A nice reminder that it's hard to know how often simultaneous discoveries happen in science. Any measures based on scholarly literature will be severe undercounts due to secrecy or simply unpublished work

New piece w/ James Evans in Science explores what we call 'science after science', an era where our ability to control nature may exceed our ability to understand it; a new struggle to sustain curiosity & understanding under AI's predictive dominance. #ai #science www.science.org/doi/10.1126/...

Now published at PNAS ‼️ w/ @innovation.bsky.social How does peer reviewer diversity affect fairness in peer review and the direction of published science? We find a "geographical representation bias" in 60 STEM journals published by @ioppublishing.bsky.social. www.pnas.org/doi/10.1073/...

This paper explores academic peer review as a source of knowledge transfer and learning for the reviewers themselves. spkl.io/63328A1crC @innovation.bsky.social @jamesmzd.bsky.social

Bad news for economists...

Wow - an important read for #academicsky #scisky #medsky, about a study on professional social media activity by female scientists vs male, performed using @altmetric.com data from 2013-2018 (so, pre-Bluesky) Summary: news.umich.edu/fewer-women-... Paper: www.nature.com/articles/s41...

Do you remember Francesca Gino's claim on her website that HBS analyzed the "wrong data file" in their investigation, and that a "real file" proved her innocence? HBS is now claiming that the "real file" was fabricated by Gino... and thus that Gino's claim was defamatory.

In 1981 NYT published a piece on the then-new citation index, and how it was being used "unintentionally" for performance evaluations

🏮NEW WORKING PAPER ALERT🏮 Scientists collectively spend tens of millions of hours peer reviewing each year, mostly for no $$. Why?? In new work w/ Charles Ayoubi and @innovation.bsky.social, we observe a private benefit for participating in evaluation: 💡💡Learning💡💡

Why do we peer review? In a new paper, we establish a self-interested reason: learning. With a quasi-experimental design applied to admin data from 55 journals, we show reviewing a paper doubles chances of citing it in future work!

Men are much more likely to self-promote their papers on Twitter/X than women

Seems relevant: politicians are more likely to vote for war when they don't have sons who would do the fighting www.journals.uchicago.edu/doi/abs/10.1...

First-ever ranking of journals by impact factor, published by Garfield in Science in 1972. Science is ranked #77, impact factor 2.99 Nature is #114, impact factor 2.34

Going to the hospital because I broke my wrist smashing the endorse button: www.understandingai.org/p/i-got-fool...

Conventional wisdom says interdisciplinary research is valuable but harder to get through peer review (need to please diverse reviewers, etc). @sdxiang.bsky.social Daniel and I partnered with @ioppublishing.bsky.social to test this wisdom and add nuance

Schism 2: electric boogaloo

Periodic reminder that economists would sacrifice half a thumb for an AER pub

Does professor quality matter for training the next generation of researchers? Or will good PhD students do good research either way? Quality matters

There used to be very little turnover in NIH sections, now there's a lot more

Hmmm, maybe enforcing antitrust is a good idea 🤔

When you ask for the data

The academic job market has changed just a bit (from an interview with Robert Barro)

Can we stop diffusion of particular AI technologies to countries we don't "like"? ChatGPT is useful case. OpenAI *still* prohibits access in major countries. But do these prohibitions actually work? New paper by @honglin-bao.bsky.social and @sunmengyi.bsky.social suggests they don't! 1/3

Marc Andreessen really, really hates universities, which seems to be a pretty common sensibility from billionaires who have spent too much time online. www.nytimes.com/2025/03/18/o...

Anecdotally, many advisors want their PhD students to go into academia and treat them worse if they want industry job. Does intention to go into industry *cause* less mentorship? Maybe confounded by something else? Inna, Austin, moi ran audit experiment to find out! (1/3)

paying $250 for peer reviews didn't have much of an effect journals.lww.com/ccmjournal/f...

Scientists in industry are more likely to rely on the papers of other industrial scientists than similar papers by academics. Could be just an awareness issue, or maybe industrial papers are more credible?

People don't like *micro* posts that much, they prefer slightly longer pubsonline.informs.org/doi/full/10....

What may seem like round #1 of this fight is actually round #2 www.washingtonpost.com/national/hea...

How do scientists choose which topic to study? Decades of studies on this, but they're dispersed across fields and not synthesized. Fortunately for us, Sidney @sdxiang.bsky.social has written a fantastic review of this literature, focusing on econ and soc lit! direct.mit.edu/qss/article/...

😳

Twitter gets more engagement per science post, but in terms of number of science posts, Bluesky has almost caught up www.altmetric.com/blog/bluesky...

Good news. The file drawer problem, of insignificant results not getting written up and submitted, is shrinking. #openscience #reproducibility #NHST #metascience 🧪 www.pnas.org/doi/full/10....

The prominent LLMs are almost surely trained on pirated content www.nber.org/papers/w33598

"We measure VC narcissism based on the size of the signature... appointment of a highly narcissistic VC leads to an overall deterioration in research and teaching performance" www.sciencedirect.com/science/arti...

Even JP Morgan Chase (!!!) thinks paying to publish is a scam

I look forward to citing this paper without reading it!

How accurately do citations reflect the original research? Do authors truly engage with what they cite? In a new study, we analyze millions of citation sentence pairs to measure citation fidelity and how it varies depending on authors’ engagement with prior literature. arxiv.org/abs/2502.20581 ⬇️

Guest post: Should Zimbardo’s Stanford Prison Experiment be retracted? asks Augustine Brannigan

Does "spinning" null results to make them sound significant work? Yes. ascopubs.org/doi/full/10....

Some scientists who use ChatGPT for coding have... odd beliefs about how it works arxiv.org/abs/2502.17348

Hello Bluesky! We are ICSSI: The 2025 International Conference on the Science of Science and Innovation Save the date! 🗓️ JUNE 16-18, 2025 ✍️ in beautiful Copenhagen, Denmark www.icssi.org

- The shortest time between a basic science discovery and a drug based on it appearing in the market is 7 years. Mostly it's 11+ years. - Only 5 of the most important drugs were developed with "essentially no input from the public sector". onlinelibrary.wiley.com/doi/abs/10.1...

there’s been a huge rise in survey experiments in these journals (1 in 5 papers in the most recent years)