Rob Davies

proud to report that the years have been unable to quench the little voice inside me that says “going on a three-day trip where you’ll be working the whole time? you’d better bring a print copy of the Personal Memoirs of Ulysses S. Grant in case you have time for a little light reading!!”

Some notable examples of long-standing errors in science from Douglas Allchin’s new article: “When is science ‘self-correcting’?”: doi.org/10.1016/j.sh... Few quotes follow…

Glad to see citations given to all this great work! McDermott (2022): doi.org/10.1017/pls.... Pham and Oh (2021): doi.org/10.1002/jcpy... Szollosi and colleagues (2020): doi.org/10.1016/j.ti... Devezer and colleagues (2021): doi.org/10.1098/rsos...

Some people were making me mad on the internet so I wrote a blog post about it. On the different reasons for people have for being left wing and how people manage the psychological tensions of the fact that they come apart. sootyempiric.blogspot.com/2026/10/the-...

Seven types of Ambiguity? Yn thys economye?

Excited about the CfP about effect heterogeneity and @markrubin.bsky.social @olivia.science @fabianhutmacher.bsky.social's project project on psychological theories, both of which are closely related to issues I work on! #metascience #psychology #philsci

Yes! Power should be a property of a single hypothesis test, even if that consists of multiple analytic tests. 1 hypothesis = 1 test is just the simplest case, but you can calculate power by simulation for arbitrarily complex analyses and decision rules.

Like if you’ve ever said something like “Claude thinks…” you better be fucking spectacularly nice to agents. You’re wrong. It didn’t think anything. But if you’re acting like it did and you’re being a dickhead then what the fuck are you practicing for? The cunt Olympics?

Thinking more seriously about the design of exploratory studies: A manifesto statmodeling.stat.columbia.edu/2026/10/09/t...

Randomly found an old poem I wrote about Microsoft

We found that the most gendered experience wasn’t work-life balance, the focus of most retention efforts. It was feeling devalued, in formal evaluations, but particularly in everyday interactions. Unlike every other topic, the gender gap in devaluation was larger among more senior faculty.

Honoured to be invited and excited to be presenting the work below and other projects I've been working on here later today on models in psychological and cognitive sciences — really looking forward to hearing from the other speakers and panel members too. ☺️

Hi all! I'm looking for people to try out my new collaborative writing platform. Think: Overleaf or GDocs for Mkdown LaTeX HTML Typst Quarto Free, open source, and full-featured. Many editors, chat, comments, track changes, history, AI, mobile, templates, etc Would ❤️ any feedback! librepaper.org

We got a lot of pushback from metascience/science reform community when we published this five years ago. It's kind of bittersweet now to observe that some of our critics have come around to invite higher methodological standards in forensic metascience. See, we could have been friends all along!

> The earliest account of [the hazards of] asbestos was provided by Lucy Deane [in 1898], one of the first Women Inspectors of Factories in the UK. [T]hese laywomen might not have been categorised as [experts] but they were competent observers [& t]heir reports were not refuted but simply ignored.

ARISE (1990) Mixed media - 20” x 20” I took the opportunity with this commission—an album cover for Sepultura—to experiment and have fun. 

#art

It’s still possible to differentiate LLM generated text from human text, and there are interesting effects if you try to make it less detectable (such as, paradoxically, the text becoming less convincing?). Only skimmed this for now, but seems interesting. www.pnas.org/doi/abs/10.1...

Are you good with numbers? We’re releasing an open dataset of numeracy scores from 75,688 people across 14 countries, gathered in 59 online surveys between 2020 and 2024 by the Winton Centre for Risk and Evidence Communication. Details and data: www.octopus.ac/publications...

...walking through the stages of science to show how what we think about people (gender, race, parenting, neurodiversity, sexual orientation, fatness) shapes how we make knowledge. More to come!

If you are interested in people’s numeracy then here’s our entire global dataset from the Winton Centre in Cambridge that you can use for analyses, along with some we’ve already run… #openscience #opendata #FAIR @octopus-ac.bsky.social

happy sickos day to all who celebrate

In 2002, Wadi Sūra II was discovered in the Western Desert of Egypt. A small rock overhang was found with painted animals of many types dating to about 7000 years ago. But it was a set of baby-size handprint stencils that presented a puzzle: because they weren't HUMAN hands.

NEW: Prime Minister Andy Burnham finally breaks a decade of UK govt silence…& reveals there was ‘clear’ Russian interference in the Brexit vote. 1/

Yesterday I ran a little drop in workshop on Governors Island where we made bird data portraits with cyanotype. Bird data can help us understand a place— and making art with it can be really fun!

*this* is why you don’t put off cleaning your coffee grinder until it’s too late. Broken ring burr holder = no coffee

Next Saturday, October 17th, for International Repair Day, hundreds of events will bring people together to celebrate the power and promise of repair. Repair thinking has long animated this journal. Read on for some of our essential reads on the repair of our built and natural environments.

There are multiple problems that we don't want to lump together completely: unbiased noise, selection bias, confounding bias, collider bias, differential drop-out, slow convergence of CLT, and more. But Meng's paradox is a basic lesson.

We really need to make Meng's Big Data Paradox central to teaching, don't we? Researchers don't know this stuff bc they have never been exposed to it. But it is fundamental. (Link to paper PDF statistics.fas.harvard.edu/sites/g/file...)

Modrian, Trees, 1908

Time for 2026 updates to my annual “opinions about solar” thread, now with storage. If you like these, you might like edition 2 of my book, Solar Power Finance Without The Jargon. A 30% discount code WSQ0437 is valid on publisher website til end of Nov 2026. www.worldscientific.com/worldscibook...

Highlight Text colours text and backgrounds in every Quarto format. Version 2.4.0 adds named styles, so '[Overdue]{.alert}' carries colours you define once. Change the red in one place and every element follows. m.canouil.dev/quarto-highlight-text #Quarto #OpenSource #QuartoPub

sheets.works/data-viz/hol...

we have a policy for the use of genAI at the @aial.ie. we reflect on values key to scientific integrity, such as replicability, and err on the side of not using it blogpost: aial.ie/blog/gen-ai-... pdf: zenodo.org/records/2322...

View from the kitchen window of a Maryhill tenement, Glasgow, 1970, photo by Nick Hedges.

Like many scientists, I’ve been mulling over the impact of LLMs on our fields. I’ve not yet seen it articulated, but I think there is a largely unspoken “invisible labor underclass” framework at play.

Fun blog, and it made me long for more connections between philosophy and applied statistics

Still upset about that highly cited paper in Psych Science you consider deeply flawed? Now is your time to shine ✨

THE TAPESTRY IS HERE #CPC26

You guys want to hear a short but adorable story that happened right in front of me at the baggage claim in NYC yesterday? I wasn't part of it and there are bits I didn't catch but holy crap, I can't stop thinking about it.

Happy Buster Keaton's birthday!

webrarian is a new R package. It turns a folder of R scripts and data into a static website where R runs in the visitor's browser, with no compute server and nothing to install. blog.thecoatlessprofessor.com/posts/introd... #rstats #webr #webassembly

Psychology people: If you haven't seen them yet, check out these upcoming special issues at Psychological Science. Submissions invited.