František Bartoš

Another meta-analyses on AI in education. > Fig. 2 provides a visual representation of the distribution of effect sizes, indicating that the effect sizes are nearly symmetrically distributed... This distribution suggests that there is no evidence of publication bias present...

Xu, T., & Wang, H. (2024). The effectiveness of artificial intelligence on English language learning achievement. System, 125, 103428. doi.org/10.1016/j.sy...

Another meta-analyses on AI in education. > Fig. 2 provides a visual representation of the distribution of effect sizes, indicating that the effect sizes are nearly symmetrically distributed... This distribution suggests that there is no evidence of publication bias present...

Huang, X., Xu, W., Li, F., & Yu, Z. (2024). A meta-analysis of effects of automated writing evaluation on anxiety, motivation, and second language writing skills. The Asia-Pacific Education Researcher, 33(4), 957-976. link.springer.com/article/10.1...

Come to our #IMPS2026 symposium on methodology of simulation studies with Richard Feinberg, @karolinehuth.bsky.social , Mirka Henninger, and David Kronthaler on Tuesdat at 12:45 to room 102. The symposium has the following talks:

@jaspstats.bsky.social released a new of the Meta-Analysis module -- if you have the latest version of JASP you should see a small blue download icon next to the Meta-Analysis module on the ribbon. You can update it via the Module Store (see screenshot below).

Currenly revising our critique of meta-analyses of AI/LLM on learning. Some of the original meta-analyses are beyond ridicolous. This one has 87 citations since published in 2025. (doi.org/10.1177/0266...)

Out of curiosity I checked a the first author's google scholar. He seemed to have co-authored additional three meta-analyses last year. Two of those show quite unbelievable effect sizes again. (I guess the d = 2.5 is fine when the rest contains d around 4)

The multilevel extension to RoBMA was just published in BRM (link.springer.com/article/10.3...). Together with @maxmaier.bsky.social and EJ Wagenmakers, we developed a 3-level version of RoBMA which allows analysts to adjust for publication bias in meta-analyses with estimates nested within studies.

Interestingly, we find that in a presence of a correlation between cluster size and effect size (as we observed in our second example), most existing methods become severely biased; multilevel RoBMA was the only method to consistently recover the true effect across all simulated conditions.

The multilevel extension to RoBMA was just published in BRM (link.springer.com/article/10.3...). Together with @maxmaier.bsky.social and EJ Wagenmakers, we developed a 3-level version of RoBMA which allows analysts to adjust for publication bias in meta-analyses with estimates nested within studies.

Accounting for the multilevel structure in publication bias adjustment is necessary for drawing valid inferences since majority of meta-analyses include multiple effect sizes per study. Ignoring this dependency results in overly narrow intervals and inflated evidence.

The multilevel extension to RoBMA was just published in BRM (link.springer.com/article/10.3...). Together with @maxmaier.bsky.social and EJ Wagenmakers, we developed a 3-level version of RoBMA which allows analysts to adjust for publication bias in meta-analyses with estimates nested within studies.

I just found out that a meta-analysis on the effect of ChatGPT on learning that we criticized last year not adjusting for publication bias and overinterpreting the conclussions (osf.io/preprints/ps...) was retracted last month (see notice www.nature.com/articles/s41...)!

A major release of the RoBMA package (4.0.0) is now on CRAN. Main highlights: - fully redesigned interface closely matching metafor - dedicated brma, bselmodel, bPET, bPEESE for single model fits - fit diagnostics features (residuals, influence, etc) (overview at fbartos.github.io/RoBMA/articl...)

These update unfortunatelly break backwards compatibility and the syntax no longer matches the previously published tutorials. Fully updated vignettes showcasing how to match the previous versions of the package are on the github pages at: fbartos.github.io/RoBMA/index....

A major release of the RoBMA package (4.0.0) is now on CRAN. Main highlights: - fully redesigned interface closely matching metafor - dedicated brma, bselmodel, bPET, bPEESE for single model fits - fit diagnostics features (residuals, influence, etc) (overview at fbartos.github.io/RoBMA/articl...)

All fitting functions (RoBMA, BMA, BMA.glmm, brma, brma.glmm, bselmodel, bPET, bPEESE) recieve following arguments: - `mods` specify moderation (meta-regression) - `scale` specify scale-models - `cluster` allows effects nested within studies (`random` like in rma.mv arrives later)

A major release of the RoBMA package (4.0.0) is now on CRAN. Main highlights: - fully redesigned interface closely matching metafor - dedicated brma, bselmodel, bPET, bPEESE for single model fits - fit diagnostics features (residuals, influence, etc) (overview at fbartos.github.io/RoBMA/articl...)

In our new preprint we show how to efficiently compute Bayes factor sensitivity analyses for any model comparison. The computation trick is simple -- it rellies on extend the model by a hyper-prior on the parameter(s) for which sensitivity is desired and using transitivity of Bayes factors.

What makes it even more efficient is combining this trick with inverse weighted marginal density estimator (IWMDE). In out settings IWMDE becomes trivial to implement and improves the density estimation tremendously so only a couple thousands posterior samples are needed.