Research
I apply computational methods — network science, machine learning, and crowdsourcing — to study political communication and the social effects of technology. My work sits at the intersection of computer science, statistics, and the social sciences.
Effects of Technology on Attitudes & Behavior
How do social media platforms and generative AI shape what people believe, how they engage politically, and how they interact with one another? My recent work investigates the downstream effects of algorithmic curation, platform design, and AI-generated content on human attitudes and behaviors — with particular attention to political polarization, information exposure, and epistemic quality. A key thread of this line of work is the development and evaluation of prosocial interventions that can redirect the harmful effects of technology at scale.
Generative AI & Society
How do large language models and AI-generated content affect trust, persuasion, and information quality? How can we detect and mitigate harms while preserving the benefits of AI?
Selected papers · full list →
- How AI ideas affect the creativity, diversity, and evolution of human ideas 2025
- Plurals: A system for guiding LLMs via simulated social ensembles 2025
- Deep Value Benchmark: Measuring Whether Models Generalize Deep Values 2025
- Seeing like an AI: How LLMs apply (and misapply) Wikipedia neutrality norms 2024
Social Media & Political Polarization
Does exposure to social media increase polarization? How do platform features (feed ranking, recommendation systems, comment sections) shape partisan attitudes and cross-cutting exposure?
Selected papers · full list →
Prosocial Interventions
Designing and testing interventions — platform nudges, conversation tools, framing strategies — that improve the quality of online discourse and reduce the spread of misinformation.
Selected papers · full list →
News Media & Political Communication
News media and social media platforms are critical infrastructure for political communication. I study how news is produced and consumed and how it shapes public opinion, and how similar dynamics — framing, agenda-setting, and persuasion — play out in social media contexts. Using large-scale text analysis, natural language processing, and causal inference methods, my work has examined ideological bias in news coverage, political framing on social media, the spread of misinformation, and election campaign dynamics.
News & Social Media Framing
Measuring ideological bias and political framing in both traditional news and social media through supervised learning, crowdsourcing, and LLMs. Topics include media bias, immigration framing, and how dominant frames shape public discourse.
Selected papers · full list →
Misinformation
Modeling and quantifying the spread of false information — including fake news and AI-generated content — with a focus on the 2016 and subsequent U.S. elections.
Selected papers · full list →
- Limiting the spread of misinformation in social networks 2011
- What happened? The spread of fake news publisher content during the 2016 election 2019
- Toward a better performance evaluation framework for fake news classification 2020
- Misunderstanding the harms of online misinformation 2024
- The dynamics of (not) unfollowing misinformation spreaders 2024
Election Campaigns
How do candidates, media, and voters interact during election campaigns? How does news and social media coverage shape voter knowledge, attitudes, and turnout?
Selected papers · full list →
- Intermedia agenda setting during the 2016 and 2020 US Presidential elections 2024
- Research note: Lies and presidential debates 2021
- Real-time analysis shows that the first debate shifted attitudes among Twitter users 2020
- Attention to Campaign Events: Do Twitter and Self-Report Metrics Tell the Same Story? 2020
Online Social Movements
Social media have transformed how collective action is organized and sustained. Using causal inference methods applied to social media data, I study how movements change participants, what motivates people to join and stay engaged, which recruitment strategies work, and how organizations structure online protest.
Participation & Mobilization
What motivates people to participate in collective action online? How sticky is activism, and what recruitment methods and networks are most effective at engaging participants?
Selected papers · full list →
Causal Effects of Movements
How do social movements change the attitudes and behavior of their participants? We developed new methodology for quantifying movement effects using social media panels constructed retrospectively.
Selected papers · full list →
Movement Organizations
What role do Social Movement Organizations (SMOs) play in online protest movements? How does organizational structure shape the reach and impact of digital activism?
Selected papers · full list →
Quality of Online Discourse
Online political conversations can be spaces for genuine deliberation — or hotbeds of incivility that drive out constructive voices. I study what makes online discussions productive, how platform design affects conversation quality, and how we can build automated tools to help participants and moderators improve discourse. This work has been supported by the National Science Foundation.
Measuring Deliberation Quality
Creating automated classifiers to measure the quality of online political talk across platforms (Twitter, Facebook, Reddit, news comment sections) in real time.
Selected papers · full list →
Platform Design & Conversation
How do design choices — threading, anonymity, moderation — affect who participates and the character of discourse? Studying natural experiments in news commenting systems.
Selected papers · full list →