News & Updates
My machine learning meta-reanalysis project “What Predicts Support for Political Violence?” with Jack T. Rametta and Alexa Federice is now available on SocArXiv. We analyze 54(!) datasets across the social sciences and find that youth, but not gender, is highly predictive of support for political violence and conditions related experimental treatments in recent data.
My frequent collaborator, Jack T. Rametta, and I will again be teaching our ICPSR Summer Program Topical Workshop titled “Causal Machine Learning for Observational and Experimental Research”. If you’re interested in our CML work, definitely check this class out!
Our related working paper “Causal Forest and Doubly Robust Machine Learning for Political Science” is available on OSF.
Sam Fuller
Postdoctoral Fellow
Machine Learning & Experiments
Polarization & Partisanship
Political Violence
About Me
Hi there! I’m a Postdoctoral Fellow in Public Opinion and Survey Methodology at the Center for American Political Studies in the Department of Government at Harvard University.
My substantive research focuses on the relationship between partisanship as a social identity and anti-democratic attitudes, including gerrymandering and support for political violence. Much of my work here focuses on leveraging measurement and scaling to better capture both the structure of partisanship and these anti-democratic attitudes. Additionally, I am also interested in a wide-array of social science questions, including those in Sociology. I am currently working on a project with Professor Caitlin Patler that leverages machine learning (specifically, Doubly Robust Machine Learning (DRML)) to estimate the causal effects of immigration detention/incarceration on mental and physical health.
My methodological research is primarily focused on creating new, and applying existing, machine learning methods for the analysis of experimental and survey data. Importantly, I care deeply about bridging the gap between methodologists who create theses methods and substantive scholars who would benefit greatly from using these methods. In that vein, I’m currently working on a series of projects with Jack T. Rametta that introduces standardized, easy-to-use ML methods for analyzing experimental data. Our first methodological and substantive papers in this series are currently under review and available here and here on OSF, respectively. I’m more than happy to field any questions and I’m definitely interested in receiving useful feedback!
I have published in journals such as Political Behavior, PLOS One, Social Science Quarterly, Politics & Gender, and Frontiers in Political Science. My previous work focuses primarily on COVID-19 attitudes, behaviors, and policy outcomes. I also have a paper on leveraging multidimensional scaling to measure ideology and populism in Western Europe (I think it’s a neat paper, check it out!).
Research interests
Machine Learning | Experiments | Partisanship | Anti-Democratic Attitudes | Political Violence
Education
🎓 PhD in Political Science | University of California, Davis
🎓 BS in Political Science & Economics | Berry College