Quality profile of Arabic final semester assessment items: A psychometric analysis

Background: The quality of assessment instruments is essential to ensure that students’ learning outcomes are measured accurately. In Arabic language learning, Final Semester Assessments (PAS) must be supported by sound psychometric qualities to function as valid and reliable evaluation tools. Aims: This study aims to examine the quality profile of Arabic PAS items at MAN 1 Gresik by analysing their psychometric characteristics and identifying items that are feasible, need revision, or are not feasible for use. Methods: This research employed a quantitative descriptive design using psychometric item analysis. The data consisted of 40 multiple-choice PAS items and students’ response sheets. The analysis integrated content and construct validity with empirical indicators, including point-biserial validity, KR-20 reliability, item difficulty, and item discrimination, using Microsoft Excel and ANATES V4. Results: The results show that content validity reached 92.5%, construct validity reached 82.85%, and empirical validity was moderate (r = 0.60). The overall test reliability was high (r₁₁ = 0.75). Item difficulty is dominated by medium-level items, while item discrimination is the weakest aspect. Based on integrated psychometric criteria, 40% of items are feasible, 57,5% require revision, and 2,5% are non feasible. The causes of the failure of the test items, content validity (7.5%), construct validity (42.5%), empirical validity (2.5%), level of difficulty (12.5%), and discrimination index (22.5%). Implications: These findings highlight the importance of systematic psychometric evaluation in Arabic language assessment. Improvements are needed in construct validity, especially Arabic language accuracy, distractor effectiveness, and item discrimination. Such an approach supports the improvement of school-based Arabic assessments to ensure more valid and reliable measurement of students’ learning outcomes.