Michael Heseltine
Post-Doc at University of Oxford. Congress, text-as-data, and political communications.
- 🚨NEW PAPER Going Local? Localized Rhetoric in Congressional Communications Is congressional comms really nationalized? Analyzing 9M+ social media posts from Members of Congress, I show that local messaging is still common & strategic. New ML + domain sharing data www.tandfonline.com/doi/full/10....
- Centrally, the use of localized messaging varies by party based on partisan control of the White House (and congressional chamber). Governing parties use messaging on localized benefits, opposition focuses more on national political developments. This results in a visually pleasing back and forth.
- Reposted by Michael Heseltine[Not loaded yet]
- 🥳 Our new paper is out at the British Journal of Political Science. We find a strong link between the ideology of candidate messaging and the perceived ideology of these candidates among voters. Evidence that voters can differentiate between candidates and are attuned to what candidates say.
- 📜Preprint!📜 GPT5 replacing GPT4 effectively renders a whole range of existing GPT-reliant research unreproduceable. Open-source LLMs offer a science-friendly alternative, but which to use? I test 28 different LLMS in 7 languages so you can pick the best models for your task. osf.io/preprints/os...
- Reposted by Michael Heseltine[Not loaded yet]
- Reposted by Michael Heseltine[Not loaded yet]
- 🚨New at AJPS to start off the year! Do partisans change their news consumption based on daily news sentiment about their team? Democrats read more news on bad days for Republicans and less on bad days for Dems. Republicans turn to more conservative news on bad days for Republicans.
- Reposted by Michael Heseltine[Not loaded yet]
- My article "Asymmetric Polarization in Online Media Engagement in the United States Congress" is published in the latest edition of IJPP. Republican media engagement and sharing on social media has increased in extremity over time, but not Democratic. Clear Trump effect. doi.org/10.1177/1940...
- ✏️New in Political Communication ✏️ w/ great coauthors. We examine perceived polarization in the US and Poland. Both exposure to partisan news and exposure to general coverage with polarization reporting frames increase perceived personal distance from other side. www.tandfonline.com/doi/full/10....
- 🚨New paper in R&P: Testing GPT-4 on a range of common political science coding tasks in 4 languages. We find generally strong results at a fraction of the time and cost. We also introduce an effective double-code and validate approach for LLM coding. journals.sagepub.com/doi/10.1177/...
- 🚨Was the COVID-19 pandemic a rhetorically unifying or divisive event for elites in the US? I test this at the federal, state, and local level. Initially COVID was a unifying event, but quickly became a divisive topic, especially for Congressional Republicans. journals.sagepub.com/doi/10.1177/...
- New paper in IJPP 🚨 How do Members of Congress engage with the media online? From news link sharing and media mentions from MoC over the last 10 years I find clear asymmetric polarization in news media engagement over time, driven by Republican Members... Polisky journals.sagepub.com/doi/10.1177/...
- 👍 Across tasks and languages, we test the viablility of LLMs like GPT-4 to replace human coders for machine learning training data, with mostly strong results! Negligable downstream effects on trained transformer models. We also offer a suggested pipeline for researchers. Polisky osf.io/cx752
- Reposted by Michael Heseltine[Not loaded yet]
