🚨 Two new post-docs at GW University!! Both affiliated with Institute for Data, Democracy & Politics, one will be more computational, the other more experimental. Links: gwu.jobs/postings/119777 and gwu.jobs/postings/119779 Spread the word!

Currently in FirstView: “Categorizing Topics Versus Inferring Attitudes: A Theory and Method for Analyzing Open-ended Survey Responses.” William Hobbs and Jon Green outline and validate a theory to separate noise from signal in open-ended responses.

Hard to express what a relief it is to have just submitted the first complete draft of my book manuscript to University of California Press (is on Bluesky) 😮‍💨🥳 5+ years in the making. 249 pages. 82,000 words. 0 ChatGPT. 😉 A million thanks to all those who gave me guidance and inspired me along the way! 🙏

Currently in FirstView: In “Crowdsourced Adaptive Surveys,” Yamil Ricardo Velez introduces a methodology (CSAS) that converts open-ended text from participants into survey items and applies a multi-armed bandit algorithm to determine which questions should be prioritized in the survey.

I've launched a new website where anyone can interactively visualize ticket-splitting rates from actual ballots (cast vote records) for any pair of contests on the 2020 Georgia ballot. Check it out! ballots.isps.yale.edu

From our new issue: "Confronting Core Issues: A Critical Assessment of Attitude Polarization Using Tailored Experiments " by YAMIL RICARDO VELEZ (Yamil Ricardo Velez) and PATRICK LIU (Patrick Liu) #APSRNewIssue cambridge.org/core/journals/…

Universities often serve as "anchor institutions" that deeply affect the character of their communities. In a new paper, we (Mike Andrews / Lauren Russell) ask how (and when) the establishment of a college influences local political and civic life. 🧵 osf.io/preprints/soca…

👇New experiments in which we aimed to map the levers and scope of political persuasion with conversational AI models. It was a tremendous privilege to lead on this work alongside the brilliant Kobi Hackenburg. The paper is packed with results and we'd love your comments!

New working paper! 🚨 Are debriefings in misinformation effective? ✅Only specific ones (with fact-checks) reduce false beliefs ⚠️Both slightly reduce belief in true info 🧠Specific ↑ learning 🟢General ↓ perceived manipulation 🔍Both ↑ transparency osf.io/preprints/psya…

Professor Yamil Ricardo Velez's article "Crowdsourced Adaptive Surveys" is available in Political Analysis FirstView: tinyurl.com/bddbmcpt Columbia University Political Science

Though strong open-weight models have been out for over a year, I’m glad that OpenAI has finally made good on its promise. This is great news for science and makes the openness-capability tradeoff less stark.

An LLM chatbot with RAG on party platforms can increase people's knowledge on policy stances but with weaker downstream effects, finds Yamil Ricardo Velez Green & Semra Sevi, preprints.apsanet.org/engage/apsa/ar…

In polling world, “synthetic sampling” (using LLMs to replace human respondents) has been getting a lot of attention recently. Last wk, Ipsos announced a synthetic panel! But does it actually work? Here’s a white paper I wrote: “Your Polls on ChatGPT” report.verasight.io/synthetic-samp…