I need to be off coffee for 48 hours for a (non-scary) medical procedure. I'm currently at the 24 hour mark. Pray for me.

Yes, unnormalized posterior density function, you _are_ enough.

“The “strongest risk factor” is not the form most related to harm. It is the form least redundant with whatever else is in the model.”

This is great - another reference for the "mutual adjustment isn't a tool you should reach for, actually".

posterior R package has now JOSS paper you can cite, too joss.theoj.org/papers/10.21... with @paulbuerkner.com , Jonah Gabry, @mjskay.com

Seems as good a time as any to spruik the #RShiny app my doctoral supervisor and I made for the #rstats and #econsky Bayes-curious among you! Useful if looking to get prior distribution parameters from published literature to use in your Bayesian or decision analytic models: osf.io/preprints/os...

Have you used Bayes factor to test point-null hypotheses in Hierarchical models and wondered: 1. Bridge sampling (BS) or Savage-Dickey density ratio (SDDR)? 2. If SDDR, which density estimation method? 3. How many MCMC samples? 4. What's the cost-benefit trade-off? This may be of interest to you.

New paper "To select or not to select: predictively consistent priors instead of model selection" with Anna Elisabeth Riha, Leevi Lindgren, @davidkohns.bsky.social, @paulbuerkner.com arxiv.org/abs/2606.22850 Model selection is not a substitute for building good models in the first place 1/

Apparently posteriors are sensitive to priors!* Specifying informative priors is a sign of a maturing field! I hope our new paper nudges psychology toward complex Bayesian models — based in formalized theories — and away from "default" priors that lack theoretical or practical grounding.

So many gems in here 💎💎💎

I'm glad we've collectively decided to ignore The Muppets' Wizard of Oz.

New paper out now! We (w/ @rmcelreath.bsky.social and @gregork.bsky.social) start from the observation that among adolescents in Leipzig, the gender gap in life satisfaction has doubled over time and ask: why?> link.springer.com/article/10.1...

There are apparently two Mattan S. Ben-Shachars in the weights! (They're both me...)

#OTD 1937 Jerzy Neyman introduces the confidence interval in a talk given at the Royal Statistical Society. Listeners are not enthusiastic: in the vote of thanks, AJ Bowley wonders aloud if confidence intervals are just a “confidence trick”, doubts their validity & if they were any possible use.

Is this a joke? Yes. Is it also an opportunity to demonstrate how extrapolations that incorporate uncertainty are far better behaved than those that don't? Absolutely.

I’ve charted American hegemony during the Iranian campaign

Finally out in Psychological Review (psycnet.apa.org/doi/10.1037/...), our update to Signal Detection Theory. We show that contrary to the prevailing Gaussian assumption, evidence distributions in recognition memory are likely minimum extreme Gumbel!

I'm excited to announce a long overdue update to {tidyclust}! This release brings support for 3 new model types: DBSCAN, Gaussian mixture, and mean shift models. We also now support parallel processing with {future} and {mirai} opensource.posit.co/blog/2026-06... #rstats #tidymodels #dataBS

#statsmeme? (Prompted by a conversation with a student who's double majoring in stats-econ and said that stats courses make more sense to him)

Is the Microsoft Office Ribbon causing people to have fewer children? A provocative new working paper explores the persistent drop in birth rates since Office 2007 was introduced nearly two decades ago.

The third edition of the Open Handbook of Experience Sampling Methodology is out! The book includes 14 chapters that cover all aspects of designing, running, and analyzing the data from an ESM/EMA study. The book is freely available here: www.kuleuven.be/samenwerking...

Accidentally color coordinated with my slides yesterday 🤭 (The actual colors were blue and green, but thanks to the projector being old, my students got a good laugh)

The 'Specialized data analysis capabilities' section rings too true - Claude is not built with co-analysis with a human in mind. It tends to build standalone scripts and functions with very low readability for the human analyst.

Me at work, every day. #RStats

From @tiffanytimbers.bsky.social & @minecr.bsky.social in @plos.org #Computational #Biology | Ten simple rules for teaching data science | #Education #Bioinformatics #DataScience #PLOSCBTSR CC/ @carpentries.carpentries.org | 🧬🖥️🧪🔓 ⬇️ journals.plos.org/ploscompbiol...

