Proceedings of the Annual Meeting of the Cognitive Science Society

Many educational approaches assume that making the learner active leads to better learning although this improvement inlearning has not be firmly quantified experimentally. The goal of this paper is to articulate a model of agency in cooperativelearning based on a predictive cognitive architecture and to explore methodological strategies as well as theoretical andapplied implications of agency in the study of cooperative learning, in this case with data on emotional processing. Resultsfrom 27 dyads (1 player and 1 watcher) who played a serious game for learning physics for 120 minutes show that agencyhas no effect on the overall quantity of emotional processing, but that the emotional processing of a watcher and playeris synchronized. A watchers emotional processing may precede or be delayed from the players. The cornerstone of thisframework is the notion of predictions, which unites top-down and bottom-up influences as modulated by the possibilityfor action (agency). The model presented is the foundation for process-oriented studies of cooperative learning from aneducational neuroscience perspective.