Investigating Engagement and Performance in Online Mathematics Courses Using Clustering Techniques

Among the various types of learning analytics that have emerged, especially in the last decade, the analysis of students’ patterns in online learning plays a prominent role, encouraged by approaches to understanding the Web in different contexts. Students’ patterns can be analyzed by examining sets of logs made by users, each set being treated as a basic unit. The study of patterns of online activity can be applied in educational contexts. In this chapter, we perform an analysis of the logs of an online mathematics course designed to allow students to follow courses at a distance, both before and after enrolling at the university. We used clustering techniques on students’ learning behavior, defined for this research as visualizations of course activities and resources, to detect differences in students’ grades according to their online learning behavior. Our results show that, in percentage, students tend to complete a similar amount of both resources and activities. There is no correlation between participation and course grades, although the most active students have higher grades. In addition, the patterns differ significantly depending on the student’s program, demonstrating the importance of a tailored pathway.