This study used blockmodel analysis technique in social network analysis (SNA) to explore the subgroups that naturally occurred in a CSCL community.

In SNA, the students within a subgroup have structural similarity in interaction with others, forming a social position, which naturally emerging in the course of students’ interaction in a virtual community. An individual's position in a social context is related to cognition.

Emergence of social positions in the CSCL community is a self-organizing process resulting from a set of underlying factors unknown to us yet, as neither the emergence of the social positions nor their constituents are pre-defined. Self-organizing system goes beyond traditional deterministic causal thinking. Self-organization informs the fact that the eventual state of the system emerges from the way the components within it interact, not stemming from some pre-established external or internal building plan. By the same token, the students’ social position in a CSCL network, which is believed to have impacts on collaborative learning, can be treated as an emerging construct arising from the interactions between network members.

Four sub-groups were found in this case that naturally emerged in a CSCL community. A logistic regression analysis was performed to test whether students' gender and learning achievement could be used to discriminate among the four subgroups. The results indicate that the aptitudes in reading, writing, and numeracy skills were statistically good discriminates of group membership after removal of gender factor as a covariate. This implies that the formulation of the social positions in that CSCL community was associated with their members’ learning achievement.

The significance of this study at least includes: 1) To our knowledge, there is no study thus far in the literature exploring emergence phenomena in CSCL environments; 2) Focusing the sub-groups instead of the whole network elicits our thinking in terms of assessment of structural and process of collaborative learning, which is that whether sub-group or the whole community is appropriate analysis unit reflecting the inherent characteristics of pattern of students’ interaction in a CSCL community; 3) The discovery of naturally emerging subgroups in a virtual community provides the classroom teacher a picture of students’ interactions that might differ from what the teacher observes in the conventional face-to-face setting.