Conceptualising Social Space in Cyberspace: A Study of the Interactions in Online Discussion forums

The study introduces an alternative analytic framework for the investigation of online discussion forums. It focuses on the social dynamics occurring in online discussion threads situated within a tertiary e-learning context, and advocated by social learning theories. Online discussion forums are perceived as conducive environments for the evolvement and support of collaborative and socio- constructivist learning. However, the literature reviewed, revealed a growing need for finding empiric frameworks for ascertaining the materialisation of these perceptions. Attempting to address the identified need, the study adopts ethnomethodological notions, complemented by Structural Analysis approach, to produce an alternative analytic frame called the Event Centre (EC) approach for the study of online discussion forums. The theoretical framework chosen in this study enables the investigation of online discussion forums as systems of relations rather than aggregations of individuals. The EC approach enables the visual representation of networks of people interacting with each other and at the same time presenting the content discussed in each interaction. Applying the EC approach to a set of 131online discussion threads, enabled the discovery of social dynamics occurring within the discussion threads. Preliminary investigations of these visually represented dynamics revealed two overarching patterns. One depicting uni directional interactions in which all participants referred to a single message and a second one depicting sequences of interactions organised in chain like patterns. The study suggests that these overarching patterns may imply different perceptions of knowledge as enacted by the participants, and hence possibly reveal different perceptions of teaching and learning through which it may be possible to detect collaborative and social constructivist processes. The study suggests that the visual patterns introduced should be perceived as abstractions of particular events, implying their generalisability and hence possible application to different data sets