@frederikaust.com
Tenured researcher at the University of Cologne, formerly Postdoc at the University of Amsterdam. Interested in Bayesian statistics, mathematical models of learning/memory, and computational reproducibility. I develop R packages and obsess about coffee
Are episodic and semantic memory really that different? Using closely matched tasks, we found no substantial neural differences between recalling personal experiences and general knowledge: https://www.nature.com/articles/s41562-025-02390-4
Our institute is hiring
1. an assistant professor (with TT) for Social Psychology (focus: environmental psychology)
ohws.prospective.ch/public/v1/jo...
2. an assistant lecturer (with TT) for Experimental Personality Psychology
ohws.prospective.ch/public/v1/jo...
Bots have made their way to Prolific experiments. Our lab has stopped online testing of adults entirely now for this reason - we want to know if what we study is real. Probably data collected 2-3 years ago are ok, but moving forward we just can't know. www.pnas.org/doi/10.1073/...
Are you passionate about teaching #statistics? Would like to have a permant #job in #academia? Are you fluent in #English and #German? Do you want to become a senior lecturer @uni-graz.at?
Apply!: jobs.uni-graz.at/de/jobs/3cb1...
Postdoc Position in Social and Organisational Psychology at University of Cologne
Application deadline: 15 February 2026
More info: jobportal.uni-koeln.de/ausschreibun...
#SocialPsyc #PhDSky
Approximate Bayes Factors Are Approximate E-Processes: https://osf.io/97e2w
On why the motherhood penalty may be overestimated. Simply put: The identifying variation comes from women having kids at different points in time, which might not be a valid counterfactual. open.substack.com/pub/plausibl...
"Theory and Practice of Bayesian Hypothesis Testing with JASP", the last of four summer workshops, is now open for registration. Join Richard Morey and myself in Amsterdam or online! (August 27 & 28)
jasp-stats.org/2026/02/09/h...
I made a map of 3.4 million Bluesky users - see if you can find yourself!
bluesky-map.theo.io
I've seen some similar projects, but IMO this seems to better capture some of the fine-grained detail
Interesting! I wonder if further supply increases (from the current low low levels) would solve both cream-skimming and the difficulty of finding a therapist for the too-ill? In other words, why in your opinion are supply increases insufficient to address the problem?
🧵 New WP with @kaimiele.bsky.social. We document 2 unsettling patterns in mental health care: 1) despite universal coverage, few individuals with mental illness receive guideline-consistent treatment, and 2) the more severe the illness, the lower the treatment uptake & the longer the wait times. 1/n
I hate this: "The Editors have retracted this article... the Mini-Mental State Examination (MMSE) instrument was used without proper permissions" link.springer.com/article/10.1... The MMSE is a series of 11 simple questions. A perfectly good paper is retracted because you can copyright 11 questions
as I'm revising my course materials, I keep stumbling upon cool @mc-stan.org developments.
Current favorites:
1. your model has funnels and you exhausted reparametrization ideas: metric = "dense_e" makes your HMC learn about covariance btw parameters. Sloooow, but effective!
1/
"Repetitions lead to better memory." Sure - but how? Our new paper led by former lab member @philippmusfeld.bsky.social shines new light on this question, and challenges some longstanding assumptions. Now published in Perspectives on Psychological Science.
journals.sagepub.com/doi/10.1177/...
+++ Teilt und unterschreibt unseren offenen Brief an die Hessische Landesregierung! +++ Wir werden die bevorstehende katastrophale Verschlechterung unseres Hochschulsystems nicht hinnehmen und fordern eine Neuverhandlung des Hochschulpakts! #nocuts #nocutsinhessen cryptpad.fr/pad/#/2/pad/...
A new study from Anthropic finds that gains in coding efficiency when relying on AI assistance did did not meet statistical significance; AI use noticeably degraded programmers’ understanding of what they were doing. Incredible.
Screenshot of Prism - I asked it to create a methods, participants, results, and discussion seconds. It did so within 54 seconds...
