A dialogue model of the resolution of inter-individual conflicts : implications for computer based collaborative learning

As a result of empirical research into the benefits of peer interaction for learning and development, there is growing interest in designing systems to support collaborative learning. How best to do this for classroom learning is still unclear, and the aim of this thesis is to marry theory with empirical study to produce a set of guide-lines for informing the design of computer based collaborative learning. The resolution of inter-individual conflicts has long been proposed as a mechanism for learning in peer interaction and research indicates that in certain circumstances it does lead to learning. How this finding can be employed in the design of systems is problematic because researchers have typically not unpacked the notion of conflict and not explained how they are resolved. To overcome this a dialogue model of the resolution of inter-individual conflicts in joint planning is proposed. There are three key components in this model; the Task Representation, the Task Focus and the Dialogue focus. Using this model I begin to differentiate between three different types of conflict. Conflicts are caused by Task Representation Differences, Intersection Differences or Task Focus Differences . All three types of interindividual difference can be resolved with a set of discourse transactions and a set of internal resolution procedures. I propose that the resolution of all three can facilitate joint planning and subsequent individual planning. Three experiments are reported which investigate these claims. The first experiment shows that pairs performed better than individuals in a planning task and this beneficial effect carries over to subsequent individual planning. The proposed model is used to investigate this facilitative effect. A corpus of inter-individual conflicts is reported. Most of the conflicts identified in this corpus are either intersection differences or task focus differences and evidence is reported which supports the claim that their resolution can lead to learning. The second and third experiments investigate the claims made from the model about task representation differences and evidence was found to support them. The research reported in this thesis is then related to the design of software to support collaborative learning by presenting a set of guide-lines for informing their design. The work also has important implications for developmental psychology and classroom practice.