Did you know that you can use the #easystats #rstats {modelbased} package detect missed modeling by adding partial residuals to your plots? Check out the vignette on plotting options that shows you how to check your model: easystats.github.io/modelbased/a...

playroom. noun. A lovely decorated room for storing the toys and games that are played with in the living room. (From my kids' dictionary, I guess 🤷‍♂️)

Did you know you can search for R packages by what they do, as opposed to just their names? Look! 👉🏼 rwarehouse.netlify.app @kylieainslie.bsky.social made The Warehouse for just this purpose. Send it to a new R programmer today ❤️ because it's the resource we all wish we'd had at some point #rstats

If anyone is interested in this but missed the talk, I'm afraid it wasn't recorded on this occasion, but the slides are now available on our webpage: ciig-seminar.github.io And Vanessa is more than happy to answer questions by email!

A new #bmm release is available on #CRAN. With this version, #bmm now covers basic response time models. This release contains three new models plus several usability features. Here's what's in it 🧵 venpopov.com/bmm/ #rstats #bayes #openscience

If you want to keep updated with all the quarto work that I do, then the best place is properly this page on my website that I try to keep as up to date as I possibly can! It contains links to all my quarto repos, talks, themes, extensions, books and more emilhvitfeldt.com/project/slid... #quarto

We've done it folks. It finally happened! Reviewer 2 is asking for a new analysis plan, *and* that the new analysis plan be pre-registered, and asking that we show that the sample size is powered for this new analysis plan. 🔥🔥🔥🔥

A very interesting read! As someone who was very active in the small community of Hebrew speaking R users, it saddens me to see the community "grow" silent.

Now that's what I call a spicy opeing: Understanding mechanisms is not a research question.

Missing data sucks. Data are MAR - FIML and multiple imputation give very similar estimates, but pooled standard errors (Rubin's rule) have much larger standard errors. Does this indicate low quality of imputation? (Uni- and bi-variate visual inspections were good) How should I proceed? #stats

"Is my toaster intelligent and conscious? Maybe!" Psychologists, philosophers, neuro scientists: well actually, we've been debating this for... "Like what do those words even mean?? lol" If you would like to engage with the lit... "Nah, I'm good. Here's a 5000 word substack with no references."

If it is indeed Microsoft Defender getting in the way, there are instructions here for fixing it (by getting a fresh set of definitions from Microsoft): github.com/posit-dev/as... Note that you may need to reinstall the Posit Assistant extension afterwards to restore the file Microsoft deleted.

🧑🏻‍🏫 Do you teach statistics? 📊 Have you ever taught ANOVA? ⏰ Do you have 5 minutes to spare? Then participate in our survey and help us improve statistics education! run.pavlovia.org/pavlovia/sur... #stats #edusky #mathsky #scisky #academicsky

Me liking this knowing damn well I too do this too often 😬

Anyone else recently updated #Positron and no longer have a working Posit Assistant? (Endless "loading") @posit.co

I teach marginaleffects in my intro stats class at Berkeley public health. I've heard some pushback that students don't learn how to eg exponentiate a coefficient from a logistic regression. They don't understand this is a feature and not a bug. Keep up the good work @vincentab.bsky.social

As a course assignment, I had students rewrite the discussion section from an actual paper (using track changes) and incorporate information from a DAG and bias analyses they ran. Some of these are among the best discussion sections I've ever seen!

The {tidymodels} hex sticker for the {censored} package is perfect 😂 (but also, I'm a little uncomfortable?)

{easystats} is not only highly functional and powerful, but their help is also a goldmine of accessible #statistics advice. Here explaining how to address high VIF values (hint: generally not by removing predictors, as often done) easystats.github.io/performance/... Thanks @easystats.github.io

Psychologists doing clustering be like: #statsmeme

Difference scores strike again! Here are those results on the full scales - what's really going on is that all religions (inc. atheism and agnostism) excluding JW should be left but also joined (??).

In the frenzy of exciting #useR2025 talks, I forgot to mention that I also gave my talk. 😆 "Adding models to {tidyclust}", slides here: kbodwin.github.io/Talks-and-Pr... PC: @healthandstats.bsky.social

Iran is anxiously checking manuscript central where its submission is listed as Reviews Complete - Awaiting Final Decision. 🤞