Eye-Tracking Analysis of Students’ Problem-Solving Behaviors for Learning Support in a Tutoring Context

Abstract

Understanding students’ visual engagement during problem solving can provide insight into learning strategies beyond performance outcomes. This study explores gaze-based indicators of learning behavior in a real-world tutoring context using non-intrusive, webcam-based eye tracking. Junior high school students solved computer-based multiple-choice questions while gaze data were synchronized with problem-solving logs. We propose a simple analytical framework based on areas of interest representing question statements, answer choices, and non-task regions, and compute descriptive gaze metrics capturing attention allocation, comparison behavior, and exploration patterns. Group-level analysis based on task accuracy suggests that higher-performing students tend to exhibit more frequent transitions between questions and answer choices and more diverse gaze movements, while time-based measures show limited differentiation. These findings indicate that qualitative gaze patterns can provide complementary insight for supporting tutoring instruction and reflective learning support in everyday educational settings.

Keywords: Eye Tracking, Eye Gaze, Eye Tracker, Education

DOI: 10.54941/ahfe1008002