ASSESSMENT OF DIFFERENTIAL ITEM FUNCTIONING USING THE MANTEL-HAENSZEL PROCEDURE

When important decisions are made based on test scores, it is critical to avoid bias, which may unfairly influence examinees' scores. Bias is the presence of some characteristic of an item that results in differential performance for individuals of the same ability but from different ethnic, sex, cultural, or religious groups. Bias comes in many forms. It can be sex, cultural, ethnic, religious, or class bias. An item may be biased if it contains content or language that is differentially familiar to subgroups of examinees, or if the item structure or format is differentially difficult for subgroups of examinees. One of the ways of detecting bias is to assess the Differential Item Functioning of the items. This study probes into the differential item functioning of the data obtained from the National Talent Search Examination (NTSE) conducted by National Council of Educational Research and Training (NCERT) in India. National Talent Search (NTS) Scheme is the most prestigious scheme of the country. The scheme has contributed in identifying and promoting the talent and is recognized globally. The Scheme was started in 1963. During the last five decades, the scheme has been reviewed and modified a number of times to keep up with the demands of the changing times. In the present National Talent Search Scheme, identification of talent is a two-stage selection process. First stage is carried out by the individual States and the second stage selection at the national level is carried out by the NCERT. At the national level 4000 candidates take the examination, which is conducted in various parts of the country. The examination consist of Mental Ability Test (MAT) and Scholastic Aptitude Test (SAT) in which there are 100 items of one mark each. Both the tests are of objective type with multiple choice items. Differential Item Functioning occurs when members of two different groups with comparable ability on a construct respond differently to a given item. Ideally, item functioning should be invariant with respect to irrelevant aspects of the test-takers, such as gender, ethnicity and socio-economic status. Items functions differently in the same ability sub groups when construct irrelevent aspects are introduced.In differential Item Functioning (DIF) the items in a test, are considered one at a time, to look for signs of interactions with construct irrelevant characteristics. In the most used Mantel-Haenszel procedure, the comparison is made between the reference and the focal groups.These groups are stratified into matching ability levels and their relative performance on each item is quantified. The ability levels are determined by the total scores on the test. Thus, the DIF analysis for one item is as independent as possible of the DIF analyses of the other items.