I am a post-doctoral associate at New York University's Social Media and Political Participation (SMaPP) Lab, and received my PhD in political science from the University of Toronto in 2018. My research finds itself at the intersection of political behavior, public opinion, social media, and statistical methodology.
I am also a partner at Vox Pop Labs, a civic technology and public opinion research firm. We work with media organizations (e.g. Wall Street Journal, Sky News UK, CBC, ABC) to provide civic engagement applications to promote political and social science literacy. We are best known for Vote Compass, an application used by millions of voters in recent elections.
How Many People Live in Political Bubbles on Social Media? Evidence from Linked Survey and Twitter Data
2019. SAGE Open (special issue on social media), January-March: 1-21.
(with Jonathan Nagler, Andrew Guess, Joshua Tucker, and Jan Zilinsky)
The Statistical Analysis of Misreporting on Sensitive Survey Questions
2017. Political Analysis, 25 (2): 241-259.
Statistical software | Software vignette | Replication data
Trying to understand Jeff Flake? We analyzed his Twitter feed — and were surprised
2018. Washington Post. (with Jan Zilinsky, Jonathan Nagler, Joshua Tucker)
Comparing Trump to the greatest—and the most polarizing—presidents in US history
2018. Brookings. (with Brandon Rottinghaus and Justin S. Vaughn)
Vox Pop Labs is a political engagement and public opinion research firm that brings social science to the public through civic education applications. Our partners include Vox.com, the Wall Street Journal, Sky News (UK), the Canadian Broadcasting Corporation (CBC), RTL (Germany), France 24, the Australian Broadcasting Corporation (ABC), and TV New Zealand. Our applications include:
Vote Compass is an electoral literacy application that educates voters about their place in the ideological landscape and provides them with information about the positions of the parties on a wide array of issues.
The Political Sentimeter is a family of applications that fit users into a discrete set of political, social, and identity-based archetypes or classes using a mixture model fit to data from large-scale surveys.
The Signal is a Bayesian dynamic linear election forecasting application developed for the most recent Canadian federal election, which ran with the Toronto Star and L'actualité.
gregory.eady can be contacted @nyu.edu or @gmail.com