<![CDATA[Understanding the misconceptions of first-year undergraduate students in solving inequality problems is crucial in the era of AI-assisted learning. This study aims to identify and analyze the types of misconceptions that students exhibit when solving inequality problems using ChatGPT, employing Newman’s framework as the analytical lens. A qualitative approach with a collective case study design was adopted to explore the complexity of misconception phenomena within human-AI interaction. Four first-year undergraduate students were selected through purposive sampling to represent diverse characteristics and challenges in mathematical problem solving. Data were collected through task-based interviews guided by semi structured interview protocols and a set of inequality problems. Data validity was ensured through triangulation by comparing responses both within and across participants. The findings revealed seven categories of misconceptions, namely misinterpreting the direction of inequality, inconsistent sign changes, confusion in combining like terms, incomplete understanding of absolute value, misunderstanding of graphical representation, lack of attention to cases, and neglecting to verify solutions. Overall, this study highlights the diverse patterns of misconceptions emerging from student and AI interactions. The results provide important insights for mathematics education and contribute to the development of AI-based learning strategies that enhance reflective thinking and metacognitive awareness, helping students strengthen their reasoning and conceptual understanding in mathematical problem solving.]]>
Copyrights © 2026
