Exploring Learners' Data of Automated Feedback System in Online Programming Course: What Can Be Learned?

• In Event: Computers and Learning: Programming and Computational Thinking

Abstract

Automated testing and feedback systems are widely applied in programming courses, providing learners with instant feedback and helping them to benefit from practice. In this study we investigated the potential of analyzing submission data, collected in ATF logs, from pedagogical point of view and in terms of learning. Examining data of an online programming course with over 100,000 learners’ submissions, we defined and calculated measures related to tasks’ characteristics, learners’ behavioral patterns and the response to feedback. Clustering methods were applied, to identify learners and tasks’ clusters. Findings suggests that ATF data analysis may expand the understanding of learning processes in a programming course, and to contribute to improving both the course and the learning outcomes

Authors

• Hagit Gabbay, Tel Aviv University

• Anat Cohen, Tel Aviv University