Item Analysis of Arabic Daily Test for Fifth-Grade Students Using the Rasch Model at MIN 1 Pasaman Barat

Abstract

Assessment in Arabic language learning plays an essential role in measuring students’ cognitive achievement; however, the quality of test instruments is often not empirically validated. This study aims to analyze the quality of Arabic daily test items for fifth-grade students at MIN 1 Pasaman Barat using the Rasch model, focusing on item validity, reliability, and difficulty level. This research employed a quantitative approach with a descriptive-exploratory design. Data were collected from students’ responses to multiple-choice items and analyzed using Rasch measurement through fit statistics, reliability indices, and logit-based item difficulty. The results showed that only a small number of items met the validity criteria based on Rasch fit statistics, while the majority were identified as misfit. The item reliability was categorized as moderate, indicating that the internal consistency of the instrument has not yet reached an optimal level. In addition, the distribution of item difficulty was unbalanced, with most items classified as difficult, suggesting a mismatch between item difficulty and students’ ability levels. These findings indicate that the test instrument requires improvement in terms of item construction and difficulty distribution. The study implies that Rasch analysis can serve as an effective diagnostic tool to enhance the quality of assessment instruments, contributing to more valid, reliable, and standardized evaluation practices in Arabic language learning at the primary school level.