Item Analysis of the Final Semester Evaluation in Madrasah Ibtidaiyah Thotiqul Huda Malang

Abstract

This study aims to analyze the quality of the final semester evaluation items for the Arabic language subject of fifth-grade students at Thoriqul Huda Islamic Elementary School (Madrasah Ibtidaiyah) in Wagir. The study employed a quantitative descriptive method with documentation techniques as the data collection tool, which included the test items, answer keys, and student responses. The data were analyzed using "Anates" software to measure reliability, difficulty level, discrimination power, and distractor quality. The results showed that the test instrument had a high reliability level with a coefficient of 0.78, and most of the questions were at a moderate difficulty level. There were 22 items with good to excellent discrimination power, 14 items with poor discrimination, and 2 items with negative discrimination power. Regarding distractor quality, the analysis indicated that some answer choices still need improvement. Based on the analysis, 19 items can be reused in future tests, while 21 items require revision and further development. The study recommends that the item development team pay more attention to question quality standards to ensure effective measurement of students' learning achievements.