Pengembangan Computerized Adaptive Test (CAT) Pada Materi Getaran, Gelombang dan Bunyi

Abstract

Problems in learning evaluation such as Paper Based Testing (PBT) and Computer Based Testing (CBT) which ignore the ability of individual variations cause the information obtained to be less than optimal. Experts have developed a learning evaluation tool, namely the Computerized Adaptive Test (CAT) to overcome these problems. In the CAT system, the items given to test takers are adjusted to the test taker's ability level. The purpose of this study was to determine the characteristics of the CAT evaluation tool on the topic of vibration, wave and sound as well as to be able to determine the profile of student learning outcomes in vibration, wave and sound materials using CAT evaluation tools. This study uses the Research and Development (R&D) method with a 4-D development model, namely define, design, develop and disseminate. The CAT developed is website-based and has a 78 item question bank with the topic of vibration, waves and sound. The result of the development obtained is that the validity of the instrument gets Aiken's V validity score with an average of 0.93 which exceeds the minimum validity of 0.87 and is declared valid. The validity of the media expert got a score of 86% in the very good category. The method of question analysis uses the Item Response Theory (IRT) one-parameter logistic model (1 PL). The level of difficulty of the questions obtained is easy category as many as 17 items, medium as many as 30 items and difficult as many as 31 items. The results showed that the feasibility of CAT media was 74% in a good category. Obtained profiles of student learning outcomes material vibration waves and sound using CAT with an average value of 47.91 categorized as being slightly below the average of the maximum value of 100.The ability level of test takers can be estimated with only 11-23 items resulting in efficiency better time than PBT and CBT. There is a relationship between the score using CAT and the score without using CAT positively with a correlation value r_count of 0.600192, this value can be interpreted as a strong level of relationship, if the score without CAT is high then the score obtained with CAT is also high, and vice versa.