Analisis Tes Pilihan Ganda Mata Pelajaran Bahasa Arab Mts Al-Hidayah Kota Batu

Abstract

This study aims to analyze the difficulty level and discriminatory power of multiple-choice questions in the Arabic subject at MTs Al-Hidayah, Batu City. Item analysis was conducted to determine the extent to which each item is able to measure students' abilities proportionally and differentiate between high-ability and low-ability students. The research method used was a descriptive quantitative approach with a sample of 25 questions tested on all eighth-grade students. Data were analyzed using the difficulty index (P) and the discrimination index (D) with classical measurement standard interpretation criteria. The results showed that 64% of the items were classified as easy, 32% were moderate, and 4% were difficult, indicating that the composition of the questions was not proportional. The discrimination power analysis showed that 56% of the items had good to excellent discrimination power, while 24% were classified as very low with six items having negative discrimination values. The relationship between difficulty level and discrimination power showed a consistent pattern, where items with a moderate level of difficulty had more optimal discrimination power. Overall, there were 5 items categorized as very good, 11 items were good with minor improvements, 2 items needed substantial revision, and 7 items needed total revision. The results of this study emphasize the importance of empirical item analysis as a basis for developing valid and reliable evaluation instruments, thereby improving the quality of Arabic language learning assessment in madrasas.