Applying Learning Analytics to Deconstruct User Engagement by Using Log Data of MOOCs

• Published in: Journal of Information Science and Engineering Vol. 34; no. 5; pp. 1175 - 1186
• Main Authors: 劉明機(MING-CHI LIU), 余禎祥(CHEN-HSIANG YU), 吳榮彬(JUNGPIN WU), 劉安之(AN-CHI LIU), 陳錫民(HSI-MIN CHEN)
• Format: Journal Article
• Language: English
• Published: Taipei 社團法人中華民國計算語言學學會 01.09.2018; Institute of Information Science, Academia Sinica
• Subjects: CAI; Computer assisted instruction; Distance learning; Online instruction; Regression analysis; Sequential analysis; system log analysis; emotion; user engagement; learning analytics; MOOCs
• ISSN: 1016-2364
• DOI: 10.6688/JISE.201809_34(5).0004

Previous research noted that software requirement analysis should move beyond usability to understand and design for more engaging user experiences. This study analyzes system logs collected from the Massive Open Online Courses (MOOCs) to capture user engagement over time. We measured engagement through mapping event logs with three components of engagement. We further analyzed the helpfulness of engagement measurement in predicting grades. The results showed that there was a significant and moderate positive correlation between the behavioral, cognitive, and emotional engagement and quiz scores. A multiple linear regression analysis also showed that higher behavioral and cognitive engagement were related to higher quiz scores. Thus, this study performed three classification methods and found that the ANN method got the highest accuracy. This study applied sequential analysis to discover the difference between video watching behavior of the high and low quiz scores students. The results indicate that it is important to take into account all of the behavioral, cognitive, and emotional engagement in understanding user engagement during the system development process.