Abstract:

MOOCs (Massive Open Online Courses) have become popular worldwide since the launch of the major platforms, Coursera and edX, in North America in 2012. Learning analytics has been used to analyze the learning logs in order to estimate learners' performance, engagement and learning trends. For example, Guo et al. (2013) analyzed clickstream data of approximately 7 million video content views and reported that the completion rate of video content views dropped significantly when the length of the video exceeded 6 to 9 minutes.

TokyoTechX’s MOOC on Modern Japanese Architecture, was released by Tokyo Institute of Technology on edX from 2017 to 2019. It had a large number of participants (over 15,000), but learners had a low course completion rate and an individual video content completion rate of less than 50%.

In this study, results are presented on a re-released course, taking into account the findings of Guo et al. Specifically,
(1) the completion rate of the entire video before and after the course reorganization and the clickstream data of the video content were compared and analyzed, and
(2) the results of the course questionnaire were also analyzed.

The results showed that the average video viewing rate for the split videos was about 10% higher than that before the reorganization. In addition, analysis of clickstream data of course participants' video viewing revealed that the number of clicks on important parts and scenes related to quiz questions increased after the reorganization. In addition, the results of the questionnaire survey showed that the split or shorter videos were viewed positively in that they made it easier to grasp important points in the video, easier to review, and reduced mind wandering while watching the videos.

Based on the above results, this study contributes to the field of MOOCs as a concrete example on how MOOCs be revised effectively and relatively quickly to increase learner engagement.

Keywords:

MOOC, multimedia learning, online education, learning analytics.