The role of emotion regulation in predicting emotional engagement mediated by meta-emotion in online learning environments: a two-stage SEM-ANN approach

Abstract Understanding the mechanism of emotion regulation and the formation of emotional engagement can improve online learning persistence and academic performance. This study was set to pinpoint the potential pathways between emotion regulation and emotional engagement through meta-emotion and develop a predictive model for online emotional engagement. The data collected from 302 college students were analysed using a two-stage structural equation modelling-artificial neural network approach. Firstly, the path analysis implied the significant linkages from emotion regulation to emotional engagement through emotional repair. Secondly, the artificial neural network analysis results suggested that emotional repair contributed to the development of emotional engagement most, and the current model predicted emotional engagement with an accuracy of 91.1%. The main contribution of the present study is providing empirical evidence to predict emotional engagement from novel perspectives through a two-stage approach.