Cognitive Games for Policy Argumentation

• In Event: 65.077 - Game-Based Assessment and Learning of Argumentation Skills

Abstract

Background: Democracy depends upon an active and informed citizenry that can use policy analysis, civic journalism, and activism—disciplines that all require policy argumentation. This work considers how we might design cognitive games, games with embedded intelligent tutors to teach policy argumentation. To create these games we must overcome several challenges including: (a) defining the skills of policy argumentation, (b) developing intelligent tutoring systems capable of understanding and providing feedback on learners’ arguments, and (c) reconciling conflicts between how tutors and games provide assistance to maximize both learning and interest.

Purpose: This work argues we can design cognitive games for policy argumentation by: (a) defining policy argumentation as a process of understanding the policy outcomes valued by different ideologies, searching for and analyzing evidence, constructing causal diagrammatic representations of policy claims, using causal models to debate the mechanisms by which policy interventions affect desired outcomes; (b) embedding intelligent tutors in fantasy environments, and (c) providing step-level tutoring and reducing penalties (as opposed to entertainment game approaches) to maximize learning and interest.

Intervention: We designed 2 cognitive games for policy argumentation. In Policy World, learners play the role of a policy analyst who must convince a senator to support policy interventions on topics like climate change, while arguing against an unscrupulous corporate lobbyist played by the computer. Learners: search for policy information; construct causal diagrams representing evidence from different sources; and debate interventions, causal mechanisms and evidence. In Political Agenda, learners assist a political talk show host who attempts to provoke the partisan guests of the show to increase the show’s ratings. To assist the host, the player must predict whether each guest will support, oppose, be ambivalent or indifferent to the policy being discussed by reasoning about the guests’ moral values, the cultural attachments of those values, the extent to which the policy supports/opposes each value, and the consistency and conflict between those values.

Setting and Participants: To test these games we have conducted several laboratory and field studies with university students.

Research Design: Several versions of the Policy World game were tested in two randomized, controlled trials, while Political Agenda is being further refined through design-based research.

Data Collection and Analysis: These studies use pre/post tests of students’ policy argumentation skill and self-reported measures of interest.

Results: The randomized controlled trials found that Policy World increased students’ policy argumentation skills, and the intelligent tutoring approach to assistance (step-level tutoring and reduced penalties) increased learning and interest relative to an entertainment game approach to assistance.

Conclusion: Cognitive games provide step-level feedback and minimize penalties in the context of a fantasy environment to increase interest and learning. Cognitive games are thus an effective approach for teaching policy argument and provide a useful tool for developing effective citizens.

Author

• Matthew Easterday, Northwestern University