We're so excited to announce that our first official CONNECT preprint is now available: "A network of networks: The next frontier for Big Team Science" 🔗 doi.org/10.31234/osf... Please read, share, and give us feedback! ✉️ connect-contact[at]manybabies[dot]org

Richard Lynn was criticised for using a sample of children with intellectual impairment to calculate the "national IQ" of Equatorial Guinea in 2002. In 2019, he calculated the IQs of Native Americans from children referred for psychoeducational evaluations, including those with learning difficulties

arXiv will no longer accept review articles and position papers unless they have been accepted at a journal or a conference and complete successful peer review. This is due to being overwhelmed by a hundreds of AI generated papers a month. Yet another open submission process killed by LLMs.

A French cyclist survived for three days after a horrendous 130-foot fall into a ravine, kept alive by the bottles of red wine he had in his shopping bag, police said.

New paper finds that selective reporting remains the most replicable finding in science: journals.sagepub.com/doi/full/10.... I especially like their new exploratory metric 'p-values per participant'. Some papers had 11 p-values per participant! 🤯

"Introducing Doublespeed, an [Andreessen Horowitz backed] startup operating a phone farm to flood social media with AI-generated slop on behalf of its clients. Doublespeed clients can expect to pay anywhere between $1,500 and $7,500 a month for access to its phone farm."

We built the openESM database: ▶️60 openly available experience sampling datasets (16K+ participants, 740K+ obs.) in one place ▶️Harmonized (meta-)data, fully open-source software ▶️Filter & search all data, simply download via R/Python Find out more: 🌐 openesmdata.org 📝 doi.org/10.31234/osf...

We don't know if these two DAGs are up to divine mediation problems like identifying the Father's Supernatural Direct Effect or Nicene Indirect Effect, but the Church is in the business of cross-world assumptions, so we're sure it will work out. 10/10 for whichever one is theologically correct 🤞

This new reporting from @jeffhorwitz.bsky.social is jaw-dropping and I think warrants a response from @cos.io given their ongoing collaboration with Meta. Jeff writes about how Meta identified content-specific harms as a key problem for teens on Instagram. 🧵 www.reuters.com/business/ins...

Hello Bluesky! We rate DAGs. Some are great. Some are... not so great. But we rate them all. Let's start with a famous powerpoint hairball a.k.a. "the Afghanisdag", presented to Gen. Stanley A. McChrystal around 2010. His own rating? 1/10 "When we understand that slide, we'll have won the war"

This is a backdoor path from the Son to the Holy Spirit that should be closed by conditioning or stratifying on the Father.

“Well-meaning” people are using machine translation to write Wikipedia articles in languages that they don’t speak themselves, accelerating the degeneration in quality of the web corpus for several languages with relatively few native speakers. www.technologyreview.com/2025/09/25/1...

New research finds that Twitter’s efforts to remove COVID-19 vaccine misinformation were largely ineffective and sometimes backfired. Skeptical communities grew more active and viral, and content quality declined over time. link.springer.com/article/10.1...

Excellent 🧵 about LLM synthetic data (silicon samples etc) and why they don't solve any particular problem in human research. FWIW, in addition to results and considerations like these, I've argued elsewhere that the entire question is ill-formed: quantuxblog.com/synthetic-su...

Psychologists running empirical studies to rediscover engineering design choices is such a strange genre of papers. By all means, run studies on LLM judgments -- but what else than lexical co-occurence and statistical priors would they be based on??

ReplicationResearch.org is now open for submissions! Submit replications and reproductions from many different fields, as well as conceptual contributions. With diamond OA, open and citable peer review reports, and reproducibility checks, we push the boundaries of open and fair publishing.

Well, good for DC search.app/PK45E

I really strongly feel that some fields of research would profit if researchers stopped collecting online data for some time and instead maybe just read a bit outside of their field.

{tinytable} 0.14.0 for #RStats makes it super easy to draw tables in html, tex, docx, typ, md & png. There are only a few functions to learn, but don't be fooled! Small 📦s can still be powerful. Check out the new gallery page for fun case studies. vincentarelbundock.github.io/tinytable/vi...

Intervening on a central node in a network likely does little given that its connected neighbors will "flip it back" immediately. Happy to see this position supported now. "Change is most likely [..] if it spreads first among relatively poorly connected nodes." www.nature.com/articles/s41...

The deadline to provide inputs into this piece of work is swiftly approaching -- 8 October. Please consider filling out the survey we're using to structure people's input! www.who.int/news-room/ar...

