my husband put on nirvana and my children instinctively invented moshing from first principles

I'm looking forward to teaching a 4-hour short course on Bayesian sensitivity analysis methods at the American Causal Inference Conference (ACIC) 2026! Register here: sci-info.org/annual-meeti...

“the proportion of SpO2 measurements >94% despite a paired SaO2 ≤92% ranged from 5.3 to 35.3 percentage points higher for patients with darker skin tones than for those with lighter skin tones (7.6-62.2% v 1.2-26.9%, rate ratio 2.3-7.1).”

Fantastic to see this published at last An objective assessment of the efficacy of home pulse oximeters on different skin tones Ping @iwashyna.bsky.social Open to comments from all #foamed #medsky www.bmj.com/content/392/...

“They who stick everything they have measured in the model will create a colliderscope of patterns.” Sayings of Confuseus

An amazing part of 2025 was the growth of ordered beta regression. The paper was cited 94 times in 2025 across many fields. Super grateful to everyone who has used it and cited it. Thank you! and don't forget your ordered beta merch 😆 #rstats rmk-designs.printify.me

‪It has a name now 😜 Many thanks to Ken for agreeing to put his good name to my...artwork. The image is in the public domain (CC 0), but citations to the linked documents are warmly welcomed. ✅ zenodo.org/records/1808... ✅ pubmed.ncbi.nlm.nih.gov/24452418/

Am I crazy? What problem does this incredibly expensive machine solve?

Just watched the new Knives Out and I think it's really important you know that the scene in the Seminary's Gym is filmed in the same place Rick Astley filmed the music video for Never Gonna Give You Up. I saw the window tracery and immediately made my friends pause the film so I could tell them.

Perfect example of the Spandex Principle TM. Just because you can, doesn’t mean you should. I imagine intubating a manikin wouldn't be too dissimilar from a human given the setup time.

Happy Christmas everyone! I'm thrilled to recieve a copy of The Ghost Map, by Steven Johnson, which tells the story of a certain 1854 cholera outbreak. The perfect #EpiSky geek gift 🤓

The interesting thing about the Table 2 Fallacy is that it works even if you're only looking at regression as a prediction model: standardized coefficients (betas) aren't equal to semi-partial correlations. #stats

Big brewery actually

THIS IS REAL

#EpiSky #Stats #PublicHealth

Looking forward to the next iteration of this work, “F- it, we’re doing four robustnesses.”

Does anyone have any good resources for the identification and adjustment for positivity violations analysis of longitudinal, continuous, time-varying exposures (binary outcome & tv-confounding)? #statsky #rstats

Did you know, that in 'honour' of wakefield, there is an #rstats package named after him that generates fake data? Makes me proud to be an R user.

Yes - on track for next year

Hello #rstats! 👋 Do you have an #rstats package or skill or strategy you love and would be willing to share with the @rladies.org community? We are looking 👀 for a couple more speakers 🔊 for the 2025–2026 R-Ladies St. Louis season. Reply here and we will contact you to find a date! 💜 💜 💜

Does anyone have any good resources for the identification and adjustment for positivity violations analysis of longitudinal, continuous, time-varying exposures (binary outcome & tv-confounding)? #statsky #rstats

personally i simply refuse to acknowledge the existence of any Star War which occurred after 1983

No. I don't teach at the moment. Search committees are free to slide into my DMs, though

As is always correct and in accordance with all the laws of the internet, I just watched Tom Holland's epic Lip Sync Battle performance, and now I am sharing it with all of you so you can do the same.

This is related to a private conversation I had recently. IMO, if you want to use Bayesian statistics in your professional work, either (a) work through at least one Bayesian textbook yourself or (b) collaborate with someone who has already done that work themselves. Take your data seriously.

Do you teach #rstats? Do your students complain about how lame and old-fashioned dplyr is? Don't worry: I have the solution for you: github.com/hadley/genzp.... genzplyr is dplyr, but bussin fr fr no cap.

once again seeking others' opinions to guide my life choices: of those who dropped out of a PhD or converted to Masters, did anyone subsequently complete a PhD with a different topic and supervisors? thoughts for or against this path?

New post, on the time Einstein had his first run in with the US peer review system: kucharski.substack.com/p/einstein-v...

"I DON'T NEED YOU TO FUCKING REWRITE WHAT I'VE JUST WRITTEN!"

This is an excellent point that generalizes. Researchers often defend suboptimal practices by referring to future studies with better designs. But: Why would anybody run those studies when you can just throw a bunch of variables into a regression and make sweeping "preliminary" claims?

My own TLDR for the message from this paper: statsepi.substack.com/p/sorry-what...

We wrote an article explaining why you shouldn't put several variables into a regression model and report which are statistically significant - even as exploratory research. bmjmedicine.bmj.com/content/4/1/.... How did we do?

This is excellent, and I'm so glad this paper was finally written, and so clearly as well. I basically write an equivalent every time I am consulting with someone proposing this type of study, and I'm so glad I can save my effort and just send them this instead!

My journey in anaesthesia research @bjajournals.bsky.social @thermh.bsky.social @anzca.bsky.social @anzcactn.bsky.social @unimelb.bsky.social topmedtalk.libsyn.com/perioperativ...

after some cool theoretical background, I think these slides sum up Thomas Lumley's talk well especially "resist mathematical defaults", "rank tests have stronger assumptions than the t-test" & "ordinal data is not the easy option" a very different take from Frank Harrell's writing on the subject!

"Pavlovian causal inference": the inherent human tendency to infer causation from mere association, even when we know we ought not to.

Some days I get a treat on my walk in to teach…

Key point from my invasive CO talk the other day #CHEST2025

The recording is now available so that you can confirm that I indeed have a German accent and color-match my outfits with my Zoom background. youtu.be/YL0co26ng-g?...

Does anyone know a workaround for modeling residual correlations in multivariate models in {brms}? It only works for Gaussian and Student distributions right now, but maybe someone has a hack… #RStats (Maybe @solomonkurz.bsky.social?)

People sometimes ask why we want to rate DAGs. Couldn't we just rate statistical models instead? We like to explain it with capybaras and bicycles.

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"

I read and write, I explore and I question, I design and script and analyse, I interpret and communicate. I do this to train my mind in the hopes of one day generating new knowledge. New knowledge that might even be useful, and that no algorithm can yet be trained on.

I feel I deserve the Nobel Peace Prize for the judicious use of a small syringe of saline with a red sticker, rather than a full on rant when the surgeon asks for muscle relaxation just before the end of the case.

yep 😞

I clearly had a lot of thoughts/feelings from reading Terry's blog post, so I wrote one of my own It ended up being a bit of a stream-of-thought post but I think it captures some of my worries benharrap.com/post/2025-10... #statssky #episky #academicsky

Just because an LLM can produce a report with various figures & charts doesn't mean it is good at statistics. Because good statistics is not about producing code. It's about deep knowledge of study design & conduct. In my opinion, 95% of all data science problems come from poor questions & design.

One reason I oppose associational or other non-causal language in research questions is it lets people think they don't need causal inference to answer causal questions. (When the underlying question is causal, of course. Which it almost always is.)