With Deborah Mayo, Kent Staley, and Wendy Parker we are editing a special issue of Synthese on "Severity and Learning from Error": link.springer.com/collections/... Deadline for submissions is the 15th of December, 2026. If this is a topic that you are interested in, consider contributing!

This is a correct argument, but in sociology it is abused by ASA Council to prevent its journals from having policies for quantitative research.

this week marks 10 years since "milkshake duck" was coined

“We provide the first large-scale, multi-paradigm evidence that people systematically overestimate the dishonesty of others.” New work by @jareef.bsky.social, Helge Thorbjørnsen, and @hallgeirsjastad.bsky.social Open Access: doi.org/10.1016/j.je... #SocialPsyc #AcademicSky 🧪

Wimsatt: Perhaps, just perhaps, the best explanation of a phenomenon at a given level is to be found at the same level. Kendler: Noone would seek to explain hypertension at the level of quarks. A few of my favourite papers on this topic of absurd biological reductionism: 🧵

🚨 A few months ago, a paper published in Science reported that “LLM adoption is associated with a large increase in researchers’ scientific output.” In a comment released today, we show that the reported effects are driven by a methodological issue in the empirical design. arxiv.org/abs/2605.17979

What was that I said about applied econometrics bing hard to publish in econometrics journals, even the ostensibly "applied" one(s)? www.sciencedirect.com/science/arti...

Interroger des LLMs plutôt que des humains, ça marche? Nous avons montré que les LLMs ne parviennent pas à prédire la valeur moyenne d'une variable dans une sous-population, que chaque modèle a une faible variance dans ses réponses, mais forte entre modèle journals.sagepub.com/doi/10.1177/...

This study of four cohorts of French schoolchildren shows that the gender gap in math (girls falling behind) begins essentially immediately as soon as kids start school. They use the natural variation in kids' ages at school start to show that it's school entry, not age, that triggers the change.

Hallelujah - my personal take on "real world evidence/data" (as I see it used) is that it is a euphemism for data missing not at random

Interesting new paper on AI coding tools and productivity on GitHub. Large gains at the code-writing stage are partly offset by downstream human bottlenecks in code review and deployment. Link: www.nber.org/papers/w35275

Why do humans punish? Our new paper, with @manvir.bsky.social, is out in Psychological Science! With experiments in a small-scale society (Mentawai) & the US, we test whether punitive cognition evolved to deter wrongdoing or restore fairness. link: tinyurl.com/msrmu7cc pdf: tinyurl.com/mvprz8fb

Now listening to another book by the same author & I actually like this one a lot better. It’s about psychosomatic illness, which seems to be *precisely* her domain of expertise, and the stories are wild. Also lots of emphasis on how ppl are not faking anything, the impairment is very real.

“I can’t believe the power pose discourse is still going on, 2026 edition”

TIL heat is really killing a lot of people in Europe. hannahritchie.substack.com/p/heat-guns-...

On the year's hottest day, a thread on why this infographic misleads (1/n)

New estimate for residual bias after study registration just dropped

🪈🐍 « Basée sur l'IA générative, elle repose sur un principe de mise en relation holistique des données recueillies lors du terrain de façon à améliorer la précision statistique des résultats par génération d'échantillons »

@dingdingpeng.the100.ci do you guys have merch? I need this on a t-shirt.

But somehow I missed that he was writing in his diary about his guilt over “flagrant distortions”. Wow

And best of all the dataset is completely open so you too can mine it for insights about open science and reproducibility in European science! And we hope you do! Get that here: doi.org/10.48804/FMK...

Excited to share our newest output from the osiris4r.eu project! Led by my wonderful PhD student at KU Leuven, Magdalena Kozula, we conducted what I believe to be the largest qualitative study of researcher's experiences with, and perceptions of, reproducibility.🧵 journals.plos.org/plosone/arti...

100% worth your 63 seconds if you have my reading speed. sinceyouarrived.world/taken

OpenAI's claim that this is a central conjecture in discrete geometry is not an exaggeration. This will I think be looked back on as the first time that AI solved a major mathematics problem (defined as a problem that all experts in some subfield had thought about). openai.com/index/model-...

