Tracking Students' Unseen Status in the Virtual Classroom: A Systematic Review

Abstract

Online learning has broadened learning opportunities but also brought with it a number of challenges to authentic student learning. A significant issue is that of the "invisible learner", the one who is attending the class but is not easy to evaluate in regard to attention and participation. This review presents an overview of world literature to identify the phenomenon of ‘unseen students’, to discuss methods for tracking engagement in low-resource settings and to explore technological solutions for engagement detection. This study followed the PRISMA framework, where 150 studies were first identified, followed by 15 studies that were included in the final study. The results suggest that the more sophisticated engagement detection models (CNN and LSTM) that use visual information are highly accurate. Stability, on the other hand, is a continuing problem with the internet, power, and limited device access to today's invisible learners. The review also finds that a combined monitoring method is necessary, which combines deep learning, lightweight recognition technique approaches and multimodal analytics. However, ethical and pedagogical principles need to inform the implementation of such systems to ensure that monitoring benefits educational quality and inclusivity.