Computer-Supported Collaborative Learning (CSCL) is typically developed and maintained through participants' social interaction, so we assume that interaction qualities are important for social construction of knowledge. For this purpose, using the social network analysis (SNA) method, we first analysed the density of the participants' interaction to evaluate participants' participation rates. Second, we performed a centrality analysis to analyse the power distribution of the participants. Third, we did a cohesive subgroup analysis to identify the solidarity of the participants and a role analysis to identify the social roles of the participants in the network.

The findings showed that the density of interaction among participants was not high, and the teacher's participation had a great influence on the participation of the group. There are, however, substantial differences in the participants' participation activity. The results also indicated that the power distributions among the participants were not even; participants only interacted frequently with several 'core participants'. The study further revealed that participants had different role positions in 'social space'.