This study investigates the quality of evaluation instruments developed to measure junior high school students’ mathematical conceptual understanding, with a focus on the topic of numbers. The research employed a quantitative descriptive method involving thirty-two seventh-grade students from a junior high school in Pekanbaru. Data were obtained through essay-based assessments consisting of three items aligned with specific indicators of mathematical conceptual understanding and cognitive domain levels. The analysis examined validity, reliability, difficulty level, and discrimination power. The findings show that all three items achieved high validity, although the overall reliability coefficient was categorized as low. The difficulty level varied, with one item classified as easy and two as moderate. Discrimination power also differed among items, with one item having poor discrimination, one adequate, and one good. These results highlight the importance of conducting systematic quality testing to ensure that evaluation instruments are valid, reliable, and appropriately calibrated for assessing students’ conceptual understanding in mathematics.
Copyrights © 2025