Why is this guy not a complete pariah?

Hot take: AI should be involved in reviewing We've read this many times Manuscripts under review should not be shared with artificial intelligence, as these tools may not be able to assure confidentiality. AI-generated reviews can contain inaccurate information and biased language I'll push back

Many journals already require authors to upload data and code. A near-future where an LLM systematically checks the analysis and the code of every published study is realistic. Believe it or not, I think this is the future of scientific publishing.

Reviewing a paper, and it is so nice that the Metacheck R package we made allows me to download all OSF Files by just typing: metacheck::osf_file_download("2w123") (not the actual repo). All files are downloaded, maintaining the folder structure, without any hassle.

New interactive simulation in my online textbook illustrating publication bias in meta-analyses, and some techniques to model bias-adjusted effect sizes. None of these work perfectly. We need an unbiased literature to draw valid inferences. lakens.github.io/statistical_...

Research by @lfitouchi.bsky.social & @manvir.bsky.social suggests the tension between (evolutionary) deterrence, and (intuitive) retribution punishment motives is resolved by a model reflecting restoration of mutual benefit between cooperative partners→compensation+deterrence buff.ly/3T6fdvd

Still a few days to apply!

Two new facts stand out: 1. 85% of hallucinated citations in preprints are also in the subsequent journal version (thanks, peer review!) 2. Fake cites more likely to use the names of (male) scholars who are already highly cited, creating a fake-citation Matthew effect. arxiv.org/abs/2605.07723

From now on, I'm going to write papers where I take an arbitrary performance metric, restrict the sample to top performers on that metric, and then hunt for negative correlations with everything else in the truncated sample. I'm at a business school, so this should be extremely lucrative.

Classic academic's wikipedia page arc

@taymalsalti.bsky.social has done trojan work trying to reproduce the results of three meta-analyses, finding the usual host of issues of effect sizes that are apparently incorrect or non-reproducible Our article is out now: doi.org/10.1111/ejn....

Speaking of crazy d values, this reminds me of this feature I was thinking about adding to the forest() function in metafor. #Rstats #MetaAnalysis

[2016] A French literature Prof was celebrated by the French national assembly, Nobel prize winners, former government ministers, MPs, decorated scientists and academics for winning award ... [2026] ... that he now appears to have made up. www.theguardian.com/world/2026/m...

This thread is the most uplifting thing I’ve seen today. My takeaway is that hard work is never wasted.

oh well oh well...

My final piece in a 4-part series on vitamin D supplementation. I think that the evidence as a whole shows quite strongly that vitamin D pills are probably a waste of time for virtually everyone. Science! gidmk.substack.com/p/theres-no-...

Writing this out in words for emphasis: five thousand four hundred and seventy eight ethnographic paragraphs coded eleven thousand eight hundred and ninety three times by eight independent coders.

wait what

Reminder: when authors observe Cohen's d = 4, no they didn't. Article now retracted: www.nature.com/articles/s41... Critique: osf.io/preprints/ps... Blog on plausibility of effect sizes: trustworthy.scientific.claims/posts/if-res...

This is an actual line that was added to the official system prompt for Codex for GPT-5.5 by OpenAI. Usually the system prompt is as minimal as possible, so I assume it would otherwise mention goblins a lot. AIs are weird.

Reminder: If researchers find Cohen's d = 6, no they didn't. trustworthy.scientific.claims/posts/if-res...

Eye opening and informative on the costs of the French system of agrégation

There are several problems, the first one being the methodology of the French dissertation, which is not codified anywhere and strongly favors literary style over argumentative style. Students spend years trying to understand this Kafkaesque methodology instead of developing argumentative skills.

People writing pre-registrations: make them shorter. Like 92% shorter. What's the data, what's the model, what are the decision rules.

People claim that psychedelics have been used by Indigenous groups around the world for millennia. But clear evidence of serotonergic psychedelic use has long been confined to the Americas. Until now. Here, we document clear evidence of Indigenous psilocybin use in Lesotho. 🍄 osf.io/preprints/so...