Inspiring talk from Dariia Mykhailyshyna about teaching and surveying the R community in Ukraine while in the midst of a war. #rstats #useR2026

#canihazpdf Löchner, J., Santangelo, P.S., Ansell, E.B. et al. Ambulatory assessment in mental health: expert consensus and recommendations. Nat. Mental Health 4, 926–935 (2026). doi.org/10.1038/s442...

The standard random-effects model in meta-analysis assumes that the amount of heterogeneity is homoscedastic. This assumption may not be true. I derived and examined various tests for heteroscedastic heterogeneity. See osf.io/cavhy for a preprint describing these methods. #MetaAnalysis #RStats

As another alternative to the standard forest plot for visualizing many effect size estimates in a meta-analysis, one can use 'orchard plots' (Nakagawa et al., 2021). Example code to create such figures can be found here: www.metafor-project.org/doku.php/plo... #MetaAnalysis #RStats

In large meta-analyses, standard forest plot can become excessively large. One solution is to group the estimates together within studies. I just added an example to the metafor website illustrating this possibility: www.metafor-project.org/doku.php/plo... #MetaAnalysis #RStats

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...

Had a fantastic time last week in Trondheim at the @eshackathon.bsky.social 🇳🇴 A full week discussing evidence synthesis & how to improve it with a brilliant and knowledgeable interdisciplinary team. Highly energising! 🙌 Thank you to @drmattg.bsky.social for hosting us & taking such good care of us

I recently reviewed a paper for a top journals in statistics that was written front to back with a LLM. To make things worse, the conclusions drawn from the simulation study did not match what the results actually showed and instead just conformed to what the authors intended to show.

I asked ChatGPT to imagine what my ideal retirement will look like. This is what it came up with.

R-folks, is the RTMB package the go-to autodiff package in R now? Limitations? Fast? Alternatives? #RStats #r4ds

Speaking of crazy d values, this reminds me of this feature I was thinking about adding to the forest() function in metafor. #Rstats #MetaAnalysis

Ian Hussey (@ianhussey.mmmdata.io) contacted me about the recurring problem of unusually large SMD values appearing in published meta-analyses. Very large SMDs can result from using SEs in place of SDs in their calculation, which is a mistake I have also seen in practice. 1/4

A new version (1.6-0) of the metadat package (wviechtb.github.io/metadat/) has been released on CRAN (cran.r-project.org/package=meta...). The package provides over 100 meta-analytic datasets, which are useful for teaching and testing purposes. #RStats #MetaAnalysis

A new version (5.0-1) of the metafor package has been released on CRAN. It includes some smaller updates, including more ways to visualize prediction intervals / distributions. Further details here: www.metafor-project.org/doku.php/new... #Rstats #MetaAnalysis

The best of both worlds: `<=` <- .Primitive("<-") x <= c(2,5,6) mean(x) #Rstats

Why has nobody ever told me that a snake fight is part of a PhD defense? www.mcsweeneys.net/articles/faq...

@lajeunesselab.bsky.social has written the paper that meta-analysts were waiting for, and I can't recommend it more: Converting and Constructing Effect Sizes With the Response Ratio doi.org/10.1111/ele.... Go check it out (and use it!)

Two new papers out! 🎉 We revisited the idea that oxytocin promotes prosocial behaviour by collecting new data (n=211) and pooling it with data from another recent large-scale replication attempt (n=321).

I rarely use Windows, but noticed that #Rstats plots under Windows (using windows()) looked very fuzzy even with a 4k resolution. This seems to be related to DPI scaling (the system was using 200%). If I override the DPI settings for Rgui, then things are nice and crisp. Is this common knowledge?

Owning a house is how I personally verified the Second Law of Thermodynamics: If you stop adding energy, money, and time into it, everything falls apart.

Whenever I read 'battery of questionnaires', I think of ...

Given that John Watts studied psychology, you can't tell me that the band name Fischer-Z is not actually based on Fisher's r-to-z transformation (and that the 'fish's head' pun was invented later).

For people in #academia -- What's the oldest email in your inbox that you still plan to respond to at some indeterminate timepoint in your life? (in other words: Please make me feel better about my inbox of shame)

You arrive at #Oslo airport, stand at the baggage claim, wonder why you have wifi, and realize that #eduroam works in the entire airport.

The plotannotate package (#RStats) is now on CRAN: cran.r-project.org/package=plot... The package allows users to annotate base R graphics plots with freehand drawing, symbols, and text. This is useful for teaching (e.g., to visually explain certain plot elements) and creating quick sketches.

The paper about splines in R that I wish I had known about when I was learning splines in R. bmcmedresmethodol.biomedcentral.com/articles/10....

Would somebody using #Rstats under macOS be so kind and do the following: x11() plot(1) getGraphicsEvent(onKeybd=function(key) return(key)) then click on the plot, and hit control-f (and the same with control-b and control-i) and tell me what it shows in the console?