OpenAI just released Prism, a LaTeX editor with embedded ChatGPT for free.
Writing a paper has never been easier.
Clogging the scientific publishing pipeline has never been easier.
It took me 54 seconds to write up an experiment I did not actually conduct.
prism.openai.com
What's Missing? 300 extensions and counting. Help shape what comes next.
What Quarto extension would make your workflow easier?
There are already almost 300 extensions browsable at m.canouil.dev/quarto-exten... and installable via Quarto Wizard at m.canouil.dev/quarto-wizard/.
Drop your ideas below.
If there's enough interest, I might just build it.
#Quarto #OpenSource
For a weekly seminar class, I'm looking for pairings of psych popularization with scholarly take-downs of the underlying research. I want students to see examples of how shoddy research gets hyped.
Very interesting. How does this differ from R Analyst and R Programming Compiler?
Promotional image for webRios showing the app icon and an iPhone displaying the R console. The console shows example R commands with syntax highlighting: basic arithmetic (1 + 1), a print statement saying 'Hello from iOS!', a warning message in orange reading 'Uh-oh, I'm in the Apple-verse?', an error message in red with the HAL 9000 quote 'I'm sorry, Dave. I'm afraid I can't do that.', and a plot command.
webRios is live. #rstats on your iPhone and iPad.
I showed native R compilation on #iOS last week. Shipping it is another story (thanks, GPL). This version uses #webR 's #WebAssembly build instead. Different tradeoffs, but this one clears App Review.
apps.apple.com/us/app/webri...
The Iowa Gambling Task is an extreme example of Jingle Fallacy and schmeasurement.
In 100 articles we found 244 different ways of scoring it, 177 were never reused. Correlations between them range -.99 to .99.
At the same time, we show meta-analyses combine these results as if they’re equivalent.
Examples of how to use the new futurize package
Oooh love this. Easiest way to parallelise anything in #rstats. Just add futurize() to your favourite function call
www.jottr.org/2026/01/22/f...
Thanks @henrikbengtsson.bsky.social HT @rstats.blaze.email
We are responding to this call for evidence:
ec.europa.eu/info/law/bet...
Do you have concrete examples of the added value of #Rstats in the public or private sectors?
Include the most important factors (risk, lock-in, security, innovation...) to assess the added value.
Thanks!
Zotero 8 is out in stable: www.zotero.org/blog/zotero-8/
A bunch of new stuff -- the new citation dialog especially is a huge (and long overdue) improvement; massively speeds up anything I do in Word/GDocs with Zotero
Bayesian model comparison implements Occam's razor through its sensitivity to the prior. However, prior-dependence makes it important to assess the influence of plausible alternative priors. Such prior sensitivity analyses for the Bayesian evidence are expensive, either requiring repeated, costly model re-fits or specialised sampling schemes. By exploiting the learned harmonic mean estimator (LHME) for evidence calculation we decouple sampling and evidence calculation, allowing resampled posterior draws to be used directly to calculate the evidence without further likelihood evaluations. This provides an alternative approach to prior sensitivity analysis for Bayesian model comparison that dramatically alleviates the computational cost and is agnostic to the method used to generate posterior samples. We validate our method on toy problems and a cosmological case study, reproducing estimates obtained by full Markov chain Monte Carlo (MCMC) sampling and nested sampling re-fits. For the cosmological example considered our approach achieves up to $6000\times$ lower computational cost.
arXiv📈🤖
Efficient prior sensitivity analysis for Bayesian model comparison
By Hu, McEwen
This was not a critique directed at you. Just noting that not all AI-usecases are problematic.
On non-natives: I agree unintelligible reviews are rare, but it takes non-natives more time. AI can help with efficiency—and equity.
„clearly used AI“ is based on the style of writing?
Depends on how AI was used. Uploaded paper and the AI wrote the review: Clearly problematic.
Used AI to turn colloquial notes/writing into appropriate prose (e.g., correcting tone, spelling, and grammar): No problem? Think non-natives.
