Maarten Marsman

Excited to lecture a full day on the Bayesian analysis of networks for the first time at the PsyNets Amsterdam workshop!

JASP Now Accepts Tax-Deductable Contributions from the US - JASP - Free and User-Friendly Statistical Software For larger donations (> $500), JASP now accepts tax-deductable contributions. To donate from the US, simple visit this UvA website. As you can read there: U.S.-based donors can make tax-deductible con...

How to donate to JASP (tax-deductable from the US)

jasp-stats.org/2025/12/19/j...

Are you interested in developing your own module in #JASP?
Join our free hackathon on 16-17 February 2026.

More information and registration:
jasp-stats.org/2025/11/21/a...

@jaspstats.bsky.social

Amerikanen hebben straks wél toegang tot DigiD – ondanks belofte van staats­secretaris Het IT-bedrijf achter DigiD komt mogelijk in Amerikaanse handen. Volgens staatssecretaris Eddie van Marum blijft DigiD desondanks ‘gewoon’ Nederlands en krijgen de Amerikanen geen toegang tot de perso...

DigiD blijft Nederlands en dus veilig. Dat verzekerde de staatssecretaris van Binnenlandse Zaken na het nieuws over de mogelijke overname van leverancier Solvinity. Die zou geen toegang hebben tot de dienst, maar daar is niets van waar. En straks kan Trump dus mogelijk ook meekijken.

🚀 Our paper on Bayes factor tests for the differences in networks (graphical models) in two independent groups is now online at Psychometrika: doi.org/10.1017/psy....

Of course, these methods are implemented in the bgms R package, which now also allows testing more than two independent groups.

GitHub - UncertaintyInComplexSystems/bamojax: Bayesian Modelling using Jax Bayesian Modelling using Jax. Contribute to UncertaintyInComplexSystems/bamojax development by creating an account on GitHub.

Bamojax (Bayesian modelling with JAX) v0.3.10 is out! Next to increased support for distributions and bijectors, several marginal likelihood estimators are now available, such as bridge sampling and truncated harmonic mean estimation (THAMES).

Check it out on github.com/UncertaintyI...!

🚀 bgms 0.1.6.0 is now on CRAN!

New in this release:
• NUTS & HMC sampling for bgm() + bgmCompare()
• Parallel chains + reproducible runs via seed
• Markov chain diagnostics (ESS, R-hat)
• New summary(), print(), and coef() methods

🔗https://cran.r-project.org/web/packages/bgms/index.html

R code and output showing the new functionality: ``` r ## pak::pkg_install("quentingronau/bridgesampling#44") ## see: https://cran.r-project.org/web/packages/bridgesampling/vignettes/bridgesampling_example_stan.html library(bridgesampling) ### generate data ### set.seed(12345) mu <- 0 tau2 <- 0.5 sigma2 <- 1 n <- 20 theta <- rnorm(n, mu, sqrt(tau2)) y <- rnorm(n, theta, sqrt(sigma2)) ### set prior parameters ### mu0 <- 0 tau20 <- 1 alpha <- 1 beta <- 1 stancodeH0 <- 'data { int<lower=1> n; // number of observations vector[n] y; // observations real<lower=0> alpha; real<lower=0> beta; real<lower=0> sigma2; } parameters { real<lower=0> tau2; // group-level variance vector[n] theta; // participant effects } model { target += inv_gamma_lpdf(tau2 | alpha, beta); target += normal_lpdf(theta | 0, sqrt(tau2)); target += normal_lpdf(y | theta, sqrt(sigma2)); } ' tf <- withr::local_tempfile(fileext = ".stan") writeLines(stancodeH0, tf) mod <- cmdstanr::cmdstan_model(tf, quiet = TRUE, force_recompile = TRUE) fitH0 <- mod$sample( data = list(y = y, n = n, alpha = alpha, beta = beta, sigma2 = sigma2), seed = 202, chains = 4, parallel_chains = 4, iter_warmup = 1000, iter_sampling = 50000, refresh = 0 ) #> Running MCMC with 4 parallel chains... #> #> Chain 3 finished in 0.8 seconds. #> Chain 2 finished in 0.8 seconds. #> Chain 4 finished in 0.8 seconds. #> Chain 1 finished in 1.1 seconds. #> #> All 4 chains finished successfully. #> Mean chain execution time: 0.9 seconds. #> Total execution time: 1.2 seconds. H0.bridge <- bridge_sampler(fitH0, silent = TRUE) print(H0.bridge) #> Bridge sampling estimate of the log marginal likelihood: -37.73301 #> Estimate obtained in 8 iteration(s) via method "normal". #### Expected output: ## Bridge sampling estimate of the log marginal likelihood: -37.53183 ## Estimate obtained in 5 iteration(s) via method "normal". ```

Exciting #rstats news for Bayesian model comparison: bridgesampling is finally ready to support cmdstanr, see screenshot. Help us by installing the development version of bridgesampling and letting us know if it works for your model(s): pak::pkg_install("quentingronau/bridgesampling#44")

De redactie verzweeg het feit dat de voorstellen van GL-Pvda volledig onderschreven wordt door het WRR-rapport Goede Zaken. Geen manipulatietechniek wordt geschuwd om de oppositie kapot te maken.

