Effects of personalised feedback approach on knowledge building, emotions, co-regulated behavioural patterns and cognitive load in online collaborative learning

Learning analytics has been widely used in the field of education. Most studies have adopted a learning analytics dashboard to present data on learning processes or learning outcomes. However, only presenting learning analytics results was not sufficient and lacked personalised feedback. In response to these gaps, this study proposed a learning analytics-based personalised feedback approach and examined the effects of the proposed approach on collaborative knowledge building, emotional status, co-regulated behavioural patterns and cognitive load. The learning analytics-based personalised feedback approach adopted a deep neural network model, namely Bert (bidirectional encoder representations from transformers), to automatically classify discussion transcripts in online collaborative learning. In total, 60 undergraduate students participated in this exploratory study and were randomly assigned into experimental and control groups. The students in the experimental group learned with the learning analytics-based personalised feedback approach, and the students in the control group learned with the traditional online collaborative learning approach. The learning analytics-based approach was found to have significant impacts and no significant difference in cognitive load was noted between the experimental and control groups.