Investigating the Learning Performance in Computer Supported Collaborative Learning Environments

This thesis concerns groupwork, Computer Supported Collaborative Learning (CSCL) and social relationships. The use of the computer, especially when it involves the web, is claimed to be one of the most powerful tools for providing teachers and learners with an interactive and independent learning environment. This claim is justified by the immediate and wide accessed to resources. Although CSCL involves many technologies and functions, it is agreed that its universal feature is to encourage students to seek in-depth learning. The main purpose of this research is to empirically investigate the influences on learning outcomes in CSCL environments, specifically to understand how affordances for collaboration contribute to user experience as well as performance in groupwork; and to study social relationships and how they may affect learning performance. The main motivations behind this research are: 1) contradictions in the literature about the effectiveness of using the technology in groupwork, and 2) the shortcomings of existing collaborative environments, such as a poor sense of presence and limited non-verbal communication. Evaluations of collaborative technology have tended to follow either an ethnographic approach to investigate the context of use in depth, or more focused experimental analyses directed towards specific questions about collaboration. However, this research followed the mixed methods approach which has been successfully applied in HCI (Murphy et al., 1999; Ormerod et al., 2004), so this approach is suitable for investigating CSCL affordances and requirements. A series of seven field studies was conducted, using both quantitative (questionnaires) and qualitative (observations and interviews) methods. Synthesising the analysis of the seven studies involved experimentally comparing the affordances of some existing collaborative technologies (Blackboard and SecondLife). Overall, the results offer four main contributions. First, a conceptual model of the factors that impact performance in CSCL environments is developed, including three main dimensions: technology, group and learner features. Second, the key theoretical findings in this research show that social relationships and overall group activities do not correlate directly with performance, so our results appear to agree with previous findings that social relationships have no positive effect on learning performance. However, some social familiarity does appear to promote group interaction and performance. Comparing the use of technologies with face-to-face collaboration produced a complex picture. The 3D virtual world did not produce the expected benefit, probably because of usability problems encountered with the avatars. In contrast, the text-based virtual world was perceived as being more usable, even though for some groups it was considered to be boring and not a stimulating user experience. Although face-to-face collaboration was expected to be most effective, and indeed it was quickest and rated best on experience and positive emotions, it did not produce more accurate results. Third, was the mixed methods research approach and the discourse analysis method used to analyse the Blackboard threads in this research. Finally, the research provides guidelines for both educators and designers of CSCL environments. Although the exploratory nature of the study resulted in certain limitations, the study enriches existing knowledge in the area of CSCL and provides theoretical, methodological and practical insights that suggest promising opportunities for future research.