Daya Pembeda Butir Soal dalam Evaluasi Pembelajaran Bahasa Indonesia: Kajian Literatur

Abstract

Learning evaluation requires high-quality test items to ensure that students’ assessment results can be measured accurately. However, test items with low discrimination power are still frequently found, making them less effective in distinguishing students’ abilities appropriately. This study aims to analyze the quality of test items based on discrimination power in Indonesian language learning evaluation through a literature review. Discrimination power is an important indicator in determining the ability of test items to differentiate between high-achieving and low-achieving students, thereby enabling a more accurate evaluation of learning quality. The research method employed was a literature review using scientific articles obtained from Google Scholar as data sources. The analyzed articles were selected based on inclusion criteria, namely studies discussing the discrimination power of test items in Indonesian language learning and published between 2022 and 2024. The search results yielded five articles relevant to the focus of the study. Data analysis techniques included data reduction, data presentation, and conclusion drawing based on findings from previous studies. The results of the review indicate that the quality of discrimination power of test items in Indonesian language learning remains varied. Several studies showed that most test items had good discrimination power, while other studies still identified items categorized as poor and very poor, indicating the need for improvement. This review demonstrates that discrimination power analysis should be conducted in the development of evaluation instruments to improve the quality of test items and to produce more accurate learning evaluation outcomes.