Estimation of Test Scores Based on Video Viewing Behavior in the Programming MOOC Course

摘要

The massive open online courses (MOOCs) are expected to offer learning opportunities to various type of learners with different backgrounds. However, MOOCs have many learners who drop out during learning process, and the completion rates are as low as about 10%. In order to improve such a drawback, it is necessary to grasp the features of the learners in the earlier stage and to provide appropriate supports to each learner. This paper estimates the difference of learners’ behavior by investigating relationship between video viewing logs and test scores in the programming MOOC course. It was observed that the repeated learning relates the higher score, the later learning relates the lower score. Even in learners who got final score between 20 to 70, there are possible learners who could be rescued by offering appropriate supports. This cluster of learners was also visible in the multiple regression analysis.