Lukas Wallrich

There is a reason for blinded peer review - not surprising that LLMs (like humans) take into account all indicators unless told otherwise or even better, prevented from doing so. There are some more interesting nuances in the paper; i hope we can move beyond discussing the trivial headlines soon.

PsyArXiv is now down to only ~40% of the backlog we started with, thanks to our amazing moderators 🎉 To all mods: thank you so much for your hard work!

Did one of your preprints get approved recently? Help us show our thanks by liking and sharing this post :)

#PsyArXiv #PsychSciSky

In the same way that we could not strictly say that x% of a student's grade is attributable to how hard they study (because for each individual factors also matter)? Just trying to figure out whether that's an (appropriate) critique of how we talk about stat models generally, or genetics specific?)

For someone really unfamiliar with the topic - what's the difference?

Very excited that hashtag#UKRI is allowing us to bring coders and meta-scientists together in Germany to build tools that make replications more visible. If you are up for three days of coding, collaboration and new connections, have a look at the details below ... and please share.

OK here it is, the moment you have all been waiting for! Our advert for 5 posts here at Glasgow: Cognitive neuroscience/psychology. Closing date May 12 - please spread the word
www.nature.com/naturecareer...

🙏 Many thanks ...
... to my wonderful co-authors: Victoria Opara, Ditte Barnoth, Miki Weselowska, and Sayeh Yousefi, the many authors of original studies who responded to our questions and many others who helped along the way! [8/8]
#Diversity #TeamPerformance #Leadership #MetaAnalysis #OpenScience

📖 Read the Full Study

Dive deeper into our findings in the Journal of Business and Psychology (rdcu.be/dSpUM) or check out the Open Access preprint (osf.io/nscd4). Explore the dataset in our interactive web app (lukaswallrich.shinyapps.io/diversity_me...) [7/8]

📚 So what? Improving Diversity Measurement & Research

Current diversity measures are often superficial or inconsistent with theory. Most research overlooks non-linear relationships and lacks context. We need more nuanced research to understand when and how diversity can deliver its promise. [6/8]

💡 So What? Rethinking the Business Case

Diversity alone doesn’t universally increase performance - so simple claims can cause disappointment. Instead, diversity needs culture shifts, and other reasons to diversify should be valued (e.g., equity, stakeholder demands, access to top talent). [5/8]

🌍 Culture Matters

Diversity shows more positive associations in countries with lower power distance and greater individualism—suggesting that inclusive leadership may improve diversity outcomes. [4/8]

🧠 Good for Innovation & Creativity
The association between team diversity and performance is stronger in complex, creative tasks, especially in R&D settings where diverse perspectives can fuel innovation. [3/8]

🔍 Weak Average Relationships

While some individual studies show strong associations, the average effects are small (and below our specific Smallest Effect Size of Interest). Overall, diversity explains less than 1% of performance variance. [2/8]

🚀 New Meta-Analysis Published: Linking Team Diversity with Performance 🚀

We analyzed 615 studies (2,638 effect sizes) to assess how different dimensions of team diversity are associated with team performance—now published as a Registered Report in the Journal of Business and Psychology. 🧵 [1/8]

🚨WP🚨
We test 9 online samples and find clear tradeoffs between attentiveness and representativeness - which sample is best depends on research q and priorities. For social/political qs I rec Bovitz/Lucid, for complex designs I rec Cloud/Prolific osf.io/preprints/ps...

New preprint:
"On the process & value of direct close replications: Reply to Shafir and Cheek (2024) commentary on Chandrashekar et al. (2021)"

Rebuttal of Shafir & Cheek's (2024) commentary of our replication now 40+ pages.

Feedback welcome!

Read:
www.researchgate.net/publication/...

👇🧵

We are thrilled to announce the Paul Meehl Graduate School for metascience at Eindhoven University of Technology. We will offer 6-8 free workshops a year for metascience PhD students. You can register and find more information at paulmeehlschool.github.io

This may well be the first: if you saw a paper in @natureportfolio.bsky.social Nat Hum Behav that you think would benefit from an independent replication, would like to do one & maybe have it published in the actual original journal, here is your chance #AcademicSky
www.nature.com/articles/s41...

So happy receiving your requests for NHB articles. We already have 30 ongoing robustness reproductions/replications with new data for NHB (2023-onwards).

Here is a list of studies available: osf.io/6mr3p/

Contact us at instituteforreplication@gmail.com if you want to participate!

New episode of @nulliusinverba.bsky.social on preregistration. I love recording with Smriti. She brings up an intriguing argument for preregistration that is different from my viewpoint, but very valid. We talked so long, part 2 will follow. nulliusinverba.podbean.com/e/preregistr...

Need a few more days to get your #SIPS2024 submission ready? No problem! We're extending submissions to February 7. #OpenScience #Metascience #Psychology #PsychSciSky #Stats #AcademicSky

Indeed! Just came across the R codebook package yesterday, which can provide a strong starting point (and be super-helpful when you need to make sense of un-FAIR data)

It's difficult for others to use your data without first reading documentation. Sharing a data dictionary or codebook alongside your data ensures that users correctly interpret variables. A few examples and templates can be found here: osf.io/e5g6t/

Just curious: how up to date is that 96%? I believe I have seen it for a while and wonder if we might at least be moving in the right direction?

⚠️⚠️New Blog Post!⚠️⚠️
Centering in Moderation Analysis: A Guide

I talk about: how to center variables (including categorical variables!), when you need to use it (moderation analysis, ANOVA) and how to do it (with easystats's {datawizard})!

#rstats

Full program online for our Workshop Perspectives on Scientific Error | 29 Feb, 1 and 2 March 2024! Program: docs.google.com/document/d/1... Keynotes by @simine.com @ianhussey.bsky.social @cdutilhnovaes.bsky.social & Judith Ter Schure
Registration (10 seats left): www.eventbrite.nl/e/workshop-p...

Interesting postdoc opportunity

"Lab² aims to be one of the most influential hubs for replications, robustness studies, and metascience in the field of economics."

wzb.hr4you.org/job/view/282...

AMPPS Tutorial Reviewer Form This form is designed to collect some brief information from potential "tutorial reviewers" for Advances in Methods and Practices in Psychological Science (AMPPS). AMPPS is committed to having our tut...

At the @psychscience.bsky.social journal AMPPS, we publish tutorials and have dedicated tutorial reviewers. TRs are go through all the code and each step of a tutorial, then provide feedback about whether it all works as it should. Grad students and post-docs are welcome to be TRs. Join us?

Paper: "Selective and (mis)leading economics journals: Meta-research evidence"

OA: onlinelibrary.wiley.com/doi/10.1111/...

Why Is There So Much Fraud in Academia? - Freakonomics Why Is There So Much Fraud in Academia? - Freakonomics

New Freakonomics radio episode with interviews of Leif Nelson, @joesimmons.bsky.social, Max Bazerman, @simine.com, @urisohn.bsky.social, and me.

Obviously about the Gino case, but also about issues of fraud more generally.

freakonomics.com/podcast/why-...