Adrian Barnett

A good argument for preprinting everything here: www.nature.com/articles/d41...

How to run a lottery Abstract. Adrian Barnett worked in a bingo hall as a boy, and now runs a lottery for research funding. Here he explains how it works and how it might be be

I once worked in a bingo hall and saw the joy of people winning money at random. Now I randomise academics in the serious business of who wins research funding. With funding lotteries growing in popularity, I reflect on how to run them and the lessons from bingo. academic.oup.com/jrssig/artic...

Access via your organisation

Genuinely once read "SSPS". Another reason to switch to R and you can't misspell that.

Screen shot of a statistical method's section from a scientific paper with Tukey misspelt as Turkey.

I think a lot about the general lack of statistical competence. Misspelling the great John Tukey's name as Turkey may seem harmless, but I can't think of how anyone whose is serious about statistics would do this.

Low quality papers are flooding the cancer literature — can this AI tool help to catch them? A large language model scans abstracts and titles for signs that an article was produced by a 'paper-mill' company.

2025. (not just cancer lit....) Low quality papers are flooding the cancer literature — can this AI tool help to catch them? A large language model scans abstracts and titles for signs that an article was produced by a 'paper-mill' company. www.nature.com/articles/d41...

Should we decide by lottery who gets a medical treatment first? Lotteries may be fair in theory, but new research shows people prefer expert committees to decide who gets scarce medical treatment.

Lotteries may be fair in theory, but new research shows people prefer expert committees to decide who gets scarce medical treatment.

How To Stop The Next 10,000 Bullshit Papers Some remarkably un-radical proposals

Free data resources in science are having the shit mined out of them to produce X,000's of bullshit papers.

What should we do?

(Warning: I sat on this draft for too long, and publishers are already doing some of it. That's what I get for engaging with the news.)

open.substack.com/pub/jamescla...

It's a needlepoint.

I think the authors would have been better off taking a walk than creating this diagram.

From the rstats community on Reddit: My mother liked my laptop stickers and did this "tidy" phone/hdd patchwork bag 😍 Explore this post and more from the rstats community

R beats Python because R has dedicated quilts, for example: www.reddit.com/r/rstats/com.... Once you've lost the quilting war, you've lost everything.

Stick figure surrounded by tree-like structures with speech bubbles about sepsis.

At last, a simple diagram that explains random forests. Go forth and make random diagrams, I mean random forests. From www.thieme-connect.com/products/ejo...

It's 2025.723 and scientists are still putting too many decimal places in their results.

Map of Queensland showing clear disease outbreak in the north

When the zombie apocalypse starts in Weipa

When the zombie apocalypse starts in Brisbane

Peer reviewers like papers that cite them New report finds that referees are more likely to recommend studies that refer to their own work

Peer reviewers are more likely to recommend accepting papers that cite their work — my latest for @cenmag.bsky.social:

cen.acs.org/policy/publi...

@aidybarnett.bsky.social

“I still yearn for the good old days…”

From the latest Private Eye, out now.

Google Scholar Is Doomed Academia built entire careers on a free Google service with zero guarantees. What could go wrong?

Since search is dead, how soon do you think Google Scholar is headed for the Google Graveyard? I'm betting it's soon, and academia is NOT prepared

A lot of government contracts in Australia have suppression clauses as a standard - that should definitely end. Media attention like this helps. And then perhaps some better training of bureaucrats to show that the truth is always the best long term solution, that's probably naive of me.

Peer reviewers more likely to approve articles that cite their own work Preprint examines how citations can influence the review process.

Reviewers are more likely to approve manuscripts if authors agree to cite their work than those who don't get cited.
Thanks to @aidybarnett.bsky.social, @balazsaczel.bsky.social and @econfeld.bsky.social for chatting with me for this story!

www.nature.com/articles/d41...

Don't miss out!

Emails raise new questions about state government's role in scientific report In May, the WA government released a report from an ongoing study into the impacts of industrial air emissions on ancient rock carvings, but new emails obtained by FOI have raised questions about the ...

Sadly familiar story on the ABC over government interference in scientific results www.abc.net.au/news/2025-08.... I fully support the statistician (Adrian Baddeley) for quitting.

@tomstafford.mastodon.online.ap.brid.gy has it. It's growing all the time.

Interesting stuff, but I don't see anything on peer review in that PDF.

Fair points. But for the uncited reviewer to be right and withhold their approval it must have been for a serious issue. Wouldn't we hope that the other reviewer (on average) would also have spotted the serious issue and withhold their approval too?

But by matching on paper the quality issues should become sidelined. We have two or more reviewers looking at identical papers, some good, some bad. If there are consistent patterns in behaviour depending on whether a reviewer is cited, then that's likely independent of the article quality.

Figure 6 from the paper, showing a model airplane that appears to have been photoshopped into a wind tunnel. The photo of the plane is taken from above, but the wind tunnel is photographed from below.

Retraction by ‪@mdpiopenaccess.bsky.social‬ of a paper in which a University of West Florida professor appears to have photoshopped his model airplane into a photo of a NASA wind tunnel
mdpi.com/2226-4310/7/...
pubpeer.com/publications...

If there's anyone listening from the Global Burden of Disease, your data is being used for trash studies and your reputation is heading down the toilet.

Care about #reproducibility in research & publishing? Join this webinar on Aug 27th, jointly organised by @aimosinc.bsky.social & @ausrepro.bsky.social on statistical reproducibility with @michelenuijten.bsky.social & @tomhardwicke.bsky.social Free - register here qut.zoom.us/meeting/regi...

Thanks for sharing. Your proposed solutions made me think of another compromise: Reviewers self-citations are in the paper for people to read, but don't get used in the indexing data. So the reviewer can contribute to scholarly debate, but can't boost their h-indices.