English

Fair Decisions through Plurality: Results from a Crowdfunding Platform

Computer Science and Game Theory 2025-09-24 v1 Computers and Society Human-Computer Interaction

Abstract

We discuss an algorithmic intervention aimed at increasing equity and economic efficiency at a crowdfunding platform that gives cash subsidies to grantees. Through a blend of technical and qualitative methods, we show that the previous algorithm used by the platform -- Quadratic Funding (QF) -- suffered problems because its design was rooted in a model of individuals as isolated and selfish. We present an alternative algorithm -- Connection-Oriented Quadratic Funding (CO-QF) -- rooted in a theory of plurality and prosocial utilities, and show that it qualitatively and quantitatively performs better than QF. CO-QF has achieved an 89% adoption rate at the platform and has distributed over $4 Million to date. In simulations we show that it provides better social welfare than QF. While our design for CO-QF was responsive to the needs of a specific community, we also extrapolate out of this context to show that CO-QF is a potentially helpful tool for general-purpose public decision making.

Keywords

Cite

@article{arxiv.2509.18343,
  title  = {Fair Decisions through Plurality: Results from a Crowdfunding Platform},
  author = {Joel Miller and E. Glen Weyl and Chris Kanich},
  journal= {arXiv preprint arXiv:2509.18343},
  year   = {2025}
}

Comments

Appearing in the Fifth ACM Conference on Equity and Access in Algorithms, Mechanisms, and Optimization (EAAMO '25)