyze (1500) randomly selected responses. The results indicated that there is DIF in terms of gender only.  Eight items, representing (32%) of the test items, were detected to have DIF, five in favor of males and three in favor of females. The results showed that Rasch trees model is sensitive to sample size, where the DIF rate increased as the sample size increased