Deliberation occurs whenever people come together to discuss an issue, weigh alternatives, or make a decision. As people engage in deliberation, they form opinions about the topic, each other, and the value of dialog for resolving disagreements. Deliberative groups can either contribute to polarization, strengthening pre-existing opinions, or they can provide an opportunity to build empathy and tolerance as individuals see and learn from people with different perspectives.

With this project, Duncan Watts’ Computational Social Science Lab at the University of Pennsylvania, aims to understand how to reduce polarization in deliberative groups and populations. This is no trivial task. Previous work has shown that the effectiveness of depolarization interventions can vary greatly depending on factors such as group size, demographics, group history, and the topic of discussion. An intervention that is useful for one group and topic may be counterproductive for another. However, the existing literature has not addressed how to systematically adapt these interventions to specific contexts.

To make progress, the project team proposes to develop a novel platform and associated methods for conducting deliberation experiments that take context into account. They also note that to fully realize the potential of this platform, they need to work with a community of researchers to explore this space and document findings.

With the first phase of their project, also funded by TWCF, the team built the infrastructure and conducted high-throughput experiments over a full domain of deliberative contexts using these tools in their own experiments and with other researchers. With this grant, they aim to build an open science platform around this infrastructure which a community of researchers can coordinate to conduct experiments and share data, expertise, and statistical models. They also plan to engage researchers to create a collaborative network that can benefit from their platform.