Investigating Demographics and Behavioral Engagement Associated with Online Learning Performance.

In recent years, online learning has become a viable alternative for learners worldwide to pursue higher education and gain advanced technical skills. In this work, we focused on data analysis to scrutinize the features associated with online learning performance and course selection. In particular, we investigated and compared how student demographic characteristics and behavioral engagement associated with academic performance based on a publicly accessible Open University Learning Analytics dataset (OULAD). We find that neighborhood poverty level, education background, active learning days and interaction times are positively associated with final learning results. In addition, students with different genders had bias in online course selection, where female students tended to favor social science courses and male had a preference for STEM. Students who performed well mainly came from learners with a well-educated prior background.