Explore Student Engagement in the Learning Management System: A Learning Analytics Approach

Students\u27 daily interaction with the learning management system generates millions of rows of data every single day, and the data has tremendous value to expand the understanding of student engagement. The empirical studies relied on students\u27 clicks to study whether a specific type of clicks leads to higher performance, but they failed to explain the relation between clicks and performance, and they did not build measurements based on students\u27 digital footprints. This dissertation study, in turn, moved from clicks to latent constructs built on the theoretical framework, providing a reliable and valid approach to measure student engagement. The present study also investigated the relation between the constructs of student engagement in the learning management system and academic performance, bridging the gap between epistemology, pedagogy, and assessment with students\u27 digital footprint data in learning analytics. This study employed the confirmatory composite analysis, a PLS-SEM approach, with 158 participants\u27 data obtained from the learning management system to address the research question. Instead of utilizing the empirical approach that simply regresses clicks on performance, the present study developed active and passive learning constructs to estimate students\u27 academic performance. By examining both measurement and structural components, the proposed model was meaningful with good fitting and moderate explanatory and predictive power. The results from the model estimation indicated that students\u27 data obtained from the learning management system was capable of measuring student engagement and predicting performance. Contrary to the empirical studies, the passive learning behaviors were no longer important if controlling for active learning behaviors. Active learning behaviors were significantly related to academic performance. While self-regulation behaviors were not a significant predictor of academic performance, they had moderate and significant effects on driving both passive and active learning constructs. The total effects between self-regulation behaviors and academic performance were also meaningful