#CancerResearchUK will stop paying #APCs for funded researchers. Good. Unfortunately its announcement is marred by false assertions and assumptions about #OpenAccess.
news.cancerresearchuk.org/2026/04/01/w...
I say more on Mastodon.
fediscience.org/@petersuber/...
I rarely get this angry, but this pisses me off.
Just because a bunch of fools started calling what they do "cognitive science" when it isn't science, doesn't mean that what we do as cognitive scientists isn't.
Go read Von Eckhardt's "What is cognitive science?" if you doubt me.
#statstab #504 Likelihood Ratio Test for Publication Bias
Thoughts: Interesting idea for a problem many of us think is present, but is hard to measure.
#likelihood #publicationbias #QRPs #metascience #metapsychology #metaresearch
freestylerscientist.pl/projects/lik...
The mathematical techniques are not rigorous btw
My priors are: the bigger the claim in the title, the weaker the arguments.
BTW. I'm quite skeptical toward Bayesian Estimation as most common replacement for frequentist statistics, but come on!
I want to thank Richard Dawkins for making my arguments about popular science writing as a vector for ideological scientism incredibly easy to develop
The “replication crisis” is partly a artefact because people started trying to do it, and turned out they’d not been clear enough about what they were doing all along. It’s a publishing problem not a science problem.
And is the same in all disciplines, I promise. It’s been studied.
Probably the only reform science needs is to dethrone its gods and heroes, and abolish and purge all elite (and otherwise) boys clubs. I strongly suspect everything else would start falling into place after that.
"when Trump announced the idea of tariffs to address the trade deficit... every economist around the country said: That's a very bad idea, that won't work....
If you look at 2025 as a whole, we have basically the same trade deficit...
So I'm going to score this one for the economist, mate."
In a sense, we have completely replaced a lot of scientific processes with the power of statistical methodology and essentially outsourced part of our scientific thinking and practice to statistical inference. And now we're confronted with the risk of the same thing happening with AI.
