Abstract:Various visualization methods have been proposed for eye movement analysis in programming education. However, most of them were utilized for the general scenario of program comprehension. In this paper, we focus on the eye movement visualization in a specified scenario of error-finding test, in which the students are asked to find multiple errors in a C-language program and submit the result by clicking the corresponding checkbox. A novel visualization scheme namely Attention Distribution Graph (ADG), is proposed to describe the line-level transition of student’s visual attention. We adopt time distance to click event and line number to describe the spatio-temporal space of eye movement, and color the space grids according to the proportion of fixation duration. Experiment results show that, from the distribution of colored grids in ADG, we can infer the student’s cognitive process in a successful error-finding task. It provides a normalized view to analyze the students’ eye movement toward a same bug, especially on the influence of preceding reading to the later decision.

Attention Distribution Graph - Visualizing Student's Attention Transition in Error-finding Tasks.