I want to thank Richard Dawkins for making my arguments about popular science writing as a vector for ideological scientism incredibly easy to develop

In our capitalist society, we worship competition But in academia it can be harmful Encouraging secrecy over collaboration, withholding of data, misappropriation of ideas, and even questionable research practices

(2) My 2024 article with Chris Donkin, in which we consider the advantages and disadvantages of confirmatory and exploratory hypothesis tests and conclude that exploratory tests can be more compelling than confirmatory tests. Open Access: doi.org/10.1080/0951...

Perhaps the best irony here, lost on the promoters, is that forensic science actually has really shoddy, mostly non-evidence based, often harmful practices. The police also is generally a beacon of fairness and justice.

Tbh, third party simulations were showing issues with p-curve for quite some time-this shouldn't have come as a surprise doi.org/10.1002/jrsm...

Does anyone know of a direct comparison between EU countries in terms of metrics used to evaluate universities and/or academic researchers? I'm especially interested in whether one can contrast countries that use N publications with those that don't. I suspect it's not so simple though

Learn about catching statistical errors before they become a problem, with two of the best in the business. qut.zoom.us/meeting/regi... (collab w/ @ausrepro.bsky.social)

@xkcd.com again provides for your meta-science slides (link xkcd.com/3117). I do wish we'd stop talking about a "crisis", bc it's been here for generations. Ppl used to pipette by mouth ffs. Enrico Fermi won a Nobel prize for a false result. But that doesn't mean we can't do better going forward.

If you’ve ever wanted to scream “JUST READ MY PAPER!” this is for you. If you don’t communicate clearly and consistently, no one hears your work. Want visibility without sounding salesy? Tell the story, show the impact, and connect with those your science can help. 🧪 #SciComm

6. That’s the dynamic I see it here. To what degree should corporations be liable for harms that they cause as judged by the best available science and a consensus among experts? If I can’t sue Philip Morris for selling cigarettes to kids without a double blind trial to prove harm, we are sunk.

"Thinking time —the time needed to concentrate without interruptions has always been central to scholarly work. It is essential to designing experiments, compiling data, assessing results, reviewing literature and, of course, writing. Yet, [it] is often undervalued." www.nature.com/articles/d41...

Honestly I think this is the most depressing aspect of the current state of science for me. The sheer effort and money that is being wasted on data collection when the study is transparently dead on arrival.>

"The problem with 'ritual science' or 'ritual statistics' is not the use of standard methods but rather a lack of understanding of the conditions under which the methods would be appropriate—or a lack of willingness to consider or interrogate the limitations of the methods being used."

1. Another Friday afternoon massacre of sorts: NIH just rescinded its Scientific Integrity Policy. h/t @lizborkowski.bsky.social

Your dignity honors the bravery of the Ukrainian people. Be strong, be brave, be fearless. You are never alone, dear President Zelenskyy. We will continue working with you for a just and lasting peace.

Conduct rigorous studies rather than 'napkin maths' to support your biggest claims.

This urge to spread bullshit statistics & results in the name of improving science does not instill confidence in metascience. Unless they start developing willingness to reflect on their own practices & take their foundational values seriously, metascientists will harm science more than they help.

AI models extrapolate human biases. In our work, we (with @hplisiecki.bsky.social ) show how easily they catch the bias of judges in sentiment analysis. Not a very surprising result (but large, very significant effect sizes), but it's an important step towards unbiased text analysis tools.

The exploratory nature of the study doesn't make the results unreliable 🙄 It's easier to understand when we realise why the confirmatory nature doesn't make the results 'reliable', unless we define 'reliability' in a very narrow and usually inappropriate scope of type I and II errors.

‘Publish or perish’ culture blamed for reproducibility crisis www.nature.com/articles/d41...

Fellow scientists, let people know you don't live in an ivory tower but a one-bedroom apartment! 🧪 #science is better trusted by people who are perceived to be humble!

The fluoride debate is reminiscent of earlier controversies, such as the debates over smoking and cancer or formaldehyde. The 'pseudo-defenders of science' and their defense of water fluoridation pose a significant threat to the credibility of science, greater than that of populist politicians.

