Psychometric Analysis of High School Mathematics Ability Test

Aim. Mathematics is essential for developing higher-order thinking skills. However, Indonesian students' performance in PISA 2022 and TIMSS 2015 shows a lack of proficiency in these skills. One contributing factor is school assessments that do not adequately foster higher-order thinking. This study evaluates the quality of a high school mathematics proficiency test designed to assess higher cognitive levels, focusing on validity, reliability, discrimination, difficulty, and variation in student responses. Methods. This quantitative study used the MEASURE approach with five essay problems on exponents, logarithms, geometric sequences, arithmetic series, and trigonometry, involving three validators and 102 high school students from three district schools in Banjarmasin. Results. The I-CVI values for questions 1, 2, 3, and 5 met the required criteria, requiring no revision. Question 4, with an I-CVI of 0.90, needed revision. All kappa values (κ) were ≥ 0.9, confirming reliability. Empirical data showed all questions had significance values below 0.05, and Cronbach’s alpha was 0.72. Discrimination analysis categorized 60% of questions as excellent and 40% as good. Difficulty levels indicated that 80% were moderate and 20% difficult. The instrument generated diverse student responses, reflecting varying ability levels. Conclusion. The high school mathematics proficiency test, designed for cognitive levels C4 and C5, met validity and reliability criteria, demonstrated good discrimination, and had varied difficulty levels. While emphasizing higher-order thinking, a more comprehensive assessment could integrate multiple-choice questions for C1-C3 and essays for C4-C5.