Quantifying Effective Scalable Governance across Thousands of Online Communities
Abstract
Quantifying Effective Scalable Governance across Thousands of Online CommunitiesOnline communities universally rely upon content moderators and community admins to function as a system of governance by setting and enforcing rules as they see fit. While online communities account for an enormous and increasing share of all human interaction, and have significant offline impact, we lack acomprehensive, data-driven understanding of how moderators decisions affect critical community outcomes such as diversity, safety, and retention of members. Existing work has demonstrated the potential of studying online governance, but has mostly been limited to single interventions or small-scale studies of specific communities. We propose computational methods to quantify principled aspects of governance and outcomes across a massive dataset of 50K+ communities from the popular platform Reddit. Using these methods we will construct a predictive model of online governance, which we will validate through active experimentation and expansion to additional sets of communities from other online platforms.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- May 05, 2021
- Source ID
- N000142112154
Entities
People
- Tim Althoff
Organizations
- Office of Naval Research
- United States Navy
- University of Washington