Under Review at Information Systems Research Please do not cite without author permission Agent-Based Modeling to Inform Online Community Theory and Design: Impact of Discussion Moderation on Member Commitment and Contribution

In this article, we advocate a new approach in theory development by translating and synthesizing insights from multiple social science theories in an agent-based model to understand challenges in building online communities. To demonstrate the utility of this approach, we use it to examine the effects of three types of discussion moderation in conversation-based communities: no moderation, in which all members are exposed to all messages, community-level moderation, in which off-topic messages are deleted for everyone in the group, and personalized moderation, in which people see different messages based on their interests. Our results suggest that personalized moderation outperforms the others in increasing members’ contribution and commitment, especially in topically broad communities and those with high message volume. In comparison, community-level moderation increases commitment but not contribution. Our results also reveal a critical trade-off between informational and relational benefits. This research demonstrates the value of agent-based modeling in synthesizing more narrowly-focused theories to describe and prescribe behaviors in a complex system, to generate novel theoretical insights that were out of scope for the component theories, and to use these insights to inform the design of online communities.