Improving the Quality of Learning Based on Evaluation Instruments

Abstract

This research aims to analyze the quality of teachers' learning evaluation instruments as final learning tests. This research uses quantitative methods with a descriptive approach. The subjects in this research were Islamic high school students in Indonesia. The data collection technique is carried out using the documentation method. The data collection instruments used in this research were exam questions, answer keys, and student answer sheets. The data obtained was analyzed manually with the help of Microsoft Excel. The analysis results show that first; regarding The validity of the end-of-year assessment questions in the Jurisprudence subject, there are 22 questions or 44% declared valid, and 28 questions, or 56% declared invalid. Second, the reliability of the final year assessment questions for the Fiqh subject is known to have a reliability of 0.7651, so it can be concluded that these questions have a high level of reliability. Third, regarding the level of difficulty of the final year assessment questions in the Fiqh subject, there are 9 questions, or 18% which are included in the difficult question category, 27 questions or 54% are included in the medium question category, and 14 questions or 28% are included in the easy question category. Fourth, the discriminating power of final-year assessment questions in Jurisprudence subjects shows that the items with very poor discriminating power totaled 4 questions (8%). There were 20 questions (40%) of poor discriminating power, 5 questions (10%) of good discriminating power, 7 questions (14%) of good discriminating power, and 14 questions (28%) of very good discriminating power. The results of this research can be used as initial data for future researchers in studying this problem in different contexts and issues and as consideration for teachers in improving their abilities and understanding in creating quality questions.