Show participants the decisive winner-loser moments of an NBA game, then immediately ask (economic) zero-sum beliefs.

Interesting paper, the experimental evidence could be oversold though. In S3 and S4, the effects are small, about .18 to .20 points on a 7 point scale, d = .15 to .16 (p-values .007 & .018 - one-tailed). Also likely, conceptual overlap between IV and DV & demand effects:

Every PI

The alt text for this image is available upon reasonable request.

Majestic, well done @rickcarlsson.bsky.social

But also the smaller the n the larger effects you can find

3 new studies on using psychedelics as treatments for mental disorders just came out; 2 in JAMA Psychiatry, 1 in Lancet Psychiatry. I reviewed one of these submissions.

Here are the core results. 🧵

In addition, "Two serious adverse reactions were reported after psilocybin, 25 mg, including 1 case of hallucinogen persisting perception disorder".

The fee to publish an open access paper at Trends in Cognitive Science is now over $7,000. Seriously, @elsevierconnect.bsky.social ??? For a 4,000 word piece? Talk about a broken system.

Mein Handy ist mir auch ins Müsli gefallen.

Ny analyse: Top 1 pct. har mere formue end 82 pct. af den danske befolkning En ny opgørelse viser, at de rigeste cirka 48.000 mennesker i Danmark har lige så meget formue som 3,9 millioner danskere tilsammen.

Denmark is often portrayed as a fairly equal society, but when looking at the distribution of wealth a different picture emerges. The richest 1% of the population owns more wealth than the bottom 82% combined. Wealth tax is now being debated. #dkmedier #dkpol #dkvalg

(oxfam.dk/publikatione...)

Coming up with a post-hoc theory section when your results are weird:

Finally! 😄

Wow the effect sizes are wild, a d of 21.5 on stress!! I'll get some Omega 3 now.

Screenshot of the "Does that use a lot of energy?" online app

Hannah Ritchie has built a fun little tool where you can compare energy usage of various products and activities.

This is super helpful imho, because it's so hard to develop intuitions even just about the scales involved here.

hannahritchie.substack.com/p/does-that-...

doi.org/10.1093/aje/...

This is purely correlational. Authors themselves state "the cross-sectional nature of our study design precludes any causal interpretations. For example, non-normative action intentions may lead people to expressing heightened threat from repression”. Much more caution is needed, also in the title.

🧵on my new paper "Synthetic personas distort the structure of human belief systems" w Roberto Cerina I'm v excited about...

🚨 Do synthetic samples look like human samples?

We compare 28 LLMs to the 2024 General Social Survey (GSS) to find out + develop host of diagnostics...

This recent RCT of an "AI stethoscope" claims the technology "shows promise" for diagnosing cardiovascular conditions.

It does not.

It is a textbook example of the risks of conducting unprincipled 'per protocol analyses'. Once again, peer review at a major medical journal has failed.

🧵 1/

198 effect sizes in ego depletion resesrch showed an effect size of d=0.62. Preregistered large replications (including some by original authors) yielded an effect size of 0. No one has been able to offer any other explanation for this huge research waste than massive p-hacking.

This too…

Sorry, I don't understand how that addresses the issue of power and the robustness of each interaction.

Cool idea, cool studies! I'm a little concerned though by the robustness of the central interactions. S1: p = .019, S2: p = .031, in S3 p = .036. Huge sample sizes are need for robust attenuated interactions doi.org/10.1177/2515...

Just an aside: 41 of the 100 meta analyses published in Psych Bull between 2023 and 2025 do not include *any* quality screening of identified studies. To the degree that those literatures include bad studies, they are being laundered together with the good ones.

Just had a similar experience at Scientific Reports. We will submit to Meta Psychology, great outlet! open.lnu.se/index.php/me...

Positive and negative self-esteem are two (or one?) such offenders that I cover in my blog post.

www.the100.ci/2023/06/13/d...

Scientific Reports is the new Frontiers, from my experience. Low quality control, high publication fees, they have a deadline for reviewers of 10 days.

Many people, unfortunately, also just don't know how bad it is (one of my colleagues had no clue; now she has and she is angry).

This is a chance for behavioral science. Moving away from weak online studies, weak experimental manipulations, and moving to studying real behavior in real life. Remember the 60s? Social psychology was real. Asch, Milgram, et al.

Researchers recently asked Americans what income level other people need in order to live a good life.

A whopping 86% of Americans reported income levels well below what others say they themselves need.

Fascinating!

Fig. 3. Scatter plots of the synthetic and empirical estimates, validation study (Stage 2). Showing N = 30,135 item-pair correlations, N = 257 scale reliabilities, and N = 1,568 scale-pair correlations for (top) the pretrained SBERT model and (bottom) the fine-tuned SurveyBot3000 model. SBERT = all-mpnet-base-v2 model.

Fig. 4. Prediction error of the synthetic estimates, validation study (Stage 2). Our prediction model allowed the error term to vary freely according to the predictor, the synthetic estimate. The thin-plate splines show that some synthetic estimates were predictably more accurate.

Fig. 5. Accuracy by domain. Accuracy differed across domains. SurveyBot3000 accuracy (colored) was always higher than SBERT accuracy (gray). Results were largely consistent whether accuracy of items was tested (left, circle) within domains or (right, cross) across domains.

Fig. 1. Multistep training procedure for the SurveyBot3000, which produces synthetic estimates of interitem correlations. (a) Pretraining base model (SBERT). (b) Fine-tuning SurveyBot3000. (c) Validation. SBERT = all-mpnet-base-v2 model.

Finally, @bjoernhommel.bsky.social's and my paper introducing the SurveyBot3000 is officially out in AMPPS. It's a fine-tuned language model that guesstimates correlations between survey items from text alone. Not perfectly, but useful for search, for example.
journals.sagepub.com/doi/10.1177/...