Proceedings of the Annual Meeting of the Cognitive Science Society

Research on human collaboration has suggested thatknowledge diversity improves group performance in complextasks such as design, problem solving and forecasting.However, in educational settings it is important to also askwhether learning and transfer for individuals within the groupis enhanced or hindered by diversity in collaborative workgroups. We compare performance in a transportation networkdesign task for two types of collaborative groups, andcompare their performance to that of individuals. In onegroup condition (Distributed Knowledge) each dyad memberhas been trained on a different subtask of a complex jointdesign problem in advance of the collaborative activity. Thesedifferent training tasks should predispose the two dyadmembers to adopt different perspectives, issues, and designstrategies, thus generating greater cognitive diversity for thegroup. In the other group condition (Shared Knowledge) bothdyad participants experienced the same training involvingboth subtasks. Task performance results show a group versusindividual advantage in performance, but a non-significantdifference in performance between the two group knowledgediversity conditions. The group knowledge manipulation didaffect group process, as measured by time spentcollaborating, number of turns taken, and number of wordsspoken. The findings suggest that group diversity canpromote individual learning and transfer when sufficient timeis allowed for discussion and group work.