Caspar van Lissa

Submit to the Theory Methods Conference by June 18! Even if you are not an expert in theory methods, submit to get some great feedback and connect with kindred spirits. Submit your proposal: https://edu.nl/mj9x6
For an overview of the conference program and keynotes: https://theorymethodssociety.org/conference.html

Shape the future of social science as a PhD student of machine learning-informed theory construction, based on patterns in high-frequency longitudinal data! You will be part of vibrant communities like the INSIGHT Lab, Theory Methods Society, and Tilburg Experience Sampling Center

https://tiu.nu/23721

EU scholars should be digitally autonomous, free from unreliable US big tech. In this article I discuss what YOU can do, which lessons the COVID-19 crisis teaches us, and how we can build on the most recent successful culture change in academia (open science) https://www.scienceguide.nl/2026/03/digital-autonomy-needs-you/

The "Theory Paper Format" is a template for writing psychological theory papers, developed by the Taskforce Theoretical Psychology (Lorentz Workshop, 2025). Use it to structure your theoretical paper, or (for editors) consider creating journal space for theory papers https://doi.org/10.17605/OSF.IO/D3JMW

Does your R-script work on your computer? Want to make it fully reproducible on ANY system, now and in the future? In this video, I demonstrate how to turn R scripts into reproducible, version-controlled projects with worcs, renv, targets, and testthat to verify reproducibility. https://youtu.be/3mgRFMr5APU

Call for Submissions for the Theory Methods Conference 2026, September 30-October 2! https://theorymethodssociety.org/conference.html

We invite you to:
1) Submit your proposal: https://edu.nl/mj9x6
2) Invite your colleagues/lab/(PhD) students, and encourage them to submit
3) Share this post

Are you open science-minded, technically savvy, and interested in mixed methods? Come build the future of mixed methods with Tamarinde Haven and Marie Stadel. Our campus is green, our colleagues supportive, and our research excellent!
https://www.academictransfer.com/en/jobs/358351/phd-candidate-on-the-future-of-mixed-methods-research-junior-lecturer-in-methodology-and-statistics/

Everyone’s talking about the "theory crisis" - but do we fix it with formal- or agent-based models, DAGs, propositions? Regardless: we need a shared medium to exchange and improve ideas. That medium is FAIR theory: Findable, Accessible, Interoperable, Reusable https://doi.org/10.1177/17456916251401850

Open source research software is powerful, but only if people know how to use it. That's why, after publishing 12 R packages, I'm shifting my attention toward open educational materials. Every month, I will record a deep-dive video on advanced stats and data science in R. First up: mixed (continuous/categorical) data Latent Class Analysis with tidySEM. What should I cover next? https://www.youtube.com/watch?v=e6lUG5f7TWs

Over christmas, I added tidySEM much-requested support for mixed-data latent class analysis (continuous/categorical)! This was technically possible, but the new implementation is easier, faster, with better convergence! Try it by running remotes::install_github("cjvanlissa/tidySEM"); ?mx_mixed_lca

Fantastic projects awarded, including two I'm part of: Hannes Datta's "ReproHub", which integrates reproducibility/checks into research infrastructure, and Terrence Jorgensen's "Infra-Structural Equation Modelling", which develops lavaan and related open educational material! Congratulations to all! https://www.openscience.nl/en/news/45-projects-strengthen-dutch-open-science-infrastructure

🟥 Ik staak tegen bezuinigingen op het hoger onderwijs. Investeren in de kenniseconomie is noodzakelijk om wereldwijde uitdagingen van klimaat, ongelijkheid, gezondheid, veiligheid, ea het hoofd te bieden. We roepen de overheid op: kies een andere koers.

Meer informatie: https://www.fnv.nl/ho

Loving @IC2S2@datasci.social , but I wonder if we're asking the right questions. Is "How can I outsource this task to a proprietary, black box, resource-intensive, often un-validated LLM?" when fit-for-purpose classifiers abound a computational a computational research question? Or even scientific?

At my preconference workshop @IC2S2@datasci.social participants will conduct a *fully reproducible* and *FAIR theory based* simulation study.
The full interactive tutorial is in the theorytools R-package docs: https://cjvanlissa.github.io/theorytools/articles/computational_social_science.html
And slides: https://cjvanlissa.github.io/worcshop/ic2s2.html

To be FAIR: Theory specification needs an update! This revised preprint discusses how making theories FAIR digital objects that others can reuse and update could help navigate the "theory crisis" in psychology. Also updated the theorytools R-package, adding tutorial vignettes showing how to make theories FAIR and what to do with them (select covariates, simulate data). Examples include Self-Determination Theory, the Dunning-Kruger effect, and Morris' Tripartite Model

https://osf.io/preprints/psyarxiv/t53np_v2

I'm hiring a PHD candidate in Machine Learning-Informed Formal Theory Construction. Please encourage talented students to apply, or reach out if you want to collaborate. Looking for machine learning, theory development, and programming skills, and interdisciplinary interests! See https://tiu.nu/22752

For @TilburgU@mastodon.world 's #sciencequest podcast, I discussed the question: Do we still need scientific models? Can we not just use AI to answer all of our questions? The interview is in Dutch, accompanying text in English: https://www.tilburguniversity.edu/magazine/overview/why-scientific-models-remain-essential

Bridging the gap between #R package documentation and #openscience Open Educational Materials, I'm embedding formative quizzes in the package vignettes (based on @debruine@tech.lgbt 's webexercises)! Check the development version of https://github.com/cjvanlissa/theorytools to see how it's done.

To be FAIR: Theory needs an update! FAIR theory can be shared, reused in analysis software, and updated based on new findings. Using existing #openscience infrastructure, it streamlines collaboration, reduces research waste, and accelerates cumulative knowledge development https://osf.io/preprints/psyarxiv/t53np

Revising papers became easy when I discovered James Conigrave's "revise" package, which inserts tagged snippets from a manuscript into a revision letter! No more back-and-forth, no more consistency checking. Now on CRAN, may it save you as much time as it did me! https://cran.r-project.org/web/packages/revise/index.html

My department is hiring an Assistant Professor of Social Data Science! I would love for this job to go to someone knowledgeable about open science, reproducibility, and meta-science. Our department is fun, supportive, and has extensive expertise in structural equation modeling, Bayesian stats, regularization, text mining, meta-science, and research synthesis. The campus has a small forest, and the city can be easily reached from Utrecht, Rotterdam, even Amsterdam. https://tiu.nu/21154