Cross-System Validation of Engagement Prediction from Log Files

Engagement is an important aspect of effective learning. Time spent using an e-Learning system is not quality time if the learner is not engaged. Tracking student disengagement would offer the possibility to intervene in order to motivate the learner at appropriate time. In previous research we demonstrated the possibility of predicting engagement from log files using a web-based e-Learning system. In this paper we present the results obtained from another web-based system and compare them to the previous ones. The similarity of results across systems demonstrates that our approach is systemindependent and that engagement can be elicited from basic information logged by most e-Learning systems: number of pages read, time spent reading pages, number of tests/quizzes and time spent on test/quizzes.