"No evidence of p-hacking" is not "no p-hacking". Especially when the methods used simply do not detect p-hacking, lol. It's like diagnosing cancer with a thermometer.

Setting aside all other related issues that might factor in such an assessment and sticking to the parameters of your framing, I wouldn't think it's any different from heavily relying on methods one doesn't even try to or cannot understand. I find it troubling in other contexts so that carries over.

There are 3 different common blue in many Slavic l. (for ex. Belarusian, Polish) Гранатавы (Be), granatowy (PL) - dark blue Сіні (Be), niebieski (PL) - blue Блакітны (Be), błękitny (PL) - light blue also, there are some regionalisms, as "modry" - Silesian, Czech, like "modra kapusta" (below)

Sometimes the combination of arrogance and ignorance is breathtaking. Yes, US health care has high admin costs; that's because it relies so much on private insurers. Govt-run health insurance has much lower overhead

📤 MetaROR offers a free, transparent platform for metaresearch using the Publish-Review-Curate model, where peer review takes place after publication rather than before. The reviews are open & can be freely reused by journals. Find out about how you can submit your work: metaror.org/guidelines-a...

These two papers, taken together, really cause a rethinking of behavioral economies. Rather than having anomalous risk preferences; it looks like people have complexity aversion to "hard" decisions, especially on valuation, which drives behavioral anomalies. Herbert Simon ftw.

It's the way I would choose if I could decide what science publishing should look like!

The upgraded (better) version of the Likelihood Ratio Test for Publication Bias has been uploaded to GitHub, MetaArXiv, and my personal website! It might be a valuable tool for meta science.

I think the replication crisis is a reality check on our necessary limitations as scientists, and our overemphasis on stats and methods is part of the problem, not the solution.

I’ve developed a new method for detecting publication bias in heterogeneous effects, with initial tests showing it outperforms existing approaches. If you’d like to support a young independent scientist in advancing meta-science, please read, share, or review—your help makes a difference! #MetaSci

Don’t ask whose fault this was. Plenty of time for second-guessing and recriminations. Ask instead, what can you do? For my part, that means telling the truth as I see it, as long as I can. The media will be under a lot of pressure to toe the line; don’t capitulate in advance

Always 💖 Don't make me tap the sign

after years of hearing colleagues & processing their arguments, i remain unconvinced that paid peer review is a solution to any of the problems w the publishing system & the publishers. i do believe it would in fact generate new problems & inequities though. for profit publishers *are* the problem.

They: LLMs can predict the results social science experiments! That's a game changer. Me: Ask better research questions, design better studies, aim to explain, do the damn research! *That* is a game changer.

Should you use Likert or VAS for EMA research? In a new paper led by @jmbh.bsky.social & Alberto Jover, we assigned n~160 in a 2 week study to either 📏1-7 Likert or 🎚️1-100 Visual Analogue Scale and compared results using Bayesian multilevel models. Brief thread with results.🧵 🧪 #psychpsysky

Very important editorial www.science.org/doi/10.1126/...

Has anyone found examples of academic fraud or significant errors in qualitative research? I’d be curious if #MetaScience could help. I see the emphasis now is on quantitative research. maybe @markrubin.bsky.social ?

something we can all do to make things better is to stop doing kitchen sink citations for random claims. let's write in a way where we only need to write things we've read, understood, and want to build on. read wider (more diverse) and deeper, cite less for the optics and more for the substance.

🧪 Scientist? 🦋 New to BlueSky? 👀 Wanna find posts by other people in your area and use the right hashtags in your own posts? 📃 Here’s a list of +200 science-related feeds and their hashtags. #AcademicSky #PhDSky

Since publication bias *exists*, you can safely assume that (some) null results will be ignored in any meta-analysis. The trick is how you deal with it :)

Theory matters. Models matter. Assumptions and assumption violations matter. Methods are not equivalent in their assumptions or statistical guarantees. We do not have the luxury of making inference in a vacuum. That's also why reproducibility/replicability doesn't mean what people think it means.