This study of intelligence in the UK Biobank is typical of a lot of current social science genomics. Impressive technically, and not over-interpreted. But still, a main result gets lost in the sauce. Within-families, the direct-effect polygenic score explains no more that 1-3% of the variance. /1

💖This paper has been ~11 years in the making - and probably my favorite project of all time. Thrilled to see it in @pnas.org! I'm so lucky that Zach decided to do a second PhD and join my lab @psychillinois.bsky.social back in 2014 - a fabulous scientist & human being! www.pnas.org/doi/10.1073/...

this is a very sharp piece on why it makes no sense to run universities as if they are businesses. They're not businesses. www.afr.com/work-and-car...

Is analytical flexibility really the biggest problem while you’re confusing the ephemeral statistical effects of psychological processes with the ephemeral statistical effects of language prediction trained on massive data sets? Hah.

Nothing. I use it for nothing at all because AI is good at zero of the tasks I do regularly Honestly I don't even know what its web address is, is it like a 2000s style ChatGPT.com or something funkier like chat.g.pt

Waiting for my preprint to be accepted, so in the meantime a teaser: here's what happens when you try to estimate a between-scale correlation based on LLM-generated datasets of participants, while varying 4 different analytic decisions (blue is the true correlation from human data):

McSweeney (2002) is required reading for anyone who studies or references Hofstede's individualism-collectivism--or Markus & Kitayama's independence-interdependence, for that matter. Just shoddy work up and down. doi.org/10.1177/0018...

Finally! 🤩 Our position piece: Against the Uncritical Adoption of 'AI' Technologies in Academia: doi.org/10.5281/zeno... We unpick the tech industry’s marketing, hype, & harm; and we argue for safeguarding higher education, critical thinking, expertise, academic freedom, & scientific integrity. 1/n

In a new paper, my colleagues and I set out to demonstrate how method biases can create spurious findings in relationship science, by using a seemingly meaningless scale (e.g., "My relationship has very good Saturn") to predict relationship outcomes. journals.sagepub.com/doi/10.1177/...

ICYMI: Google removes net zero goal from website.

Now on CRAN, ggdiagram is a #ggplot2 extension that draws diagrams programmatically in #Rstats. Allows for precise control in how objects, labels, and equations are placed in relation to each other. wjschne.github.io/ggdiagram/ar...

How should the behavioral sciences be mainstreamed into public health? How would we know if this goal is achieved? With the WHO Behavioural Insights Unit, my team has been working on these questions. Curious what we came up with? Check out the public consultation below www.who.int/news-room/ar...

How should the behavioral sciences be mainstreamed into public health? How would we know if this goal is achieved? With the WHO Behavioural Insights Unit, my team has been working on these questions. Curious what we came up with? Check out the public consultation below www.who.int/news-room/ar...

Not for nothing, whenever someone conflates search algorithms, LLMs, and whatever the fuck AI is, I send them this article.

This is what I've been saying since 2023 (image below) "prediction: use of "AI" [...] will come to be broadly associated with cheating, deception, lack of respect for other people, and low quality work that cannot be trusted in important settings"

AI-adjacent people working in cognitive science should pay attention to this horrifying lawsuit about ChatGPT helping (encouraging!) a child to commit suicide

realistic Star Trek

“The study authors asked GPT 4o-mini to evaluate the quality of 217 papers. The tool didn’t mention in any of the reports that the papers being analyzed had been retracted or had validity issues. In 190 cases, GPT described the papers as world leading, internationally excellent, or close to that”

Just sharing a public thank you to the SIPs + PsyArXiv volunteers for maintaining such a successful resource I understand the frustration of those with preprints that were temporarily removed but I think this is a good chance to reflect on the volunteer work of others that we often take for granted

AI stocks drop sharply after an MIT report says 95% of AI investments generate zero (0) return. Even AI prophet Sam Altman acknowledges there's a bubble:

Large Language Models Do Not Simulate Human Psychology arxiv.org/pdf/2508.06950

Racial diversity in healthcare providers reduces disparities in healthcare outcomes.

Check out our new pre-print in which we argue something that might be obvious but doesn't seem to be obvious to everyone: Large Language Models (such as ChatGPT) do not simulate human psychology.

Systematic reviewers have a 10.7% error rate when screening papers for potential inclusion. In other words, about 1 in 9 abstracts are categorised incorrectly (i.e., false inclusion or false exclusion). This really drives home the benefit of having at least two screeners

using AI to review a paper is one thing (and there's a lot to be said abt that), but submitting a paragraph-long AI summary as a review is a whole 'nother level of laziness. editors, flag that shit & never pass it along to the authors. also don't accept any papers or reviews from that reviewer.

Post-tenure scholarly productivity falls way off for most scholars if you're not in a lab field www.pnas.org/doi/10.1073/...

New working paper led by Jaspreet Singh: osf.io/preprints/ps... Some of you may know about repliCATS, in which (mostly Global North) experts used structured peer review to judge judge the credibility of behavioral research. We used the same method with behavioral scientists from around Nairobi 1/