Fascinating paper on supply-side factors of negativity bias, misinformation and misperceptions. Also interesting measurements of accuracy of reporting and negativity bias of specific media orgs (e.g.
CNN, Fox News).
Conditionally accepted at the APSR (w/ @scottclifford.bsky.social & @patrickpliu.bsky.social):
Why does political information so often change beliefs but NOT attitudes? We highlight the role of belief relevance, or the extent to which beliefs bear on attitudes.
When you collect data online, are the results from humans or AI? In a project led by Booth PhD student Grace Zhang, we estimate the prevalence of AI agents on commonly used survey platforms:
osf.io/preprints/ps...
🧵
I'm also curious about this!
Screenshot of claude just writing a design no trouble
Writing simulations in DeclareDesign just went from "I should do that, but it's kind of a lot of work" to extremely easy
#rstats Here's a useful guide to creating publication-ready #ggplot figures to journal specifications, which is often quite fiddly.
jaquent.github.io/2026/02/crea...
interesting paper on effect of paternity leave reform on gender beliefs and norms. #WinningHeartsAndMinds
www.nber.org/papers/w34862
Russia, Venezuela, Iran, China, the Sahel region, the United States ...
Want to know why state agents carry out brutal repression — or participate in illegal coups?
Our new book "Making a Career in Dictatorship" provides answers — it just got published by @academic.oup.com:
tinyurl.com/ystwm3tf
Paper on statistical power necessary for interaction effects
doi.org/10.1177/2515...
"Acceptable or Not? Understanding Attitudes Toward
Citizens' Discrimination Against Frontline Workers" by @halling.bsky.social, @mathildececchini.bsky.social, & Benedicte Gronhoj shows that language-based requests are viewed as more acceptable than religious ones.
doi.org/10.1111%2Fpa...
dplyr 1.2.0 is out now and we are SO excited!
- `filter_out()` for dropping rows
- `recode_values()`, `replace_values()`, and `replace_when()` that join `case_when()` as a complete family of recoding/replacing tools
These are huge quality of life wins for #rstats!
tidyverse.org/blog/2026/02...
Is meat consumption becoming political? @willemboterman.bsky.social & @eelcoharteveld.bsky.social examine Dutch surveys showing meat eating aligns with right-wing ideology & climate scepticism. Read OPEN ACCESS: buff.ly/HU5Mdec
@polstudiesassoc.bsky.social #polsky #polsci #FoodPolitics
We’re organizing a workshop at Aarhus University. Please share and consider submitting!
🗓️ 13–14 April 2026 | 📝 Deadline: Mon, 16 Feb 2026 (extended abstract) — junior scholars prioritized
🎤 Keynotes: @stefwalter.bsky.social (Univ. of Zurich) & @hhuang.bsky.social (Ohio State)
Survey experiments have become a popular methodology among social scientists. Has it been effective?
In POQ, Rauf et al. study the efficacy of 100 survey experiments. Their results show that a majority of hypotheses were not supported.
Read now: doi.org/10.1093/poq/...
Here comes another aviation analogy.
I sincerely hope that these types of tools will be used to help us do _better_ research first and foremost.
I fear instead it will be used to help us do _more_ research _faster_.
The magic of autopilot is that it helps pilots fly better, not more.
Will you incorporate LLMs and AI prompting into the course in the future? No. Why won’t you incorporate LLMs and AI prompting into the course? These tools are useful for coding (see this for my personal take on this). However, they’re only useful if you know what you’re doing first. If you skip the learning-the-process-of-writing-code step and just copy/paste output from ChatGPT, you will not learn. You cannot learn. You cannot improve. You will not understand the code.
In that post, it warns that you cannot use it as a beginner: …to use Databot effectively and safely, you still need the skills of a data scientist: background and domain knowledge, data analysis expertise, and coding ability. There is no LLM-based shortcut to those skills. You cannot LLM your way into domain knowledge, data analysis expertise, or coding ability. The only way to gain domain knowledge, data analysis expertise, and coding ability is to struggle. To get errors. To google those errors. To look over the documentation. To copy/paste your own code and adapt it for different purposes. To explore messy datasets. To struggle to clean those datasets. To spend an hour looking for a missing comma. This isn’t a form of programming hazing, like “I had to walk to school uphill both ways in the snow and now you must too.” It’s the actual process of learning and growing and developing and improving. You’ve gotta struggle.
