A Narrative Review of Student-Centered, Data-Driven Approaches for Improving Success in Mathematics Education
Abstract
Recent international evidence indicates that mathematics achievement remains a challenge across educational settings. At the K–12 level, many students experience low mathematical proficiency and confidence, while higher education institutions continue to face challenges related to student engagement, retention, and academic performance in mathematics. This narrative review examines peer-reviewed studies published between 2020 and 2026 on student-centered learning and the use of artificial intelligence (AI) in mathematics education. Relevant studies were identified through searches of IEEE Xplore, SpringerLink, MDPI, Crossref-indexed journals, and other open-access academic journals, while OECD (PISA 2022) and UNESCO reports were consulted to provide global context. The search used combinations of keywords related to student-centered learning, artificial intelligence, mathematics education, learning analytics, formative assessment, personalized learning, student engagement, and learning outcomes. Only English-language, peer-reviewed studies directly related to mathematics education were included. A total of 15 peer-reviewed studies employing quantitative, qualitative, and mixed-methods designs were analyzed using a thematic synthesis approach, while two institutional reports were used only to provide global context. The findings consistently showed that integrating student-centered teaching with AI-supported tools, including learning analytics, adaptive learning systems, and formative assessment, improved student engagement, motivation, personalized learning, and mathematics achievement. The review also identified common implementation challenges, including limited technological infrastructure, insufficient teacher preparation, ethical concerns, and unequal access to digital resources. Overall, the evidence suggests that AI is most effective when integrated with student-centered pedagogical practices rather than used as a standalone technological tool.