Feedback systems and control
Studying how feedback, incentives, and control structures shape collective outcomes.
Online communities and platforms rely on feedback mechanisms—such as ratings, incentives, moderation, and governance—to coordinate behavior at scale. This project studies how feedback systems shape collective outcomes: when they improve learning and coordination, and when they introduce distortions, bias, or inefficiency.
We conceptualize feedback as a form of control. Feedback can guide exploration, stabilize norms, and align individuals toward shared objectives—but it can also amplify popularity bias, create runaway cascades, and suppress innovation. Our goal is to identify general design principles for robust feedback systems that perform well across tasks, populations, and institutional settings.
What we do
- Design controlled experiments that manipulate ratings, incentives, and social information to isolate their causal effects on behavior.
- Measure how feedback reshapes exploration, convergence, efficiency, and inequality in collective systems.
- Model feedback loops using computational approaches to predict when systems produce effective coordination versus systematic distortions.
- Connect experimental results to problems of online governance, institutional design, and collective intelligence.
Why it matters
Feedback systems underlie modern social platforms, markets, and institutions. Understanding how feedback shapes behavior is essential for designing systems that promote learning, fairness, robustness, and innovation, rather than polarization, manipulation, or inefficiency. This work informs the design of online platforms, recommendation systems, and governance mechanisms, with implications for social media, education, science, and digital democracy.
Related publications and links
- placeholder: feedback systems paper
- placeholder: experimental codebase
- placeholder: datasets
- placeholder: online demos