<![CDATA[This study aimed to analyze the quality of assessment instruments on ecosystem material based on the aspects of validity, reliability, difficulty level, and discriminating power of test items. The study employed a quantitative descriptive research method. The research population consisted of all tenth-grade students at SMKN 1 Onohazumba, with the research sample comprising one agricultural major class. The instrument used was a multiple-choice test consisting of 30 items developed according to the learning indicators. Data collection techniques were carried out through learning achievement tests and documentation. Data analysis was conducted using validity testing with the Product Moment correlation formula, reliability testing using the KR-20 formula, difficulty level analysis, and discriminating power analysis of test items. The results showed that all test items were declared valid, with the highest r_xyvalue of 0.506 and the lowest value of 0.368, where all r_xyvalues were greater than the r-table value of 0.367. The reliability test produced a reliability coefficient of 0.844, which falls into the very high category. The difficulty level analysis indicated that all test items were categorized as moderate, with difficulty index values ranging from 0.32 to 0.68. Meanwhile, the discriminating power analysis showed that 20 test items were categorized as good, while 10 test items were categorized as excellent.]]>
Copyrights © 2026
