Thinking about and discussing this more with colleagues, I'd really like a continuous time solution, like the Gillespie algorithm but for perfect conterfactuals as defined in thread below. I can't find that this has been done, but somehow believe some rogue chemist has already worked it out

The most romantic sentence in the German language: Möchtest du mit mir eine Kartoffel essen gehen? (Today and tommorrow in Leipzig at Augustusplatz) kartoffelverband-sachsen.de/saechsisches...

Just finished reading this *excellent* article by Gabriel et al. which discusses which effects can be identified in randomized controlled trials. With DAGs!> link.springer.com/article/10.1...

Another record month for bioRxiv - and further evidence the pandemic spike+dip was just that and growth continues. Thanks to all involved and that includes 🫵

Are we doing simulations wrong? This paper convinced me we are. doi.org/10.1098/rstb... Usually we run 2 sets of "worlds" w and w-out intervention. Gives large uncertainties that include negative (harm) effects of interventions that are actually always positive (beneficial)!

I recently read Egan's Incandescence. This is a book that people hate or love. I liked it! It doesn't condescend to the reader, but expects you to pay attention and figure out the major plot links on your own. Also endless pages of aliens talking about geometry. Not going to be a movie anytime soon

Was asked about collinearity again, so here's Vahove's 2019 post on why it isn't a problem that needs a solution. Design the model(s) to answer a formal question and free your mind janhove.github.io/posts/2019-0...

RDM Weekly Issue 15 is out! 📬 - Data Tracking in Neurodivergent Samples Guide from @jcbullen.bsky.social and colleagues - Science as Amateur Software Development from @rmcelreath.bsky.social - Project Management from @manybabies.org and more! rdmweekly.substack.com/p/rdm-weekly...

This covers some important sources of bias that arise from naive control strategies. Of value even if you don't care about the life-satisfaction ~ age association.

@rmcelreath.bsky.social be like

Was reminded of this classic today - we are all Olaf when reading the scientific literature

This only happens to you once

Some of you monsters will like this: Brutalita Sans - a custom font editor. I recommend a font in which every vowel has a random diacritical mark. As a treat. brutalita.com

Just back from this - I was the misfit invited speaker who knows nothing about psychology. It was really educational for me, expanded my view of theory construction problems in the human sciences.

My default explanation for sudden changes in a time series is change in measurement. So glad to see another example, sent to me by a colleague who shares my cynicism. Short, clear explanation of Liquid Chromatography-Tandem Mass Spectrometry. Which is important to the story.

I am preoccupied with a workshop this week, gave a plenary this morning on scientific workflow. I am going to write this up as a paper - but for now here are some slides. Nobody break anything while I am offline please

I cast a wild parlor weasel into the timeline. It’s a long September, but it will pass

Our software paper on SpeciesDistributionToolkit.jl is now published in @peercomjournal.bsky.social - peercommunityjournal.org/articles/10....

🔵 The Carpentries have recently published a new version of Building Better Research Software, a lesson in The Carpentries Incubator designed to serve as a “next steps” workshop after Software Carpentry. Find out more about it and how to get involved at: www.software.ac.uk/news/calling...

Time for something completely different. "This film creatively illustrates the process of protein synthesis through dance, symbolizing the assembly of amino acids into proteins." www.youtube.com/watch?v=3GfW...

Lots of unrealized value in researchers using packages like this to structure and manage projects. We need to integrate with causal inference and calibration/testing tools, so that stat models can be derived in pipeline from questions and mechanistic models, and then tested on synthetic data.

rixpress is now an @ropensci.org package! Link: docs.ropensci.org/rixpress/

SquidSim is the coolest package - it let's you build complex hierarchical data structures and then simulate data from the world you create. The best tool for doing proper power analyses and testing how well your models can uncover the 'truth'. I've been recommending it to everyone!

A quick reminder to all those who care about open research to consider joining us for the FOR2026 Conference in the Future of Open Research at the Technical University of Munich in May 4-6, 2026! Deadline for paper and poster proposals: 30 September opensciencestudies.eu/for-2026-con...

Despite what some adults seem to think, teenagers are fully human.