Happy to finally share the preprint for our work on how affective inertia differs within-days and overnight (with Stacey Scott, Martin Sliwinski, Josh Smyth, @wviechtb.bsky.social chtb.bsky.social‬ and @aneubauer.bsky.social) 1/4

Interested in research synthesis, meta-analysis, systematic reviews, methodology, software, guidelines, best practices? 😀 📊 This #YouTube channel is really the best spot to learn and meet our community! 👇

Planning to do a systematic review? Do watch this very informative presentation first! Julian Higgins talks about improper use of: PRISMA reporting guideline I-squared statistics Test for asymmetry in funnel plots Random effects model #ESMARConf2025 m.youtube.com/watch?v=wRA0...

I love auto translations. "The Tasmanian-led approach to random-effects meta-analysis". 😕 Didn't know that there was a Tasmanian school for meta-analysis! Learning something new every day. #ESMARConf2025

Super interesting talk by James Pustejovsky (@jepusto.bsky.social) and Martyna Citkowicz at #ESMARConf2025 on the metaselection package, which provides cluster-robust inference methods and clustered bootstrapping to account for dependencies in selection models: youtu.be/2qGcVeEqtBY

#ESMARConf2025 is live (www.youtube.com/@esmarconf), starting with a talk by @asanchez-tojar.bsky.social on "Practising what we preach: Rethinking standards in meta-analysis for ecology and evolution". At 11am CEST, I will give a talk on how to visualize the amount of heterogeneity in forest plots.

On Thursday evening during my weekly live stream on R and statistics, I will do something meta-analysis related in connection with #ESMARConf2025. Further details about my live streams can be found here: www.wvbauer.com/doku.php/liv...

Just a reminder that #ESMARConf2025 is starting tomorrow! This is a free online conference about evidence synthesis and meta-analysis using #Rstats. Further details can be found here: esmarconf.org/2025/

Tuberculosis still kills around 1.3 million people every year. It was once a massive killer in Europe and the US, but not anymore. How did these countries manage to tackle it? Our first article on TB in a three-part series: ourworldindata.org/tuberculosis...

Every scientific institute should follow this path. There is no moral stance for spending so much public money on bad software that sacrifices your sovereignty. Word/Excel are not the worst of it. e.g. The Max Planck Society spends millions of Euros annually for shitty MS calendar/email.

Hadn't heard about Sci-Net (sci-net.xyz) as a platform for requesting / sharing papers (essentially a #CanIHazPDF website). Not sure how I feel about the use of cryptocurrency tokens for this purpose (more on this at: sci-hub.se/sci-net).

Another case of the #RStats package ecosystem blowing my mind. I needed to find the nearest ellipse given a set of (x,y) coordinates. Of course there is an R package for that: cran.r-project.org/package=coni...

10/10; super cool resource!

I will teach my advanced #MetaAnalysis course this year (September 24-26, 2025): www.wvbauer.com/doku.php/cou... Since my regular meta-analysis course is also overdue, I decided to extend the advanced course by an extra day to cover the standard methods more thoroughly.

I will do an extra 'Open Online R Stream' (www.wvbauer.com/doku.php/liv...) next week Thursday (2025-05-29) at the usual time (6pm CEST). I will talk about syntax style guides, syntax formatting, and take a look at the styler and lintr packages and the new Air formatter software from Posit.

Here is something for 'base R graphics' aficionados: github.com/wviechtb/plo... The package allows you to annotate plots by freely drawing on the plot device with the mouse (works for 'windows' and 'x11' plotting devices). Various symbols and text can also be added. #RStats

Please join us next week Friday for our final seminar of the 2025 season! We are thrilled to be joined by Dr. Richard Riley (@richarddriley.bsky.social), a leading expert in individual participant data meta-analysis, to discuss the rationale, concepts, and pitfalls in conducting IPD syntheses.

Finished up chapter 6 from 'Statistical Rethinking' yesterday during my online stream. I really liked the chapter - using DAGs and simulated data, it covers some of the ways regression models can lead you astray in making (causal) inferences (via post-treatment bias, colliders, confounding).

Just read Correll et al. (2020; doi.org/10.1016/j.ti...), who recommend avoiding the use of Cohen's benchmarks (i.e., small, medium, large effects) for power analyses. Generally agree and the discussion on the inconsistencies in the definitions was quite interesting.

📢 The SAA program is live at www.saa2025.com/Program and it’s almost too good to be true — but it is true! So many incredible symposia, paper sessions, and posters to choose from that you’re guaranteed some FOMO! 🤯

A new review on different types of repeatability (ICC) and how you may get it ecoevorxiv.org/repository/v... With my good #squid friends (Holger and my 10th variance-related paper).