Unidimensionality Analysis and Differential Item Functioning in the Applied Mathematics Final Examination Test of Politeknik Negeri Bali Students Using the Rasch Model

Abstract : The objectives of this research are (1) to analyze the fulfillment of unidimensionality assumptions on Applied Mathematics FSE instruments by the Rasch Model, (2) to identify and analyze the existence of items that have differential item functioning (DIF) among study programs in the polytechnic, and (3) to interpret as well as recommend improvements to instruments based on results from analysis about unidimensionality and DIF. This survey is quantitative research with a cross-sectional design. The sample consists of 206 students coming from three study programs: the Mechanical Engineering, the Refrigeration and Air Conditioning Engineering, and the Utility Engineering Technology, Politeknik Negeri Bali (PNB), selected by the total sampling method. The research instrument comprises five questions constructed according to an outcome-based education (OBE) curriculum where content validity was checked through Aiken’s V method, giving a value equal to 0.91. Data were analyzed using Winsteps 5.9 software based on the Rasch model. Results of the analysis indicated that the test has outstanding reliability and separation indices both on respondents and items. Most items fall within the fitting criteria of the Rasch Model. The unidimensionality test proved that it can consistently measure a single construct in applied mathematics competency. Based on DIF analysis, most items work well across different learning categories, but a few items have large and significant DIFs. These imply that this instrument is suitable to be used as an assessment tool for learning outcomes where several parts still need improvement to enhance fairness in measuring. It raises awareness about conducting comprehensive evaluations on assessment tools within the OBE context in vocational education.