Fair coins tend to land on the same side they started: evidence from 350,757 flips.

That's the title of our paper summarizing ~650 hours of coin-tossing experimentation just published in the Journal of the American Statistical Association.
doi.org/10.1080/0162...

New paper with @richarddmorey.bsky.social now out in JASA, where we critically examine p-curve. Below is Richard’s excellent summary of the many poor statistical properties of p-curve (with link to paper). I wanted to add some conceptual issues that we also tackle in the paper.

This week, I had the pleasure of teaching the Bayesian approach to network analysis at the Network Psychometrics summer school at Lake Como School of Advanced Studies.

Organized by Giulio Constantini, Michela Zambelli, and Semira Tagliabue with @briganti.bsky.social and @anastasiapsy.bsky.social…

Ben jij gedreven om psychische klachten te voorkomen voordat ze beginnen of terugkeren? Dan is dit promotietraject iets voor jou!

Shift Left @Arkin @amsterdamumc.bsky.social

lnkd.in/di4vWujy
@uvapsychology.bsky.social

"So EJ, say I am tossing a coin..."

Hello world! We are a new company that provides support for organizations and industries who use the JASP open-source stats program. Check out our website www.jasp-services.com and our first blog post www.jasp-services.com/first-post/
Know companies using commercial stats software? Share this info!🙂

🚨 New preprint: A Stochastic Block Prior for Clustering in Graphical Models

We introduce an SBM prior to detect/test clusters in Bayesian network models for binary & ordinal data. Includes R code & tutorial.

📄 osf.io/preprints/ps...
📚 Blog: www.nikolasekulovski.com/blog/post2/

I am looking forward to expanding the scope of my professorship by combining cognitive and statistical modeling with LLMs😊

There will be two job openings for postdoc positions soon - one starting in September 2025 and another one a year later.

Home - Network Psychometrics: theory, methods and applications Network Psychometrics: Theory, Methods and Applications 4-8 August 2025   Join us to explore the forefront of psychological research! Network Psychometrics has rapidly gained prominence across various...

I am excited to teach at the Summer School on Network Psychometrics this August in Como, Italy! Theory, methods, and hands-on R for analyzing cross-sectional and longitudinal data. Full information and application (deadline April 24th): ntps.lakecomoschool.org

An introduction to Sequential Monte Carlo for Bayesian inference and model comparison—with examples for psychology and behavioral science - Behavior Research Methods Bayesian inference is becoming an increasingly popular framework for statistics in the behavioral sciences. However, its application is hampered by its computational intractability – almost all Bayesi...

My tutorial on Sequentual Monte Carlo for psychology and behavioural science is out now in Behavior Research Methods! Check it out at link.springer.com/article/10.3...

I could collect more data, and the Bayesian approach allows me to monitor the evidence (e.g., the BF) as the data come in. I could also include theory or results from earlier research and update my knowledge. This is the kind of cumulative science I like to see! :-)

The phrase about results "not providing a solid basis for cumulative science" is about edges with anecdotal evidence. If I do not have enough evidence to draw a conclusion about an edge, then any decision is "risky". I would be happy to learn this if my theory or intervention builds on the edge.

Interestingly, we also found evidence for the absence of many edges, a result that doesn't match the predictions of the unidimensional factor model, suggesting it wouldn't fit the data well. Thus, we likely need network or equivalently higher-order factor models to describe these data!

Hi Miri, I don’t think it’s about partial correlations but about model complexity, as Karoline said. Unidimensional factor models are also based on partial correlations and are often robust, but they also have far less parameters than network models.

Are psychometric networks sufficiently supported by data such that one can be confident when interpreting its results? We analysed 294 psychometric networks from 126 papers with the Bayesian approach to address this question @jmbh.bsky.social Sara Ruth van Holst @maartenmarsman.bsky.social 🧵

Emotions are reactions to situations we encounter in daily life. In our new paper in Psych Review (psycnet.apa.org/fulltext/202...; with @oisinryan.bsky.social and @fdabl.bsky.social), we take a first step towards building a generative model for emotion dynamics based on this simple principle 1/4

Je was me net voor!

This is a collaborative effort with amazing colleagues: Lourens Waldorp, Nikola Sekulovski, and @jmbh.bsky.social. 🙌 Thanks to this team for their hard work in advancing Bayesian methods in network analysis!

bgms: Bayesian Analysis of Networks of Binary and/or Ordinal Variables Bayesian variable selection methods for analyzing the structure of a Markov Random Field model for a network of binary and/or ordinal variables. Details of the implemented methods can be found in: Mar...

Ready to try it out? 🚀 Our method is implemented in R package bgms. Install it now and start analyzing your networks with Bayesian advantages! 🔗

cran.r-project.org/package=bgms

Why is this exciting? 🌟 Our Bayes factor test helps you distinguish the absence of evidence from the evidence of absence of a difference effect. 📊 This means you can actually quantify the support for the null hypothesis of parameter equivalence! 🎯