Pinned
#QOTD. ”In God we trust. All others must bring data.” -- W. Edwards Deming.
Josh Kalla
1,288 posts
- 🚨NEW PREPRINT w/ @dbroockman. We conducted a real-world experiment to study the effects of partisan media. In Sept 2020, we paid people who normally watch Fox News to instead watch CNN. Median person was paid $15/hour for 7 hrs/week for 4 wks to watch CNN. Thread on results 1/
- Important and provocative finding from @seth_j_hill: "when cities raise their minimum wage by 10%, relative homeless counts increase by three to four percent." Full paper at osf.io/z2fqj/
- New at @PNASNews with @dbroockman, @chriscaballero_ & @easton_matty: "Political practitioners poorly predict which messages persuade the public." As we head into the last week of the campaign, a good reminder that political practitioners have poor intuitions as to what persuades
- New working paper: "Which narrative strategies durably reduce prejudice? Evidence from field and survey experiments supporting the efficacy of perspective-getting" (w/ @dbroockman) Draft is here: osf.io/z2awt/ Comments welcome! A brief thread...
- Lots of smart people are dunking on McCain for not supporting paid leave sooner. This strikes me as bad politics. McCain has a compelling story modeling opinion change that she wants to share on national TV. Progressives should be embracing her rather than scoring Twitter points
- Seems like an important update to this claim: using the voter validated CES, if everyone had voted, Trump would have done slightly better tufts-pol.medium.com/trump-built-a-…
You've probably heard the claim that non-voters are Republican and that Trump would have won by more if everyone had voted. All the publicly available data says that's wrong. New from me, Bonica, @rfunkfordham, & Ernesto Tiburcio 1/n - Replying to @j_kallaWe found that shifting people's media slants from Fox to CNN: (1) made them care more about COVID; (2) learn different information about current events; (3) become more supportive of vote-by-mail; and (4) become more negative towards Trump/GOP 4/
- Time-sharing Experiments for the Social Sciences (TESS) is launching a grant for grad students and those <3 years post-PhD. Winning projects will be ~2x as large as standard TESS study. Great opportunity to collect survey experimental data. More details at tessexperiments.org/html/yic2023.h…
- NEW WORKING PAPER Many people warn that LLMs pose threats to democratic societies through their persuasive capabilities. How should we evaluate these claims? In this project, we discuss limitations of existing evaluations and present our own framework (1/6)
- Why am I pro-abundance? Let me tell you a story about a small park across the street from my place. It's got a slide, swings, monkey bars. But it's been stuck in renovation limbo for over 2 years. (1/5)
- Yale Political Science is seeking several faculty (open rank) in American, Comparative, Pol Econ, and/or IR. Broad search, but we particularly encourage those with Japanese or Middle East expertise to apply. #PoliSciJobs apply.interfolio.com/110488
- Replying to @j_kallaThis study shows more than how CNN benefits Dems and Fox benefits GOP. It shows how partisan media hides and filters information from viewers, thereby undermining the electorate's ability to hold their elected officials accountable. 10/
- Oh dear. There have been two major cases of scientific fraud linked to Dan Ariely in the past month. Is Duke University investigating for a broader pattern of fraudulent research? This is serious (1) datacolada.org/98 (2) retractionwatch.com/2021/07/29/pro…
