Exploring the Interplay of Topic Complexity, Emotional Engagement, and Cognitive Engagement in MOOC Discussions: Using Deep Learning and Topic Modeling.

This article focuses on investigating how topic complexity in Massive Open Online Course (MOOC) forum discussions influences learners' emotional engagement, cognitive engagement, and academic achievement. Using a novel two-step methodological approach that combines Bidirectional Encoder Representations from Transformers (BERT) for engagement detection and a Joint Emotion and Cognition Topic Model (JECTM) based on Bayesian networks, the study analyzed over 27,000 discussion posts from a psychology MOOC involving 2,857 learners. Findings indicate that higher topic complexity correlates with increased higher-order cognitive engagement but also with more confusion and negative emotions, while lower complexity topics elicit more positive emotions and lower-order cognition. Importantly, learners who engage with high-complexity topics while maintaining positive emotions and higher-order cognitive engagement are more likely to achieve academic success. The study offers implications for optimizing MOOC discussion design and suggests potential applications in classroom dialogue and AI tutor interactions.