Hey #statstwitter, why do PPCs seem to be such a Bayesian thing when you can easily do the same thing with frequentist output?

Today would have been Carl Sagan's 90th birthday. He wrote this in 1995.

Fun day of flipping in the rainforest. Coastal carpet python sharing a sheet of tin with a red-bellied black snake.

Hey #statstwitter, since I've never had a single account engage with me on Bluesky, how do we do posterior predictive checks with GPs? I assume you want to do the brms-equivalent of allow_new_levels = TRUE?

Currently seeking PhD candidates to help me with exciting #ecologicalforecasting problems using Stan and #rstats ecogambler.netlify.app/opportunities/

My third blog post, once again on fitting ecological models in Stan.

Generalized linear neural network models statmodeling.stat.columbia.edu/2025/02/04/gen…

After several delays, Department of Government Efficiency has finally posted its purported savings. Why did it take so long to create a simple webpage with a 1000-row table? Who knows! Let's dig in. Headline number: $55B saved. They list the savings per nixed contract. This should be easy to verify then. 🧵

I was just told that eBay is aware of this, but is choosing to not do anything about it. First time I have felt discriminated against since going free. Gee, maybe I should start my own ecommerce website. What'll I call it...?

Milk Boat.

Hey #statstwitter people, what do we think about interpreting coefficients in cloglog binomial regressions? Since everyone hates odds ratios I wonder if interpreting hazard ratios is easier.

Long before LaTeX, most mathematical theses relied on hand-inserted symbols. But in 1968, Dennis Ritchie - creator of C - typeset his entire thesis, subscripts and all, with astonishing accuracy. How exactly he pulled it off remains a mystery. Learn more in this week's paper

I just honest to god, don't even know what we're doing anymore. Every day is just the dumbest most cargo cult slurry getting interpreted however anybody wants. It's so unbearably dumb, and it's literally the state of the art for humanity.

Almost all of bad observational research works this way. Fields have vague priors drawn from lots of real and sometimes hallucinated indirect evidence and then they go back and paper those beliefs with trash regressions. Then they point to the regression to rationalize the prior.

I'll never not find this entertaining.

Me: Decimal points are missing, and please check the identical data. Correction: Decimal points added. Identical numbers are identical numbers. Scientific Reports: Cool. Me: Bro...... pubpeer.com/publications/7…

I'm going to repeat my concluding takeaway: science is hard. And the profession should be constructed in such a way – taking replication and reproducibility seriously, and institutionalizing it, rather than punishing it – to reflect that reality.

The pretty draft is now online (free)