Investigating Variation in Learners’ Behavior Through the Lens of Learning Design, Process Mining and Learning Analytics

Abstract

The analysis of data generated by Massive Open Online Courses (MOOCs) platforms has been of great relevance in recent years, since it has made possible to identify patterns of student behavior, success factors associated with course completion, as well as opportunities for improvement in course design. However, data collected without an adequate context does not allow to understand and improve the learning sequences planned in a MOOC. This requires the use of appropriate frameworks to understand and explain the behavior of students in an online course. In this sense, the objective of this work is to address the analysis of student behavior in MOOCs, using Learning Analytics (LA) and Process Mining (PM) techniques to examine the impact of Learning Design (LD) on the participation and progress of students. Specifically, we seek to investigate variations in student behavior in a 6-week programming MOOC. For this, using PM and LA techniques under the umbrella of LD analysis, data from N = 38,838 students enrolled in a MOOC was analyzed. The results revealed that students who passed the course generally spent more time in their study sessions throughout the course, routinely exceeding one hour. They also demonstrated strong engagement with the summative assessments. In contrast, students who did not pass spent less time per session and showed a decrease in the number of weekly sessions, especially from the third week onwards. This emphasizes the importance of designing course content in a way that maintains student engagement and motivation beyond the first few weeks, highlighting the importance of course design in terms of learning sequences that influence course completion.