An Analysis of ChatGPT's Instructional Explanation Performance for Teaching Mathematical Connection

This study explores the educational potential of ChatGPT in developing mathematical connection competencies, focusing on its instructional explanations for difficulties related to linear functions. Four types of connections were categorized as conceptual connections, representational connections, mathematization connections, and interpretation connections. Cases of difficulties in each type of connection were extracted from the literature, and ChatGPT’s explanations for each case were analyzed based on five criteria for effective explanations. The analysis revealed that ChatGPT provided accurate diagnoses and effective explanations for conceptual connections and mathematization connections. However, explanations for representational connections and interpretation connections tended to emphasize procedural aspects, suggesting areas for improvement. Based on these findings, this study discusses the strengths and limitations of ChatGPT and proposes implications for its effective use in teaching mathematical connection competencies.