Predicting the Emotional Engagement in Online Learning: A Hybrid Structural Equation Modeling-artificial Neural Network Approach

Abstract

Emotional engagement is identified as a critical issue for assessing online learning performance. With the increasing attention to emotional engagement in online learning, there has been a concomitant rise of needs and interest in investigating the prediction of it. As significant antecedents of emotional engagement, the study explored the effects of the emotion regulation and meta-emotion on online emotional engagement by newly employing a hybrid structural equation modeling-artificial neural network (SEM-ANN) approach with data of 302 students. The SEM analysis indicated that emotion regulation was highly related to meta-emotion, and two dimensions of meta-emotion, emotional clarity and repair were significantly associated with emotional engagement. Besides, emotional repair also fully mediated the link from emotion regulation to emotional engagement. Likewise, the ANN model suggested that emotional clarity and repair predicted emotional engagement, with more than 71% accuracy. The findings implied the critical importance of paying attention to learners’ meta-emotion for promoting emotional engagement and academic achievement. Theoretical and practical implications are discussed.