How do social networks and voting incentives interact to produce and sustain political polarization? This work combines formal models, network simulation, and public writing to examine the mechanisms behind ideological sorting.
The median voter theorem predicts convergence — so why does polarization persist? Bringing abstention and ideologically motivated voters into the model changes the math, and the predictions, considerably.
Why do online communities sort themselves into echo chambers? A model of how cognitive biases interact with network structure to drive polarization in public discourse.
A public-facing overview of the echo chambers research — what the model shows, what it means, and why it matters for understanding online political discourse.
How do things — technologies, behaviors, information — spread through networks? And how do we measure the properties of those networks accurately? This work addresses both questions, from co-diffusion dynamics to sampling methodology.
How do technologies rise and fall together? Examining co-diffusion patterns on GitHub to understand which tools complement each other and which compete — with implications for how innovation spreads.
GitHub →A new method for estimating how much birds of a feather flock together online — using dyadic predictions to measure homophily even when demographic data is incomplete.
Using machine learning classifiers to infer group-level demographic properties in online networks — a practical tool for studying populations where direct measurement isn't possible.
Respondent-driven sampling is widely used to study hard-to-reach populations — but how biased are its estimates, and why? A systematic decomposition of where the errors come from.
How do people coordinate, compete, and enforce norms — especially in communities defined by their informal rules? This work spans professional hockey, organizational diversity, and the sociology of stigmatized careers.
How do players build careers around fighting in a sport that's slowly eliminating it? Tracing the enforcer role across 15 years of NHL data to understand how deviant specializations achieve legitimacy.
GitHub →How do attachment preferences in diverse teams shape collective outcomes? A formal model of how individual-level choices aggregate into system-level performance in problem-solving groups.
Seven decades of NHL fights mapped as a social network. What emerges is a surprisingly structured system of informal enforcement — with implications for how communities regulate themselves without formal rules.
GitHub →How do careers built around stigmatized activities — like professional fighting — acquire social structure and legitimacy over time? The dissertation that launched the hockey networks work.
How do shared expectations emerge in small groups, and how do they enable coordination? An examination of the micro-level mechanisms behind collective behavior.
GitHub →Online reviews as informal sanctioning — what drives people to take the time to write them? An analysis of how service characteristics shape review behavior on platforms like Angie's List.