@boryslaw.bsky.social
Intro to causal inference for psychologists: https://czasopisma.uwm.edu.pl/index.php/pp/article/view/9731/7171 A causal-theoretic definition of measurement invariance (see p. 14): + A new ordinal regression family https://osf.io/preprints/psyarxiv/e7a3x
The question which always bothers me, though, is whether the research is good enough to justify the time, effort, public funds, and the reaction of the mostly severely incompetent readers who will then try to do research based on the results.
A point of contrast.
Well, yes, people can use regression models to answer questions. The larger models let them do that at larger scales. It's still the people who ultimately provide the answers, though.
Wasn’t joking
"I got my PhD by writing prompts instead of doing research, I'm winning"
got some bad news, there still no jobs and now you also know nothing
David Perell's YouTube conversations with writers are one of the best things I have seen in a while. They are personal, warm, filled with surprises and insights, and let us peek into the science and art of one of the most important forms of thinking that is storytelling.
every quantitative measure is actually a stack of qualitative assumptions in a trenchcoat
A picture is worth more than a thousand words. However, sometimes a short sequence of words describing a beautiful scene can beat the picture.
"Their fundamental design flaws are completely hidden by their superficial design flaws" is something I think about way too often.
In psychology, that's because most researchers ask the wrong questions. If you are a teacher, ask your students to raise their hands, and you will see d=infinity. We should focus on these kinds of effects rather than pretend we can do science like physicists or chemists.
Also, you can and should be doing both things in science
I love love love doing those kinds of seminars. And I will seize every opportunity to do them in the future. I do the same thing with some paper I coauthored in the past, though.
My employer asks me to complete a survey on AI usage for which this is the first question (required):
We're thrilled to announce the Ctrl-Z Award, a US$2,500 prize for researchers “who discover substantial errors in their published work and take meaningful steps to correct the scientific record."
Covered by @nature.com today; read more here: centerforscientificintegrity.org/2026/03/10/a...
I'm not saying you can't possibly generate a worthwhile hypothesis from your data. I'm just saying that generating a hypothesis from the entirety of human knowledge that preceded your data is a much safer bet.
Curious about how psychological theories are built, tested, and refined in practice? This summer, the Theory Methods Society is launching the very first edition of the summer school Theory Building in Psychology at the University of Amsterdam (July 6–10, 2026).
theorymethodssociety.org
The Norwegian Consumer Council with an amazing video on enshittification and how to resist it.
www.youtube.com/watch?v=T4Up...
One funny effect of pretending that methodological issues can just be ignored because of “future studies” is that it probably prevents those future studies from happening. Like why bother actually addressing hard issues when you can just get away with hand waving?
bro just one more future study bro, bro I swear just one more future study and it'll fix the inference bro
I have got the data
and I figured out exactly how to do my cluster analysis
all I need is a relevant question that my cluster analysis is going to answer
There are no top journals; there are only silly, outdated, and grotesquely inadequate methods of disseminating alleged knowledge that we call journals.
Leibniz, looking at the universe: "Why is there something instead of nothing?"
Me, looking at my Outlook calendar: same
Every time I am forced to use Matlab I can only conclude it is a language designed by and for sociopaths
Shout out to this @olivia.science blog post that I sent out on a neuro listserv back in grad school, that a prof there is apparently still mad about
neuroplausible.com/matlab
“No front-page stories in NYT, WSJ, WaPo. No mention on NBC and CBS Sunday Shows. No stand alone segments on Evening News shows. A death count on par with the OKC bombing is relegated to the back page.” www.columnblog.com/p/corporate-...
ReasBook is a Lean 4 project for formalizing mathematics from textbooks and research papers. github.com/optpku/ReasB... #LeanProver #ITP #Math
I joined the boycott. Consider joining yourself. All AI is dangerous but ChatGPT's developers are truly devoid of conscience.
There's a discussion on my TL about whether you'd expect studies from the 90s to replicate in the 20s given that the situational context has changed, and it feels like everyone involved is *so* close to getting it but the hegemony of experimental methods in US social psych is proving a mental block.
The greatest risk of AI in Higher Education isn’t cheating — it’s the erosion of learning itself
If we eliminate the friction that builds expertise, what’s left? Struggle isn’t inefficiency. Ignore this, and we’ll produce credentials while quietly dismantling HE
theconversation.com/the-greatest...
