The quality of test items is essential for producing accurate and reliable learning assessments. High-quality instruments not only measure students’ learning achievement but also provide valuable information for improving instructional practices. This study aimed to analyze the quality of essay test items on the topic of Changes in the States of Matter for fourth-grade elementary school students based on validity, reliability, difficulty level, and discrimination power. A quantitative descriptive method with an item analysis approach was employed. The participants were 22 fourth-grade students from a public elementary school in Pati Regency, Central Java. The instrument consisted of 20 essay questions administered during the second semester of the 2025/2026 academic year. Data were analyzed using Microsoft Excel to calculate validity coefficients, reliability coefficients, difficulty indices, and discrimination indices. The findings showed that 70% of the test items met the validity criteria, while 30% required revision. The reliability coefficient of 0.83 indicated high consistency of the instrument. Difficulty level analysis revealed that 80% of the items were easy, 10% were moderate, and 10% were difficult. The discrimination power analysis showed that 35% of the items were poor, 20% satisfactory, 30% good, and 15% excellent. Overall, the test items demonstrated adequate quality for classroom assessment. The findings suggest that revising items with low validity and discrimination power can improve the quality of assessment instruments, enabling teachers to evaluate students’ learning outcomes more accurately, objectively, and fairly.
Copyrights © 2026