Leveraging smartwatches to estimate students' perceived difficulty and interest in online video lectures.

Choi, Jinhan, Jeongyun Han, Woochang Hyun, Hyunchul Lim, Sun Young Huh, SoHyun Park, and Bongwon Suh.

In Proceedings of the 11th International Conference on Education Technology and Computers, 2019.

Online videos have become a popular medium for delivering educational materials. Analyzing video interaction log can provide valuable educational insights. However, for small-sized online courses, due to the small size of samples, analyzing online log is often not enough for modeling students’ learning behaviors. In this study, we aim to explore the feasibility of utilizing commercial smartwatches to augment building of such models. We collected online video interaction log as well as physiological data from smartwatches and built models to estimate the perceived difficulty and interest of students while watching online video lectures. The results show that smartwatch data could significantly improve the amount of explained variance in their perceived difficulty and interest by 100% and 64% respectively. We hope the result could inform the application of a smartwatch for students’ in online video learning.