To be clear, I'm not against it. It's a bit annoying to have to read so many verbose AI reports, but I still think this is preferable to the regime of before, where most scientific work was never rechecked and many errors languish in code, polluting the record.

An interesting thing is happening to me at the moment. I have a standing bug bounty policy for my first authored scientific work. That means, I've committed to pay people who report bugs in my code, up to 750€ for severe errors. Since GPT-6 Astra came out, I'm now getting several bug reports daily.

100%. There is no reason to treat a supplement differently than code and data. Some journals insist on hosting the supplement on the article website. I would still include a DOI for the supplement in the paper to be on the safe side.

Daniel Lakens@lakens.bsky.social · 3d ago

Journals should stop offering to store supplemental information. Everything that is shared with a paper should have it's own DOI, and live in a proper repository. As authors, link to files (e.g., Zenodo), but never add them as supplemental files.

Special Issue: Explanation and Causality in Sociology: link.springer.com/journal/1157...

KZfSS Kölner Zeitschrift für Soziologie und Sozialpsychologie

Kölner Zeitschrift für Soziologie und Sozialpsychologie / Cologne Journal of Sociology and Social Psychology  Founded in 1948, with a focus on theory ...

link.springer.com

Als ich mir in den 90ern Unis in #Berlin anschaute, war mir nie bewusst, dass die Humboldt-Universität bis 1933 das #Harvard der Welt war. Ich glaube, vielen Deutschen ist nicht bewusst, wie fundamental unser Uni-System 33-45 zerstört wurde – und dass es diese Spitzenstellung nie wiedererlangt hat.

🎉 COS has established an independent subsidiary organization in Brussels! While the Belgian organization is legally independent from COS in the US, it will not operate separately—it's part of one integrated, global COS organization with a shared mission, strategy, & brand

A New Chapter for COS in Europe

The Center for Open Science (COS) has officially established an independent subsidiary organization in Brussels, Belgium. This presence gives COS flexibility to engage with European institutions, organizations, funders, and communities.

cos.io

What counts as data in qualitative research? blogs.lse.ac.uk/impactofsoci... Good post around the question of "what is lost when qualitative data is treated mainly as text?" and how this relates to the use and limits of AI for some types of qualitative research. 1/

Noah is leading our effort to do a naturalistic randomized trial on a variant of the Registered Reports publishing model. Registered Revisions implements randomization at Revise-and-resubmit stage for new data or analyses. Here's a thread and preprint with the latest.

Noah Haber@whaleactually.com · 6d ago

... and that's where Registered Revisions comes in The Registered Revisions Meta Trial Pilot Report is OUT NOW!! osf.io/preprints/me... But new(ish) policy involves developing a new(ish) multi-headed trial design, workflows, infrastructure, piloting, mistakes, and a super team. And it's all here.

Nate Breznau, Hung H. V. Nguyen: Metatheoretical multiverse analysis: Improving the reliability of research with theories-as-data https://arxiv.org/abs/2609.09190 https://arxiv.org/pdf/2609.09190 https://arxiv.org/html/2609.09190

The "prediction" that "AI could kill all humans" in the next decade was actually (10, 100]% probability btw – in an X post!¹ 😆 😭 If you're into this genre of doom, but with thought put into it, you could try the National Risk Register.² ¹ x.com/EvanHub/stat... ² www.gov.uk/government/p...

Introducing REWIRE: Advancing Reproducibility Across the European Research Area www.openaire.eu/introducing-... #OpenScience A laudable project, let's hope it will have broad impact in Europe.

`webexercises` creates interactive exercises for pkgdown, Rmarkdown, and with this update, Quarto (books), Latex output, and #R package vignettes! My hope is that this will support open educational materials, especially in R-package documentation. cran.r-project.org/web/packages...

webexercises: Create Interactive Web Exercises

Functions for easily creating interactive web exercises in 'R Markdown', 'Quarto', and package vignettes that students can use in self-guided learning.

cran.r-project.org

Rogue Scholar has now made its >50K blog posts available as PDF files.

This week, @nature.com features the World Bank's reproducibility project, under which the Bank is publishing 'reproducibility packages' alongside all their reports, papers and technical notes. www.nature.com/articles/d41...

