<![CDATA[The implementation and development of formative test items often still face various challenges that impact the accuracy of evaluation results. This study aims to analyze the quality of formative test items on physics problem-solving, particularly in the topic of dynamic electricity, by applying the Item Response Theory (IRT) approach in senior high schools in Bulukumba Regency. The study focuses on four main aspects: validity, reliability, item difficulty, and item discrimination. This research employed a quantitative descriptive design with data consisting of test responses from 369 twelfth-grade students. The analysis was conducted in three stages: content validity testing using Aiken’s V to assess the alignment of items with the measured competencies; reliability testing using the Kuder–Richardson 20 (KR-20) coefficient; and item characteristic analysis based on the IRT approach to obtain indices of item difficulty and discrimination. The results indicate that all items demonstrated good content validity, with Aiken’s V values above 0.75. The reliability coefficient showed high internal consistency, indicating that the instrument is reliable as a measurement tool. Regarding item difficulty, the 15 test items ranged from easy to difficult, with the majority categorized as moderate (10 items), while a smaller number were classified as easy (2 items) and difficult (3 items). This distribution suggests that the test adequately represents students’ abilities, particularly at the average ability level. The discrimination index ranged from moderate to high, with five items categorized as high and ten items as moderate. No items fell into the low category, indicating that the test items effectively distinguish students based on their ability levels. Overall, the instrument is considered valid and reliable, with proportional levels of difficulty and good discrimination power, making it appropriate for measuring physics problem-solving competence, particularly in the topic of dynamic electricity. The implication of this study is the need for wider application of IRT-based item analysis in research on the development of educational instruments at the high school level. ]]>
