ScholarWorks@숙명여자대학교

초록

This study aimed to analyze research trends in studies published between 2022 and 2024 in English that measured student engagement using multimodal learning analytics (MMLA). Using a systematic literature review, 33 journal articles and conference proceedings were analyzed based on an analytical framework established based on previous studies. The results showed that most MMLA-based studies focused on proposing measurement models, with deep learning and machine learning as the most frequently used methods. Among the engagement sub-dimensions, cognitive engagement was most frequently investigated, followed by behavioral and emotional engagement, and the majority of studies addressed only one sub-dimension. Regarding data types, visual and facial expression data were the most commonly used. By engagement type, physiological and visual data were frequently applied to measure cognitive engagement, body-movement and visual data for behavioral engagement, and facial expression data for emotional engagement. Based on the main findings, this study derived implications for establishing measurement indicators grounded in the concept of student engagement, diversifying data types and engagement sub-dimensions, and broadening research participants and learning contexts, and suggested directions for future research.

키워드