Kualitas Butir Soal dalam Evaluasi Bahasa Arab : Kajian Empiris pada PAS Program Intensif

This study evaluates the quality of multiple-choice test items in the Final Semester Assessment (PAS) of the Arabic Intensive Program at UIN Sunan Ampel Surabaya. Using a descriptive quantitative approach, the research analyzes difficulty index, discrimination index, and distractor efficiency of 25 questions in the writing skill (maharah kitabah) domain. Results reveal that 52% of the questions are categorized as easy, 40% as moderate, and 8% as difficult. Regarding discrimination, only 4% of the items are highly effective, while 48% are categorized as poor, with 8% showing negative indices. Distractor analysis shows that 45% function effectively, while 30% fail to meet the criteria. The findings highlight imbalances in difficulty distribution, limited discrimination capacity, and inconsistent distractor efficiency. Recommendations include revising poorly functioning items to enhance evaluation accuracy and fairness, ultimately supporting effective Arabic language learning and assessment development