Daniel Lakens

I also created a version of the 'guess the correlation' game. Students always enjoyed it, but the original website (with fun retro design) started to require a google account login. So I created my own version. Play it here: https://lakens.github.io/statistical_inferences/06-effectsize.html#correlations-and-variance-explained

I grind my coffee beans by hand, listen to music on vinyl, do stats on my fingers, read the literature on paper, I have my meetings in person in a cafe smoking a pipe, I teach with the blackboard, and write down ideas in Waterman Serinity Blue.

I produce 100% artisanal science.

We are hiring an assistant professor in Metascience! Join our lovely department of Human-Technology Interaction, with a very active Metascience group, at Eindhoven University of Technology. Apply here before August 14, or share the vacancy in your network!

https://www.tue.nl/en/working-at-tue/vacancy-overview/assistant-professor-in-metascience

I'm working on a new Metacheck module: Datacheck. Repocheck already lists all files on Github, Zenodo, OSF, Researchbox. Datacheck will download them and.... which automated checks would you like for datafiles? Can be general (e.g., flag outliers), or file specific (e.g., excel files). Let me know 👇

Yesterday we recorded a Nullius in Verba podcast episode in which we critically reflect on multiverse analysis, it's misuse, and it's very limited uses in science. I bring up some points of our recent preprint 'There is only one correct analysis' (https://osf.io/preprints/psyarxiv/hvjxf_v1). Out this Friday!

New blog post: Evaluating Dr. Cuddy’s Claim that the Debunking of Power Posing is a Myth.

https://daniellakens.blogspot.com/2026/05/evaluating-dr-cuddys-claim-that.html

On an AI generated description of a non-existent study, incorrectly citing findings from studies, and the importance of scientific criticism.

New preprint: There is only one correct analysis. https://osf.io/preprints/psyarxiv/hvjxf_v1

We take a critical look at recent proposals for multiverse or many analyst procedures, and strongly argue that epistemic uncertainty should be reduced, not “embraced”.

Multimodel inference is largely descriptive: It can't be used to evaluate theories. We discuss '12 Angry Men' to illustrate why we don't do science by a majority vote. Robustness is not a goal of science - it is a consequence of severe tests.

>

If a scientist uses AI to make strong claims in public about the scientific literature, and some references do not exist, and the generated text does not correcfly reflect the findings, should we inform them privately to fix the text or should we share this information publicly?

A year ago Sajedeh Rasti and I reached out to 20 Dutch metascientists and asked: Instead of competing for this upcoming grant, how about we collaborate, determine priorities, and combine our expertise? Both grants were funded. We discuss our experience in a new preprint.

The preprint is available on Psyarxiv: https://osf.io/preprints/psyarxiv/fmnc5_v2

We had not heard of people trying collaborative grant writing like this before. If you have experiences with this format you would like to share, let us know!

The first time a serious journal asks me to review a paper that is completely written by AI.

How do we deal with this? Is it fair to write 'this is either written by AI, or mirrors it's lack of original content'?

How do I deal with the 1% probability a human wrote it anyway?

It probably took way too long, but I have becomes such a fan of Obsidian over the last weeks! The open .md format makes it so versatile! Combined with python scripts, you can automate workflows, build dashboards, track to-do's, and keep all info organized. https://obsidian.md

Statisticians are trained in how to compute statistics, but not why to compute statistics. At least 10% of their training should be in philosophy of statistics. They should also spend 10 credits on their own empirical research project to understand how data is acually collected.

Almost in Paris for a talk on 'Which Research is Worth Doing Well' at the University of Paris Nanterre. It is so convenient to be able to walk to the train station, get into a train, and end up 2 countries away - all while working on the revisions of a paper.

198 effect sizes in ego depletion resesrch showed an effect size of d=0.62. Preregistered large replications (including some by original authors) yielded an effect size of 0. No one has been able to offer any other explanation for this huge research waste than massive p-hacking.

A new set of papers, sharing the long-awaited result of several reproducibility and replicability projects, including commentaries, is published today. I look forward to reading the studies, and re-using the data generated! https://www.nature.com/collections/idajfifcfg

FAIR data and code is both the most boring part of Open Science, and the part that will most fundamentally change how science is done.

This gap is regrettable, but understandable. Most scientists can't imagine what their science would be like if research outputs were FAIR.

Some people think that 'the null is never true' is a killer argument against p values. It isn't. If you believe this, test against a meaningful null - e.g., r > 0.1, instead of r > 0. Moving to range predictions is a big improvement - also if you like p values. https://doi.org/10.1177/1745691620958012

New Nullius In Verba episode on miscitations:
https://nulliusinverba.podbean.com/e/ep-77-miscitatio/
How often are citations to the scientific literature outright misleading? Do we really need to spell out that people are supposed to read what they cite? And should we review citations as they do in law?

