The data is in: the NIH goalposts have shifted. What were once almost certain fundable scores have become coin flips and what used to be likely grants have become aspirational, leading to fewer awards. Another manifestation of how HHS policies have led to fewer awards and less science.

I tried to tell y'all.

This recent RCT of an "AI stethoscope" claims the technology "shows promise" for diagnosing cardiovascular conditions. It does not. It is a textbook example of the risks of conducting unprincipled 'per protocol analyses'. Once again, peer review at a major medical journal has failed. 🧵 1/

OpenSAFELY is open from today! Huge thanks to all who supported this vast collaboration: whole population GP data; in a productive platform; innovative privacy protections; unprecedented support from professions, privacy campaigners; &c Now it's over to users! www.bennett.ox.ac.uk/blog/2026/02...

Happy 🎂 Richard Price b #OTD 1723 (d 19 Apr 1791) Best known for popularizing the essays of Thomas Bayes, the basis of what is now called Bayes's Theorem. Price’s contributions were more substantial than he has been credited with. 1/4🧵

Can anyone tell me which crypto coin I can cash all this ‘Enhanced Reputation’ in at? Because I’m still carrying £15k worth of debt from altering my UK terrace so I can live downstairs when my MS gets to that stage… and I don’t have a job.

New submission format at SBE: “Replications as Registered Reports” link.springer.com/journal/1118... You can get "in-principle acceptance" before data collection even begins; final paper gets published regardless the results, if the study is conducted rigorously. #EconSky

On the perils of open data. Archive link: archive.ph/ZavpU

Emory libraries cancel Elsevier journals due to rising journal costs 📚 www.emorywheel.com/article/2026...

Machine learning research is not serious research and therefore hallucinated references are not necessarily a big deal, agrees a prestigious group of machine learning researchers statmodeling.stat.columbia.edu/2026/01/26/m...

We are always excited to see the QUOD Biobank grow, with samples from over 9,000 donors. We are grateful, as always, for the generosity of our donor families and to the teams across the UK who make collection possible, helping us to support research into improving transplantation outcomes.

We are so proud to present 9 posters at the NHSBT Clinical and R&D Conference this year! @liseestcourt.bsky.social @transfusionlib.bsky.social @stemcellevidence.bsky.social @btjo.bsky.social @thrinaline.bsky.social

#SpecialIssues have fueled the growth of some of the largest #OpenAccess publishers. Does a journal that allows a #GuestEditor to both plan a special issue and write many articles in it have a conflict of interest? #scicomm #peerreview @science.org www.science.org/content/arti...

Here's a suggestion for a New Year's resolution: If you see influential bad research, say something. One part of the whole replication crisis story is that a lot of psychological researchers privately knew that a lot of stuff was bad, but it wasn't discussed publicly.

Point of no return: academic #funding is at a tipping point. When the effort and money spent applying for grants equals or exceeds the funding awarded, the system stops serving science and starts draining it. Time to rethink how we fund #research. 🧪🌐🌍 www.nature.com/articles/d41...

Good thread. My concern is, CI is presented as a solution to bad stats when in reality it suffers from many of the same limitations. We can almost never be certain of most reqs in any case -- exogeneity; data integrity; missingness mechanisms; selection bias to name a few (1/2) #statsky #econsky

WTF is this? I did not write or submit this paper. Are journals now run entirely by AIs who hallucinate entire publications? Knowledge Of Pregnant Women About The Benefits Of Pregnancy Books In Ottawa | International Journal of Health Engineering and Technology share.google/ArcgxhUDfSY9...

These claims about what AI can do truly need citations. Checking statistics isn’t routine work, nor is whether citations to the literature are sensible. Whether statistics are appropriate and render reasonable inferences is inherently a question of human judgement.

Finally, if you can get through the paywall this piece from earlier in the week illuminates how the experience above is just a tiny part of an industrial-scale, global, and sector-wide catastrophe of AI and citations to nonexistent papers in academic publishing bsky.app/profile/mile...

Does e-cigarette use cause stroke? ~2 years ago, Gal Cohen & I alerted MDPI to fatal flaws in a published paper which claimed nicotine vaping causes stroke. This morning, the journal finally retracted the flawed article: doi.org/10.3390/neur... ... #tobaccocontrol #harmreduction #statistics

You remember that Nature Aging paper about how multilingualism protects against accelerated aging? Well…

I made an error that I found thanks to @lukaswallrich.bsky.social. The code and preprint have been corrected/updated. Effects on false positive rates are smaller and only for alternative and correlation method but I think the overall argument still works.

They're eliminating the Statistic Department as the Admins apparently can't do basic math or data scrubbing. "Other errors they found included numerous data errors, such as misplaced decimal points and incorrectly typed data." #stats

🎯 statsepi.substack.com/p/the-review...

