Using rasch model to identify differential item functioning of teachers' job satisfaction scale with respect to gender

Abstract

Making sure that a test or scale is not biased is one of the essential factors to consider in selection and use of psychological test. This means that it is important that a scale is fair to all respondents in a population. Item Response Theory (IRT) standards show that a scale should be independent of the properties of the sample. Differential Item Functioning (DIF) means the difference between psychometric properties of an item between groups that have the same ability. Specifically, this research determined differential Item Functioning of Teachers' Job Satisfaction Scale (TJSS) with respect to gender, using the Rasch model. The sample comprised of 972 teachers from 36 public secondary schools in eight Local Government Areas in Enugu State of Nigeria. The researcher developed a 90- item instrument. This was trial-tested and factor analysis was run but only 55 items survived. In order to answer the research question, Conditional Maximum Likelihood Estimation Technique of the Win steps 3: 80 Rasch software (Linacre, 2014), was used to analyze the data. From the result, some Q/ the items in the scale functioned differently. This is an indication of DIF effects. It means that some items are not measuring what they are expected to measure. It was recommended that psychometricians should adopt IRT techniques so that a scale will be fair to all respondents in a population.