Hidden Markov Models - Lecture B10 of Stat Rethinking 2026. Hidden state models, inference of latent strategies, time series, is the president dead?, capture-recapture and demographic inference, Guerilla Bayesian Workflow. This is the final lecture for 2026. www.youtube.com/watch?v=fuon...

This is nice. > library(broman) > library(rethinking) > data(Monks) > excel_fig(head(Monks))

A student's model was sampling a scale parameter poorly. A common problem. This time I fixed it by *making the model more complex*. The intercept and slopes were stratified by a 3-level categorical var, but the individual effects were not. Thousands more parameters later, chains are better.

Is there some open source spreadsheet that enforces and supports all these recommendations? I'd love to read in a methods section: "We entered and managed all data using OpenSheet with Broman-Woo guardrails."

It's St Patrick's Day. Again here is my pocket guide to the Peoples of the British Isles: Anglo-Saxon | Gael

Happy St. Patrick's Day www.youtube.com/watch?v=c-4a... Chris McMullan, a great piper and (for a time) my teacher.

This is exactly how I feel when I cycle

Thanks to all for the HMM examples. The stats community here has grown to be very useful. I appreciate you all.

“we are trapping 21st-century minds inside 19th-century academic hierarchies.” This essay gets a lot right about problems with German science. I would add that the hierarchies and precarious contracts lead also to systemic abuse and scientific misconduct. open.substack.com/pub/realimag...

“we are trapping 21st-century minds inside 19th-century academic hierarchies.” This essay gets a lot right about problems with German science. I would add that the hierarchies and precarious contracts lead also to systemic abuse and scientific misconduct. open.substack.com/pub/realimag...

I only knew Chris Sims' work, and I wish I had got to meet him. But to the extent that we know something of each other through what we put out in the world, here is my personal biased favourite of his: Why Econometrics Should Always and Everywhere Be Bayesian sims.princeton.edu/yftp/EmetSoc...

The @dagophile.bsky.social has a new short (10min) video up about race and gender in causal analysis. It's short and clear, echos same themes as in the two causal analysis of discrimination lectures from my most recent course. www.youtube.com/watch?v=UzpR...

Do you struggle with missing data in your statistical analyses? Come join our Summer School "Solving Missing Data Problems in R"! We will tell you when you can safely use complete case analysis, and when more dedicated strategies are required! utrechtsummerschool.nl/courses/data...

My wife bought us a new painting by Pia Sinkkonen named "Taasko yövuorossa?" which translates to "On the night shift again?", but in my head I call it "1. Draw some circles. 2. Draw the rest of the funny owls."

It may be the cold medication talking, but I think I've found my hidden Markov model example for my new lecture.

Another robo-grouse drop. If you didn't see the thread yesterday (quoted below), I am looking for unusual examples of hidden Markov models to use as examples in a new lecture I am writing

History is full of regime-change attempts that assume removing a leader will spark quick transformation. More often, it collapses authority and produces fragmentation, conflict, and long instability. smallwarsjournal.com/2026/03/12/w...

📢 REMINDER: We're hiring a #DigitalHumanities Scientist to join our interdisciplinary DH Lab in Rome! 🔎 Focus: Digital art history, AI & machine learning, cultural heritage data, collaborative research ⏳ April 30, 2026 🔗 www.biblhertz.it/en/opportuni...

So what's the most interesting, or weirdest, application of hidden markov models (HMMs) that you've seen? I once built an HMM for sage grouse mating behavior, fit to data collected by a robotic grouse (pictured doi.org/10.1086/703518). Looking for other examples for a new lecture on HMMs.

Catching up on reading and this paper is pretty cool. I need to take another slow pass and really do the derivations myself. But I love it when there is yet another way to derive/justify a procedure I am already using.

Here we go again: Lecture A10 of Stat Rethinking 2026 - Unmeasured confounds, confounded mediation analysis, collider bias (again), masked discrimination, sensitivity analysis and the importance of making strong assumptions, how to defend yourself against a hostile scientific literature.

With @saramitri.bsky.social, @sonjalehtinen.bsky.social and L. Lehmann we’ve launched the UNIL Center for Theory in Ecology and Evolution @unil.bsky.social🇨🇭 To kick things off, we’re offering short visiting fellowships for theorists in ecology & evolution. Apply & pls RP 😀 tinyurl.com/2wem36zz

Curious about how psychological theories are built, tested, and refined in practice? This summer, the Theory Methods Society is launching the very first edition of the summer school Theory Building in Psychology at the University of Amsterdam (July 6–10, 2026). theorymethodssociety.org

Last week, I mentioned this in passing in a workshop: In 1938 Enrico Fermi won a Nobel Prize for discovering two new elements of the periodic table. Lise Meitner shortly showed that Fermi was mistaken and instead had produced known lighter elements by fission. She did not win a Nobel prize.

