We actually show in the model that because our best estimates of statistical power are fairly low, even if literally 100% of our hypotheses are in fact correct we still must have extreme selection against nulls

Yes, but have you considered that I — a very normal Bluesky user — have a strong belief that AI is bad? I am very smart

This is it. I guess we could do a diff-in-diff on the role of diff-in-diff on the evalaution of outcomes, before and after the diff-in-diff revolution.

When I pitch academics on my paper on nulls one common and understandable reaction is "but they're probably noisy and thus uninformative nulls." This is true, but it misses the key realization that WE PUBLISH THE RESULT WHEN THE NOISY TEST IS P<0.05.

We spent months training grad student RAs and GPT-5 mini still beat them by a lot

This is why I was praising JEPS a few days ago. What they've done is impressive.

LLMs are very good at extracting information from academic articles. They are much better than even highly trained humans (our grad RAs had hundreds of hours of practice). And of course they're ~1000x cheaper and faster

I have a new paper. We look at ~all stats articles in political science post-2010 & show that 94% have abstracts that claim to reject a null. Only 2% present only null results. This is hard to explain unless the research process has a filter that only lets rejections through.

I just love this image for teaching. "There is only one test" (image from Modern Dive, but original from @AllenDowney, allendowney.substack.com/p/there-is-o...)

I made little pandan crème brûlées. If you’re okay with a more rustic look, you cannot beat the ease and consistency of doing them sous vide in mason jars.

Finish up a paper guesstimating selection on significance across all journals in political science and the Journal of Experimental Political Science does not get nearly enough credit. They've created the best institution in the discipline for publishing unbiased research.

Causal inference nuts: Radiolab has a great podcast this week about the pitfalls of using correlational evidence in policymaking. The empirical case is the self-esteem movement of the 1980s. Would be a good listen for those who teach causal inference.

Crazy snowstorm means throwing the kids outside with cookie rewards

We're hard at work on new AI tools for RStudio that will support EDA and we hope generally accelerate data science. If you're interested in trying them out, join the private beta waitlist for a sneak peek: posit.co/products/ai/ #rstats

blog post: Total Margin of Error margin of error = 2 x (reported margin of error) and how much of this error is "bias" vs "variance" ?

NEW PAPER! We know that lead exposure is bad, but where is it all coming from? We show that used lead-acid battery recycling is likely much more important than previously thought - perhaps 1/3 of exposure in LMICs. LINK: www.cgdev.org/publication/... w/ @theomitchell1.bsky.social & James Hu

Waterloo: snow day Wilfred Laurier: snow day McMaster: snow day Guelph:

On Feb 16 I have a spare evening around the SF to Palo Alto area. Where should I eat?

I just finished a three-year term as an editor at an international relations journal. I began at the start of the LLM era but ended right in the middle of it. Our volume of submissions tripled and our desk reject rate rose to 75%. I have some thoughts. open.substack.com/pub/hegemon/...

100%

Working with big spatial data sets in #rstats? You should try {duckspatial}. The dev version of #duckspatial (soon on CRAN) uses #duckdb to perform super fast and memory efficient spatial operations cidree.github.io/duckspatial/... In a benchmark against, {sf}....

Some news: we're committing $175m to @GiveWell in 2026, growing our >$1B committed to date. Based on GiveWell’s estimates, we think this funding has saved >100,000 lives so far. 🧵 coefficientgiving.org/research/al...

My latest: Contrary to popular perception, American antisemitism is not primarily a partisan problem—it is a generational one. "The research collectively suggests that America is becoming more anti-Semitic because its young people are becoming more anti-Semitic." Gift link:

I can’t get past this. The (imperfect) analogy in my head comes from living in Rio: you’re always in a state of soft alert, which becomes a permanent background noise to life, bc you know that, even if low probability, very bad things could happen at any moment, and you must know how to respond.

I promise if you’re not Jewish and don’t know any Jews quite well you really do not understand the level of justified paranoia baked into most public Jewish gatherings: Security at swimming lessons at the JCC, off duty police outside a bar mitzvah as a favour, etc. It shocked me.

Synagogues were safe until they weren’t. JCC were safe until they weren’t. Public beaches, public parks…risk is something I think about every single time I take my family to school, pray, exist as Jews

Again, even though the Japanese government leading up to 1945 committed terrible atrocities, none of us think that Japanese internment was caused by the Japanese government. It was caused by American racism.

2 weeks ago, 4 blocks from my synagogue local police had a shootout with a man holed up in his house with an AK47. It took multiple armored vehicles to breach the house and I could hear the shooting and concussion grenades. The man had a nazi flag and other material in his house.

The chabad rabbi killed at Bondi Beach

Antisemitism has obviously been getting worse for years. People in my social group tend to downplay it in various ways and those with power like the mayor of my city have mostly ignored it.

It’s Boulevardier season!

New paper! @william-dinneen.bsky.social @guygrossman.bsky.social Yiqing Xu and I use GPT to code 91k articles from 174 polisci journals (2003–2023)and track research designs, transparency practices, and citations. How has the credibility revolution reshaped the discipline? doi.org/10.31235/osf... 🧵

Yeah bsky.app/profile/ryan...

preparing some slightly deranged slides for last class

A student recently asked me for academic job market advice and I pulled up a slideshow from a few years ago. I don't think I've shared it, but it might be broadly useful. I think the advice almost entirely holds up. First part is about my time on the job market 1/3

So then the researchers divided the academics into two groups, the Eagles and the Rattlers <sorry> Twitter and Bluesky, and then

Some people are unaware but the ideal wok is made of steel and costs about $15 at your local Chinese grocery store. Every time you use it grows in power.

Adding a pinch of sugar to food should be way more common in western home cooking

At the Ag Fair & @Uofg has a presence. I was talking to the DNA barcoding people & found out that Steve Newmaster has been effectively banned from teaching & research but still gets his salary. If we can’t fire people for research fraud, what are we doing? www.science.org/content/arti...

Help me think of an engaging physical way to demonstrate the problem of selection on statistical significance. My best idea right now is to take a Galton board and drill holes in the bottom so some of the balls with |z| < 2 fall out.

Like 500 social scientists just got a Nature publication today. This will eventually be a DID for someone.

Say I wanted ~9k USD to buy out the teaching time of a prof (a brilliant coauthor of mine) so he can work on creating a validated ground truth dataset for a challenging data extraction and labelling task that current public LLMs (e.g. GPT5) have not yet saturated. Who do I pitch?

Snacking cakes by Yossy Arefi is the most dangerous cookbook. It’s entirely 1 bowl cakes that you can make in under 1 hour, like this chocolate sesame one. Perfect for when you need to eat cake at 9pm on a Sunday.

Such a happy example of people practicing evidence-backed living: www.nytimes.com/2025/10/20/w...

Quarto friends! I need your help: We are implementing a new QMD parser in Quarto. It will be super nice. But it's a big change, and we want to minimize the impact. 1. Can you share a link to your Quarto project at github.com/quarto-dev/q... 2. repost this and let your Quarto friends know too?

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...

Great news on @malengo.org: With a new large investment from The Shapiro Foundation (theshapirofoundation.org), we'll be able to bring several hundred vocational training and university students from Kenya, Uganda, and Rwanda to Europe over the coming years! More here: malengo.org/malengo-secu...

I'll be in SF (well, Mountain View) February 16ish to 22ish. Who should I meet? What should I be doing? My time is booked Feb 20-22, but I'll have time before. Relatedly, anyone want me to talk on statistical power or using LLMs to code ~all political science articles?