Effects of collaborative filtering‐based peer recommendation mechanism on in‐service teachers’ learning performance, knowledge construction, and social network during online training

Background Collaborative learning has become a crucial approach to promoting online in‐service teacher training. Appropriate peer recommendation for group composition is the basis to ensure productive learning outcomes of collaborative learning. However, there is a lack of understanding of the impact of peer recommendation on in‐service teachers' online training. Objectives Therefore, based on the collaborative filtering approach, a peer recommendation mechanism for in‐service teachers was constructed in this study. We investigated the effects of the constructed mechanism on in‐service teachers' online learning performance, interactive knowledge construction behavioural patterns and social networks. Methods In this study, 82 in‐service teachers were recruited to participate in the study. Participants under the experimental condition (n = 41) were invited to apply collaborative filtering‐based peer recommendation mechanism for grouping, while participants under the control condition (n = 41) were invited to use random grouping method. Participants' interaction data were collected from online collaborative discussion activities in a 5‐week asynchronous online course. The pre‐post knowledge test, content analysis, lag sequential analysis and social network analysis were used to analyse the differences between groups. Results and Conclusions The following findings were revealed: (1) the experimental group performed better than the control group in terms of online learning performance; (2) interactive knowledge construction behavioural patterns generated in collaborative activities showed that the experimental group achieved deeper level of knowledge construction in collaboration than the control group; (3) social networks generated in collaborative activities showed that the experimental group had tighter interaction relationships than the control group. Therefore, the peer recommendation mechanism could be useful to improve peer recommendation for group composition and had positive effects on in‐service teachers' online training. Implications This study can shed light on the construction of peer recommendation mechanism in recommending appropriate peers for in‐service teachers and different learners in online collaborative learning.