“Overall, the message from this survey is positive; evolutionary biologists are readily employable outside of academia, generally well-prepared for those jobs, and report high levels of satisfaction” www.biorxiv.org/content/10.1...

💡 Explore how practices from distributed, open-source software can make science more professional, open, and collaborative! Attend the “Science as amateur software development” lecture by Prof. Dr. @rmcelreath.bsky.social on Wed 17 Sept, 9:00. 👉 Register here: www.pretix.osc.lmu.de/lmu-osc/OSSS...

I realize the world is burning, but here is the abstract of a talk I will give multiple times in different formats this month. With the feedback, I hope to write up something substantial on the theme in the remainder of the year.

September has arrived and I do not like it

Dreaming Horse by Franz Marc, 1913 #artbots #guggenheim https://botfrens.com/collections/212/contents/137668

Humans are in fact a venomous species (Figure 6 from "A global database on blowguns with links to geography and language" Aguirre-Fernández et al doi.org/10.1017/ehs....)

Preprint and the website by @dingdingpeng.the100.ci and @vincentab.bsky.social are great, and I got inspired to illustrate Bayesian workflow for model checking and comparison *before* model interpretation following their friendship importance example users.aalto.fi/~ave/casestu...

Nice looking paper

This is one of those probability facts that drives my usual advice to people seeking intution for probability theory: Stop seeking intuition! It's not intuitive, and that's why it is so useful. You can learn examples and reform your intution in time. But better to just trust the axioms and compute.

This "dynamic borrowing" thing is interesting. Like a kind of mixture? I need to read up on it more. Also always nice to see ppl using my software to fit exotic models cleanly that simpler interfaces can't manage. >>

In my dept today we discussed deriving estimators from causal models. This can be opaque, but as a simple e.g. instrumental variable Z for estimating X on Y. Z –a–> X –b–> Y, and X <– U –> Y, where U is unmeasured, a and b are path coefs. We want to know b. U prevents direct approach. >>

Interesting paper that I'll have to spend slow mathy times with later, but had a real record-scratch-freeze-frame moment while skimming...

Here is a PDF of chapter 17 - it's just 3 pages and will fill you with hope and optimism, or so I've been told share.eva.mpg.de/index.php/s/...

It is maybe a little unfair, but as an intro paragraph it is great writing. From Hacking's Representing and Intervening (link in quoted)

Rounding to zero is often a problem in machine learning, which is why we do computations on log scale and use various composition functions. This below (from materials in a paper I am reading) is not how to do it. If you are getting zeros in your prob vector, back up and do the calculations right >

I welcome the death of google scholar so that h-index can also die - a weird proxy metric ("the largest h such that h articles have at least h citations each") that become ubiquitous seemingly because some google engineer added it to a side project

Systematic reviewers have a 10.7% error rate when screening papers for potential inclusion. In other words, about 1 in 9 abstracts are categorised incorrectly (i.e., false inclusion or false exclusion). This really drives home the benefit of having at least two screeners

This in the latest PNAS. Phylogenetics has gone all in on Bayes but knowledge & norms have lagged. I mean posterior often has good frequentist coverage but can't assume it. Also no chain diagnostics in the paper or supplement. Where are our standards?

Ecology needs to overcome siloed modelling 🌐🌎🧪https://www.cell.com/trends/ecology-evolution/abstract/S0169-5347(23)00242-2

It's like an episode of black mirror in which, instead of fixing the poisonous two-sided market that favors oligopoly publishers, funding orgs decide to try out *price controls*.

A cat lives here

Some light reading for the weekend — this is from David Hilbert's 1900 lecture on 23 problems in mathematics. It is comforting somehow that academic German has not gotten much better or worse in the last 125 years.

I've largely ignored these P-curve tests tbh. I posted something a few years ago about how p-values are not in general uniform under the null and found psychologists getting upset at me bc this relates to P-curves, but I never dug into the history. So another paper to add to my reading pile.

Several ppl sent me this extraordinary chart crime. Thank you I love it

🐒🕸️ New preprint! Confused about how to model animal social networks? ASNA can be confusing—but also full of opportunity. We break down 5 common misunderstandings in animal social network analysis and share solutions from behavioural ecology, anthro, stats, & network science. Hope it helps! A 🧵