Sentiment Evolutions in Blended Learning Contexts: Investigating Dynamic Interactions Using Simulation Investigation for Empirical Social Network Analysis

Sentiments evolve over time driven by diverse interactions, especially in blended learning contexts. However, there is scarce empirical research to investigate sentiment evolutions with respect to interactions in the longitudinal data of blended learning. Therefore, by gathering the longitudinal data of five time periods from 38 postgraduate students, this study applied simulation investigation for empirical social network analysis (SIENA) and lag sequential analysis (LSA) to analyze sentiment evolutions with respect to interactions in blended learning contexts. As the knowledge constructions progressed in the blended learning context, students tended to deep interactions and sentiments had the tendency of diversity. Additionally, students experienced the confusion and insightful sentiments accompanied by the deep interactions. In the stable stage of interactions, students may express positive and insightful sentiments along with the agreement interactions. Most notably, although joking sentiments and social-emotion interactions occur less than other types of sentiments and interactions, there is correlation between them.