When we started to promote replication studies (e.g. Koole & Lakens, 2012) we had a scientific culture where publishing a direct replication with a null result was practically impossible, and most didn't even consider trying. That culture has changed - although still not enough.

I remember the widespread pushback when the Peer Reviewer Openness Initiative started in 2017, where we would only review papers that share data and code (or explain why that is not possible) https://www.opennessinitiative.org/the-initiative/

Now, funders and journals are requiring it.

It is a joy to see people use and appreciate our paper on exploratory research, how it is a crucial step in the research process, how to be transparant about it, and how to do it well. It has 'sport science' in the title, but is relevant for many fields! https://www.tandfonline.com/doi/10.1080/02640414.2025.2486871?url_ver=Z39.88-2003&rfr_id=ori:rid:crossref.org&rfr_dat=cr_pub%20%200pubmed

We are pleased to announce that the Paul Meehl Graduate School is launching the PMGS Dissertation Award, recognizing outstanding PhD dissertations that advance meta-research.

See the eligibility criteria and apply before June 1st, 2026 at:
https://paulmeehlschool.github.io/award/

My PhD Sajedeh Rasti has published a World View piece in Nature today: "Don’t compete, collaborate: why collective funding applications are the future" https://lnkd.in/eC28rwpw She discusses how she successfully brought together a large team of meta-scientists to collaboratively apply for funding. I'm super proud of this amazing young scholar!

If you prefer to read open access texts, then read our more extensive overview (submitted to Meta-Psychology) here: https://osf.io/preprints/psyarxiv/fmnc5_v2/

My Talk 'What's the benefit of preregistration for big team science' is online, and you can watch it here: https://www.youtube.com/watch?v=31lZG6iN0gQ

Hope to see you at the Conference on Transparency, Technology and AI in Peer Review on June 5, 2026, which is organized in collaboration with the Center for Humans and Technology at Eindhoven University of Technology.

Register at: https://meta-eindhoven.github.io/events/

The first of a 2 part episode on 'Disagreement' in science of our podcast Nullius in Verba is out: https://www.podbean.com/eas/pb-3vw2d-1a8de28 . Why is disagreement and criticism so important in science? Do we need to interact with all our critics? And is criticism, as Popper wrote, how 'our mind grows and transcends itself'?

Interesting interview with Anthony Edwards on his experiences working with Ronald Fisher, diving into the topic of Eugenics, with a much more nuanced take than you'll nowadays find on social media. Worth watching: https://www.youtube.com/watch?v=kqLB5a8op5c

It is very funny to see that authors who in 2019 argued that "Preregistration is redundant, at best", are now co-authoring papers where their collaborators want to preregister all studies. I guess they weren't even able to even convince their direct co-authors!

Second keynote of the day, Lisa Spitzer talking about how to foster preregistration, with a lot of insights into how it was to be a metascience PhD researcher!

If you are new here, remember you can get updates about new preprints from @PsyArXivBot@botsin.space , @EdArXivBot@botsin.space , @MetaArXivBot@botsin.space and @SocArXivBot@botsin.space !

Went bouldering with Noah van Dongen and @cjvanlissa@mastodon.social . Noah got a Veni grant this year, Caspar a Vidi, and I received a Vici, so we had to make this picture 😊 We will also work together in our projects on improving theorising in psychology!

Re-reading Hasok Chang's call for pluralism in Science in his 2012 'Is water H20?'. It is so badly argued for, his discussion is so one-sided, and his rejection to engage with criticisms is so annoying, I find it hard to complete. Will discuss it in an upcoming podcast episode.

New Paul Meehl Graduate School workshop announced: Computational Cognitive Models

https://paulmeehlschool.github.io/2026-03-10-cognitive-model/

Please note that the workshop will be exclusively in-person on April 20, 2026. 

New (free) workshop by the Paul Meehl Graduate School on agent-based modeling: https://paulmeehlschool.github.io/2026-04-24-agent-based-models/ The workshop will be exclusively in-person on May 27, 2026, in Eindhoven. Register before May 18th!

With Deborah Mayo, Kent Staley, and Wendy Parker we are editing a special issue of Synthese on "Severity and Learning from Error": https://link.springer.com/collections/ebjdhfadcd Deadline for submissions is the 15th of December, 2026. If this is a topic that you are interested in, consider contributing!

"20% of the studies presented causal claims that were not supported by their designs, opening the risk of their results being misinterpreted and their effect sizes being exaggerated [..] exacerbated by a lack of method transparency and construct validity" https://journals.sagepub.com/doi/abs/10.1177/01492063261424849