Evaluasi dan Rekonstruksi Butir Soal Pilihan Ganda sebagai Asesmen Formatif dalam Pembelajaran Berbasis Deep Learning

Abstract

Formative assessment plays a crucial role in supporting the learning process, particularly in developing higher-order thinking skills (HOTS). However, the creation of multiple-choice items used in schools is still dominated by lower cognitive levels and does not yet fully support deep learning-based instruction. This situation indicates a gap between curriculum requirements and assessment practices in the field. This study aims to analyze the pedagogical quality of multiple-choice items in supporting formative assessment and to reconstruct the items so that they align with HOTS requirements and deep learning-based instruction.

            This study employs a qualitative approach of a descriptive nature using document analysis techniques. The research data consists of 10 multiple-choice items found in the Indonesian language teaching module on popular scientific articles at SMP Negeri 10 Semarang for the 2025/2026 academic year. The analysis was conducted based on several aspects, namely the cognitive levels of Bloom’s revised taxonomy, the function of formative assessment, the quality of item construction, and alignment with deep learning principles. The conclusion of this study indicates that multiple-choice items can be developed into more meaningful formative assessment instruments if designed based on HOTS and deep learning, thereby fostering deep understanding and developing students’ critical thinking skills.