Application of the Rasch Model in Analysis of Exam Questions at the Faculty of Psychology of Universitas Islam Bandung

Abstract

In knowing student achievement in a particular course, students often use examinations conducted at least twice a semester. It is expected to be able to measure accurately the extent of student competence in the course concerned. But in reality, not all students get good quality grades. However, the first thing to be ascertained when student achievement is low is a measuring tool used to measure learning achievement itself whether it is already of good quality or not. When the results obtained come from a low quality achievement test kit, then the wrong conclusions will be obtained. The method used to analyse the quality of the performance measuring instrument is the Rasch model. Based on the research that has been done on the exam questions at the Faculty of Psychology Universitas Islam Bandung which is for Neuropsychology courses it is found that the quality of the exam questions is not good. This can be seen from the number of problem items that need to be revised in total or replaced with an average of 50%. The number of questions that need to be replaced illustrates that these questions cannot distinguish students who have high ability from students who have low ability. The low ability of questions to distinguish the ability of respondents in dealing with problems can be caused by the unequal level of difficulty of the item problem.