"If a double-blind trials is good, a triple-blind trial is better." The "triple-blind trial" in Greg Egan's short story "Blood Sisters" raises the ethical problems of double-blind trials (withholding effective treatment, deception) to next level: Don't even tell people they are in a trial.

Lecture B09 of Stat Rethinking 2026: Generalized Linear Madness. Two examples of analyses that build up from scientific models first. #1 allometry predicts weight from height with ZERO parameters. #2 Vibrant ODEs for time series instead of cold dead lag models. Fully coded and animated with love.

Like many methodologists, I often find myself arguing with someone that their method/epistemology makes no sense. That someone thinks I am arguing against their *conclusions*. But I am arguing they have no justification for their conclusions. Is there a philosophical term for this problem/confusion?

"Evolution Evolving is a landmark book written by stellar scientists who have brought together numerous interesting and relevant examples that show why development matters for evolution." Thankyou @kampourakisk.bsky.social for your wonderful review in American Biology Teacher. See shorturl.at/2Gso8

Blog post on a new paper (w/@dingdingpeng.the100.ci & @statmodeling.bsky.social) reflecting on the value of multiverse analysis 10 years in. Like many attempts at science reform, it provides clear conceptual value but is tricky to theorize & can even be counterproductive if taken too seriously

Max entropy is the statistical version of "Measure twice, cut once." Multiverse analysis in practice is "Cut a bunch of times, and pick the ones you like the best."

Every time I am forced to use Matlab I can only conclude it is a language designed by and for sociopaths Shout out to this @olivia.science blog post that I sent out on a neuro listserv back in grad school, that a prof there is apparently still mad about neuroplausible.com/matlab

“la statistique bayésienne” — everything sounds classier en français

A biologically parameterized model of growth and honest measurement error in the mix. This paper is a quantum leap in handling diverse and messy cross-cultural and archaeological data on human growth, where papers usually fit arbitrary polynomials and ignore measurement error. Link in quote

Has anyone tried teaching *only* Bayesian inferential #stats for undergrad psych (or other UG sosci programs)? Like completely foregoing NHST, confidence whatevers, p values, CLT... I'd be very interested in seeing a syllabus of such a course/ program!

So I gave this workshop today and I think it went pretty well. The best comment afterwards was "thanks for presenting statistics as someone who is not dead inside". I will aim to write this up as a blog post or preprint when I get some time (after teaching finishes later this month).

*drums in the deep* work flow . work flow . work flow *growing closer* Work Flow . Work Flow . Work Flow *menacing now* WORK FLOW WORK FLOW WORK FLOW

my course notes on a bayesian workflow for (single agent) cognitive modeling are now fully revised and online: fusaroli.github.io/AdvancedCogn... Predictive checks, updating checks, sensitivity analyses and simulation based calibration in @mc-stan.org Feedback is very welcome!

*deep inside the pentagon* Analyst: Sir, all is go but we still need to disable the statisticians. They might be trouble in some tail scenarios. Chief: Agreed. What do you suggest? Analyst: I know just the thing... *goes online and starts posting about interpretations of confidence intervals*

New postdoc opportunity in our lab!

A Rock Pigeon with a threatening aura

I know armageddon is starting but I am giving a workshop tomorrow

Lecture A09 - I get mildly ranty about the low quality of papers on discrimination and somehow also introduce generalized linear models for events and illustrate post-stratification. The theme continues next week with modeling sensitivity to unmeasured confounding. I will try to be less ranty.

Yes please let's normalize errors, their discovery, their correction, and their integration into the literature. "Corrections Not Found: Post-Publication Integrity in Usable Security and Privacy Research" (osf.io/preprints/ps...)

🎉 It's livestream day 🎉 Join me at 1700 (CET; 1600 UCT) today for two hours of GAM goodness 🤤 📽️ Youtube: youtube.com/live/A9U8e1K... Hit the Notify me 🔔 to get a reminder when I go live #RStats #mgcv #statistics #GAMs #DataScience 🧪

good morning everyone project your personal imposter syndrome onto this gif ur welcome

Red Oval by Vasily Kandinsky, 1920 https://botfrens.com/collections/212/contents/137250

What a pre-conference workshop on Bayesian statistical methods—a powerful toolkit for our community! 📊 Many thanks to Daniel Redhead and Ramona Roller for guiding us through STRAND and to Michael Chimento for the introduction to STbayes. And thank you to all participants for their energy! 🙌

I told several non-German colleagues today that the German word for placenta is "mother cake" (Mutterkucken) and they all hate me now

Sometimes I assign applied statistics homework problems that are pretty difficult. Students spend hours over days on them. But you know what is much much harder? Research