Interlacing Gaze and Actions to Explain the Debugging Process

Abstract

Debugging is an indispensable skill of successful programmers. As such, teacher’s should not overlook it when teaching programming. The main aim of the study is to use gaze data combined with measures at different temporal granularities to show how these measures are related to the outcomes (in particular debugging success as a learning by doing outcome) students have at the end of the debugging task. The results delineate that combining gaze data with actions (reading, writing, scrolling) and unit tasks (main method and JUnit test) gives new insights to further understand the cognitive actions in debagging a program. Moreover, this study also focuses on discovering debugging patterns of successful students in order to improve the design of learning activities to teach students how to debug. Finally, with the analysis, the authors have shown an automatic way of detecting successful action-gaze patterns.