The World Bank is right to prioritize reproducibility

The development bank and others are starting to publish and verify the raw data and code accompanying their research-based reports and papers. Other agencies should do the same.

nature.com

Curious about Zenodo as an OSF Project alternative? Here's a short video to get you started: File Uploads, Communities, and generating links for Anonymous Peer Review. youtu.be/jW84PxuJTMw

Getting Started with Zenodo (transitioning from OSF Projects)

YouTube video by AMANDA MONTOYA

youtu.be

1/ What is the prevalence of personally identifiable information (PII) in author-collected data publicly shared in the social sciences? We audited 327 replication packages from 11 leading journals in economics, political science, and psychology Answer: surprisingly high 🧵

In diesem in der @pvs-journal.bsky.social veröffentlichten Aufsatz verorte ich Manows Buch "Spaltungslinien" in der Debatte zum Postliberalismus. Der zentralen Annahme, dass rechtspopulistische Parteien heute wirtschafts- und sozialpolitisch links stünden, widerspreche ich. rdcu.be/swtwc7WqF9ta

Scientific datasets are riddled with copy-paste errors www.sciencedetective.org/scientific-d... 18 datasets out of 600 are reported as having serious issues. This is bad enough, but not as bad as the title may suggest. The approach is simple, but powerful and interesting. 1/

One of the better arguments that Claude is human-level intelligence is that it doesn't understand tikz either.

$100k is a lot, even for what they offer. Who can afford this? (not that you should use it even if you could afford it) And what guarantees are there they produce legit research and do not simply spend a fraction on a fully AI-produced paper with fake data?

Academic Gossip@academicgossip.bsky.social · 2w ago

Spotted: The modern Paper Mill. Price tag: $100k. FastCode AI is offering Research-as-a-Service (RaaS): for $100k, you bring the problem and the data, they do the research, run experiments, write the manuscript. Target venues: NeurIPS, CVPR, ICLR, AAAI. Timeline: 3-4 months. #AcademicGossip

Ask yourself: What would be the consequences for a student in your course doing this? Now ask yourself: What should be the consequences for a professor?

Data Colada@datacolada.bsky.social · 2w ago

Colada[139] Artificial Deadlines Part 2 Evidence of fraud in the 25 year old gradesheet behind the key finding in an influential paper on deadlines and procrastination. datacolada.org/139

The Journal of Trial and Error is looking for a new Editor with an interest in meta-research, open science, or academic publishing. More info: www.linkedin.com/posts/were-l...

Job ad- Assistant Professorship in Political Science at The University of Gothenburg web103.reachmee.com/ext/I005/103...

🧐 What do citizens think about "null results" in scientific research? To find out, @johnholbein1.bsky.social & I fielded a large-scale online experiment. We gave respondents straightforward info about a study, asked them some Qs, showed them the results, then asked them Qs again.

Canada has the most ambitious program to poach academic talent from the U.S., with a budget of over 1 billion CAD, but the first intake only has 61 people listed — with just 7 in the broader social sciences.

My paper with Sam Pimentel on worst-case confounders is now out at Observational Studies in a special issue on Cornfield sensitivity analysis! (1/5) muse.jhu.edu/pub/56/artic...

Psychological Science special issue call 📣 Ever had your research denigrated bc it's "only descriptive", "what's the mechanism though?" and "but what are the theoretical implications?" Now has come your time to shine. Send your best descriptive work! www.psychologicalscience.org/publications...

Ein paar Worte zu „Linke Politikwissenschaftler wie Tarik Abou-Chadi…“ Eine ungute und gefährliche Entwicklung, dass sich solche Formulierungen immer häufiger finden. Gerade in Situationen, in denen wissenschaftliche Evidenz der eigenen Argumentation entgegensteht. Das hier ist ein Paradebeispiel.

New paper alert! "Mini-Many Analysts: Using Registered Reports to Learn About Analytical Flexibility" in Teaching of Psychology! Want to teach about #OpenScience in undergrad psychology classes? We propose a project that integrates #RegisteredReports and #ManyAnalysts! doi.org/10.1177/0098...

Mini-Many Analysts: Using Registered Reports to Learn About Analytical Flexibility - Amanda Kay Montoya, Gareth Ong Yu, John Tawa, 2026

Introduction Analytical flexibility describes the multiple defensible analytic strategies researchers can use. Registered reports reduce analytical flexibility ...

doi.org

I read The Economist article about Acemoglu. It feels like someone let out his/her reviewer 2. www.economist.com/finance-and-... I think there is an interesting question in it: "whether his reputation at the very apex of the economics profession is justified by his scholarship". 1/