Developing of Computerized Adaptive Test (CAT) Based on a Learning Management System in Mathematics Final Exam for Junior High School

Abstract

Purpose: The research aimed to develop a valid, feasible, and efficient computerized adaptive test (CAT) to measure students’ abilities on the end-of-semester exams. Design/Approach/Methods: This research was a 4D research and development with four stages, namely, definition, design, development, and dissemination. Stratified random sampling techniques were used in the study by randomly sampling based on high, moderate, and low levels of academic achievement ability. The study sample was 637 junior high school students in Yogyakarta. Finding: The results showed that the developed CAT program was considered very good and feasible to use. The validity of the contents was obtained by 200 valid questions, and in the characteristics of the item, all items were declared to fit with the Rasch model. All problem items with item difficulty index were relatively good. The reliability of the item estimate was in the very good category while the reliability of the case estimate was in the good category. The lowest and highest ability test results obtained were in the range of 1.1 to 9.85. The ability of test takers was diverse and had scattered distributions. The feasibility of using CAT media was obtained from teachers and students indicating that CAT media was in a very feasible category to use.