Leveraging Expert Insight to Define Multifaceted Student Engagement Levels in Higher Education Online Learning Using a Belief Rule-Based Framework

Abstract

Educational Data Mining (EDM) is an interdisciplinary field that aims to address education issues by using vast amounts of educational data. Meanwhile, the Ministry of Education Malaysia desires to promote online education to raise teaching and learning standards and increase efficiency. Therefore, it is crucial to increase student engagement for improving the teaching and learning process, institutional effectiveness, as well as EDM research. Although student engagement is getting more important with the growth of online learning, there is no clear consensus to identify and differentiate student engagement into distinct levels. Moreover, student engagement is a multidimensional construct that is made up of behavioural, cognitive, emotional, and social engagement. The purpose of this paper is to collect and analyse the professional view and opinion of three educational experts on student engagement so that they can be utilised to develop the rules that will distinguish between levels by implementing a belief rule-based system. According to the experts' findings, it is critical for the lecturer to identify students who are having problems engaging with online learning early on so that they can be helped to engage more fully and, as a result, do better academically. Besides, prediction of student engagement level depends on multiple dimensions, while statistical methods like mean, median, and quartile can be utilised to translate numerical data into verbal terms. Lastly, the best time to extract data for the student engagement level prediction is the week before mid-semester break.