Systematic Analysis of Learning Models to Improve Students' Mathematical Representation Skills

Abstract

This study aims to systematically analyze the learning model used in improving students' mathematical representation skills. The research approach used is Systematic Literature Review (SLR) with reference to the PRISMA flow, which involves the process of identification, selection, and synthesis of Scopus indexed articles published in the range of 2020–2025. From the results of the screening of data sources from Scopus, ScienceDirect, SpringerLink, Taylor & Francis, Wiley, and MDPI, five core articles were obtained that met the criteria of relevance and quality. The analysis was carried out through thematic coding techniques to identify the focus of learning interventions, representational strategies, and the achievement of students' mathematical representation skills. The results of the analysis showed that learning models such as Problem-Based Learning, Realistic Mathematics Education, Brain-Based Learning-based TANDUR, Think Talk–Write, and GeoGebra-assisted Discovery Learning consistently improved students' verbal, symbolic, and visual representations. Improvement occurs when learning provides meaningful context, structured representational scaffolding, and visual-interactive manipulation as a medium for concept exploration. In conclusion, the effectiveness of improving mathematical representation skills is not only determined by the chosen learning model, but also by the quality of pedagogical design and the consistency of teachers in systematically integrating representational strategies in the learning process.