This Tumblr post puts it well (it’s about art specifically, but it applies to coding and data analysis too): Contrary to popular belief the biggest beginner’s roadblock to art isn’t even technical skill it’s frustration tolerance, especially in the age of social media. It hurts and the frustration is endless but you must build the frustration tolerance equivalent to a roach’s capacity to survive a nuclear explosion. That’s how you build on the technical skill. Throw that “won’t even start because I’m afraid it won’t be perfect” shit out the window. Just do it. Just start. Good luck. (The original post has disappeared, but here’s a reblog.) It’s hard, but struggling is the only way to learn anything.
You might not enjoy code as much as Williams does (or I do), but there’s still value in maintaining codings skills as you improve and learn more. You don’t want your skills to atrophy. As I discuss here, when I do use LLMs for coding-related tasks, I purposely throw as much friction into the process as possible: To avoid falling into over-reliance on LLM-assisted code help, I add as much friction into my workflow as possible. I only use GitHub Copilot and Claude in the browser, not through the chat sidebar in Positron or Visual Studio Code. I treat the code it generates like random answers from StackOverflow or blog posts and generally rewrite it completely. I disable the inline LLM-based auto complete in text editors. For routine tasks like generating {roxygen2} documentation scaffolding for functions, I use the {chores} package, which requires a bunch of pointing and clicking to use. Even though I use Positron, I purposely do not use either Positron Assistant or Databot. I have them disabled. So in the end, for pedagogical reasons, I don’t foresee me incorporating LLMs into this class. I’m pedagogically opposed to it. I’m facing all sorts of external pressure to do it, but I’m resisting. You’ve got to learn first.
Some closing thoughts for my students this semester on LLMs and learning #rstats datavizf25.classes.andrewheiss.com/news/2025-12...
I was going to just share an excerpt of this great @donmoyn.bsky.social piece, but there are too many excerpts worth sharing. Just read the whole thing. If you care about governance, democracy, and the rule of law, these issues are crucial. open.substack.com/pub/donmoyni...
I consider this piece on information spillover/information equivalence a must-read for survey experimentalists. It's addresses a key limitation for many survey experiments (especially the hypothetical scenario/vignette type) www.cambridge.org/core/journal...
Treatment can also be (new) information focusing on effects on belief updating and downstream outcomes. See also "information provision experiments" in econ lit www.aeaweb.org/articles?id=...
Today I published a replication outlining concerns with "Instrumentally Inclusive" by Turnbull-Dugarte and López Ortega (2024, APSR).
I document seemingly idiosyncratic and ad hoc choices made by the authors that create a pattern of statistically significant results consistent with their theory.
🚨 SynthNet is out 🚨
Researchers propose new constructs and measures faster than anyone can track. We (@anniria.bsky.social @ruben.the100.ci) built a search engine to check what already exists and help identify redundancies; indexing 74,000 scales from ~31,500 instruments in APA PsycTests. 🧵1/3
Selective reporting
I recently had a similar issue and ended up using a solution combing geom_step with a histogram as detailed by @kjhealy.co here: kieranhealy.org/blog/archive...
📄 New WP version out - full overhaul!
The Politics of Evidence Selection (w/ @jesperasring.bsky.social )
Comments welcome!
🔗 osf.io/preprints/so...
Interesting new study on the elusive connection between organizational performance and user satisfaction
After a huge post-election flip in economic perceptions, I thought Democrats and Republicans might be lying to pollsters to send a partisan message — but I was wrong!
New in the Journal of Experimental Political Science (open access): doi.org/10.1017/XPS....
Title: A Justification for 80% Power Abstract: Cohen’s heuristic reason for choosing 80% power (balancing Type I and TypeII errors) conveniently arrives at approximately the same number as an approachwhere one maximizes the marginal gain in power per standard error reduction. Ihave yet to see someone point this out, and this is interesting because it providesa non-arbitrary justification for 80% power.
a derivation of the result
I think this is kind of neat and I don't think anyone else has noticed it (I've looked and I can't find anyone who has) osf.io/preprints/so...
Maybe I should back off "justification" language, but it's at least a remarkable coincidence. I still think someone else *must* have noticed it...
When there is a random way to do something, there is a less random way that is better but requires more thought. In this case, regression models that make no sense don't belong in a multiverse analysis. An inferential regression without a causal justification is like an opinion without reasons.
We need to have a conversation about random seeds. Don't use 42.
blog.genesmindsmachines.com/p/if-your-ra...
We are hiring PhDs and postdocs to work on the ERC project GETGOV, where I am the PI.
We will investigate governing elites since 1789. I am sure that it will be a lot of fun and result in great research!
Postdocs: www.jobbnorge.no/en/available...
PhDs: www.jobbnorge.no/en/available...