A Threat to Evidence-Based Vaccine Policy and Public Health Security at the FDA: former FDA commissioners call out Dr. Prasad's departure from previous standards in @nejm.org www.nejm.org/doi/full/10....

Diederik Stapel was a massive watershed moment in psychology. However, he was -- and let's be slightly glib here -- some guy from The Netherlands who wrote social psychology papers. The full accounting of the Eysenck case is approx, at minimum, TWO STAPELS. retractionwatch.com/2025/12/03/n...

“A great deal of effort has been invested to evaluate if factors like questionable research practices (etc.) affect replicability and effect heterogeneity. We would like to see more of that effort going into critical examinations of the interac-tion between common measures and methods.”

[Pic] is such an important point. If you're never wrong, you can't learn anything. And if you're afraid to be wrong, you're easy prey for plausible-sounding misinformation. Education, no matter the subject, is about how to think. The facts/ideas/materials change, the skills required do not.

15-year-old bug 😐 in an industry that has the highest profit margins possible pretty much www.theguardian.com/science/2017...

The most important part of this article, to me, is the end. It challenges a false binary between “long COVID” and “everyone else.” Maybe it's a spectrum... 🧵 (1/5) www.bmj.com/content/390/...

Domain knowledge is always critical for the appropriate application of statistical methods. Here, it would have prevented this embarrassing episode occurring at all.

Not sure where this is from, but assuming it’s right, there are clear implications for countering the replication crisis and “false positive” psychology: Insufficient sample sizes are a primary culprit, and should be a primary target of reforms.

Unfortunately, they can fake data just as easily, and do much more damage to the literature with it. Need checking of grants, ethical approval, etc to verify that the research actually took place.

Publication bias is greater at the top five journals: "56% of statistically significant results [in leading econ journals] were selected to be statistically significant. Selection bias is greater at top5 journals, where 66% of significant results were selected to be significant."

We have progressed from data collection to data analysis.

"Two major publishers have begun to automatically reject the vast majority of papers based on public health data sets, following revelations that unscrupulous actors use these data sets to churn out nonsense scientific papers." www.science.org/content/arti...

A lot of psych is already conducted with online convenience samples & ppl are probably excited about silicon samples bc it would allow them to crank out more studies for even less 💸 How about we reconsider the idea that sciencey science involves collecting own data. www.science.org/content/arti...

People are running stats on LLM-generated participants and think they’re being social scientists when in fact they’re technically just playing a very strange video game. This is like saying you’re doing math research because you’re playing sudoku. www.science.org/content/arti...

I walk my dog by the river. Properly managed sewage systems are sexy. top-of-the-poops.org/waterway/uni...

This is a failure to explain herd immunity properly. Some parents hear only a demand that they vaccinate to protect others. But their (unvaccinated) kids are most likely to be harmed in an outbreak. It's their (unvaccinated) kids who pay the price of herd immunity breaking down. Spell it out.

A US senate subcommittee was told this unpublished vaccine study was 'the most important ever.' Henry Ford Health now confirms why it was never published: it failed scientific standards. Here's what's actually wrong with the study: jakescottmd.substack.com/p/why-the-senates-most-important-vaccine

Academic authors, here's a peek into the black box of journal publishing from an journal editor if you can bear it:

This can't be true because academic papers are all peer-reviewed by experienced experts who would be able to easily identify these shenanigans and stop them in their tracks.

🚨 Relaunch of CRAN Task View: Clinical Trial Design, Monitoring, and Analysis #rstats Complete rewrite by an int'l team of experts in #clinicaltrials Ya Wang, Thomas Jaki @thomasjaki.bsky.social, Laura Pascasio Harris, Orla Doyle, Elias Laurin Meyer, Wilmar Igl cran.r-project.org/view=Clinica...

Nice example of the ecological fallacy, with a conclusion so absurd it's hard to understand how it got written down. #stats

Recent paper on paper mill activity in the Taylor & Francis journal 'Bioengineered', which we have published in Taylor & Francis journal 'Bioengineered'. ~25% of the investigated 878 papers were 'problematic' Bioengineered has now halted submissions. www.tandfonline.com/doi/full/10....

A new expression of concern was published on this randomized clinical trial of cinnamon for diabetes. The study itself has been cited, according to Google, 1,700 times. I first complained about it in 2023, so let me tell you the story of one of the most problematic papers I've come across 1/n

Pretty striking for the BMJ to run this piece. “[Immunity debt’s] explanatory power has faded as the number of non-covid infections has kept rising each year.”

It really is quite stark, mapped out like this. (Assuming the underlying data is correct, I can't vouch for it but it smells right.)