Using Transformed Item Difficulty Procedure to Assess Gender –Related Differential Item Functioning Of Multiple-Choice Mathematics Items Administered In Nigeria

Abstract

This study investigated gender-related differential items functioning in multiple-choice mathematics items administered by National Examination Council (NECO) in 2011 in Nigeria. The method employed was Transformed Items Difficulty (TID). The samples used for the study were drawn from data set containing responses of 1600 examinees from 15 secondary schools randomly drawn from 20 co- educational senior secondary schools in Nsukka Local Government Area of Enugu state, Nigeria. The sample was made up of 774 males and 826 females. The study provides evidence that there are gender differences in performance in multiple-choice test items in mathematics that vary according to contents area even when the content is closely related to curriculum. However, very few items (8 out of 50 items) were in favour of males while non was in favour of females. The implication of this is that most of the items neither favour males nor females when subjected to analysis using TID. However, test items may carry with them demands other than those associated with gender. The significance of this study is that the results show an incorrect picture of the quality of education for different groups and this may likely lead to the resources for education being distributed in an unfair manner.  From these results it is therefore recommended that test developers need to balance some aspects of content to reduce form-to-form variation in the effects of item properties on the performance of male and female examinees.

Keywords: Transformed item Difficulty, Differential Item functioning, Perpendicular distance, item difficulty, Delta value, Mathematics.