e-Learning Interactions and Academic Outcomes: an Analysis of Undergraduates in Sri Lanka





e-learning interactions, student behaviour, learning analytics, online learning


The shift to online teaching and learning following the onset of the COVID-19 pandemic resulted in a large increase in the usage of online learning platforms. Student interactions with these platforms provide an important source of information on student progress on a course during periods of distance learning. However, in resource constrained settings where students face difficulties in accessing stable and quality internet, it is unclear to what extent interactions with the learning platform influence academic outcomes, especially for programmes that were not originally intended to be delivered online. This study makes use of data on four cohorts of undergraduate students at the University of Moratuwa, where each cohort was exposed to different periods of online teaching and learning. The data covers interactions with Moodle course pages and assessment marks for multiple courses with diverse subject contents offered in two faculties. Interactions with the learning platform are measured using clicks on different types of objects on the course page as well as the distribution of clicks over the course of semester. To account for the confounding effect of prior ability on the relationship between learning platform interactions and academic outcomes, we also use results from courses that the students have taken a priori. We find that while the switch to online teaching and learning led to a dramatic increase in the level of student interactions with the learning platform, the level of interaction has remained above the pre-COVID levels even after resuming on-site, face-to-face delivery. While differences among the subjects and cohorts exist, results from a multiple regression model suggest that the association with certain types of learning platform interactions and academic outcomes are significant even after controlling for prior ability and differences in course design. Specifically, the volume and consistency of access both improve outcomes with the timing of clicks more strongly associated with the final examination mark whereas the type of content clicked on is more associated with continuous assessment marks. The findings have important implications for the continued adoption of blended learning methods and course design for the programmes under study.