Exploring Relationships Between Distractibility and Eye Tracking During Online Learning

More than half of students think their attention is easily shifted when they’re learning online. Distractibility, to a certain extent caused by visual stimuli is the main impact to decrease their academic performance. In addition, eye-tracking technology has been widely applied to explore distractibility in many “look” tasks, such as reading, viewing advertisements, and watching online videos as well as measure the efficiency of visual cognition. Therefore, this paper aimed to discuss the relationship between distractibility with eye movement indices and academic performance. Fifty high school students (30 girls) were recruited to complete experiment that was divided into two groups, which are the experimental group with distractions and controls with no one. The result showed that three of traditional eye movement indices were significantly correlated with distractibility (\(p < 0.05\)). Then we introduced the network accessibility model and the gaze transformation entropy to create two composite indexes according to the complexity and directivity of distractibility characteristics. The result revealed that the two composite indexes are significantly correlated with distractibility (\(p < 0.05\)). Finally, we constructed the mapping model about eye movement metrics about distractibility and online learning performance with a machine learning algorithm. The result ration was \(R^{2} = 0.799,\) and the error was \(Re < 0.1\), which proved the model was feasible and accessible. The research from the perspective of distractibility can provide valuable support for physiological indicators testing tools of academic performance and highlights the applications of eye movement dynamics.