#Replications
Nice to be in California for a whole week for the #psa2026
I am gonna be talking about what replications have to do with test severity, and what test severity has to do with underdetermination of hypotheses by data, and ultimately why need a more principled approach to replications.
November 18, 2026 at 1:11 AM
Successful replications
November 13, 2026 at 6:47 PM
First time going to a hackathon alone and it was an amazing experience!

So many smart, creative people, all working towards #MakingReplicationsCount
Special thanks to @aufdroeseler.bsky.social, @lukaswallrich.bsky.social and @josefinaw.bsky.social for organizing.
November 14, 2026 at 10:34 AM
the dominance of man over woman, of adult over child, of white over nonwhite: all mirror the dominance of capital over worker, they are functional and necessary replications of this central antagonism! The abuse of the worker is tantamount the abuse of the child.
November 14, 2026 at 10:19 AM
Task: find replications of previous experiments in the multidisciplinary field of research on charitable giving.

A comprehensive manual search identifies 48 replications of experiments. How many replications from the set of 48 can AI-powered tools identify?

Not many, it turns out. The best […]
November 17, 2026 at 11:58 AM
Following this account has made me think about the shift that happened in ideaspace between the VHS era and the HD era. Maybe replications of reality were always supposed to be imperfect and we sinned when we made them too clear, too indistinguishable
November 14, 2026 at 4:35 AM
When you see this in action its so much more mind blowing.

See an article mentioned on a webpage and be given links to replications of that study, pubpeer comments, retractions. Wow.

Incredible work by the @forrt.bsky.social folk.
November 11, 2026 at 9:25 PM
Lukas is a trained psychologist and meta-scientist. He is the managing director of the Münster Center for Open Science, Editor-in-Chief of the scholarly-led journal @r2journal.bsky.social, & co-founder of the world’s largest curated database of replications. More info: lukasroeseler.github.io/home/
Dr. Lukas Röseler – Lukas Röseler
lukasroeseler.github.io
November 17, 2026 at 4:29 PM
I remember them too, which is why I'm glad that today there are plenty of options for those who want something that feels like the software from the old days. One of them recently made the news...

www.tomshardware.com/software/mic...
Creator of free Adobe clones unveils open-source Microsoft Office Excel, PowerPoint, and Word replications — Autodesk tools and other subscription-based services may come next
It's a protest now.
www.tomshardware.com
November 13, 2026 at 5:30 AM
Our website and submission system are currently slowed-down due to bot attacks on scholarly-led journals in North-Rhine Westphalia (a part of Germany). Our mirror website including all articles is unaffected: www.replicationresearch.org

If you want to submit an article, please send us an e-mail.
Replication Research
A diamond open-access journal publishing replications, reproductions, and meta-scientific research.
www.replicationresearch.org
November 17, 2026 at 3:08 PM
New post on what using AI for peer review or replications might mean. I argue that it depends on whether peer review takes place in adversarial or cooperative contexts. computingandsociety.substack.com/p/the-future...
The future of replications and peer review in the age of AI is not great.
Not unless we make fundamental changes to how scientific publication is organized. Otherwise, LLM-driven peer review will generate more heat than light, and significantly less trust from the public.
computingandsociety.substack.com
November 17, 2026 at 2:19 PM
Would love an episode about replications of the self (clones, replicas, uploaded consciousness, etc)
November 11, 2026 at 5:21 PM
Damian Pembroke kept Viogenix separate from DEP, and Bob, for a reason.

Maybe a two-factor authenticator would have been nice to have in this situation?

#TheAechFiles #BeforeSilence #EatingLife #DystopianFiction #PsychologicalHorror #HorrorFiction #SpeculativeFiction #IndieAuthor #BookSky
November 18, 2026 at 2:23 PM
Very excited about the new open source browser plugin that informs you about free PDFs, pubpeer comments, retractions, registered reports, and of course replications. This is one of the outputs for a UKRI project and was led by Keegan Vaz: forrt.org/marco/
Has this paper been replicated? Retracted? Is there a Registered Report?
FORRT ORE, our free and open-source extension for Chrome and Edge, shows you in Google Scholar, PubMed, Scopus and more.
The first version is out now. A thread on what it does:
forrt.org/chromium-ext...
November 10, 2026 at 11:05 AM
For each paper, ORE shows:

• Replications and reproductions, colour-coded by outcome
• Retractions and expressions of concern (Retraction Watch)
• Registered Reports, with the Stage 1 protocol and Stage 2 report linked
• PubPeer discussions
• Links to free full text
November 10, 2026 at 10:46 AM
This is our first release, so we’d like to hear what works for you and what doesn’t. Use “Report an issue” in the extension.

Install for Chrome or Edge:
chromewebstore.google.com/detail/ffncp...
FORRT ORE — Open Research Extension - Chrome Web Store
Shows Replications, Retractions, Registered Reports, PubPeer comments and open-access copies for the papers you search for and read.
chromewebstore.google.com
November 10, 2026 at 10:46 AM
Sounds right, proper delicacies are just what they are, and in the current moment best understood as positional goods. The fall off to replications is huge and we’d all be better off just enjoying what we have, well prepared
November 10, 2026 at 5:53 AM
Article highlight 📰

Replications are still rare in behavioral epigenetics. In 116 formerly out-of-home placed young adults, we preregistered a replication of shared and type-specific DNA methylation signatures of childhood maltreatment reported by Cecil et al. (2016). #neuroskyence 🧪 #OpenAccess
Epigenetic signatures of different types of childhood maltreatment in a high-risk sample: a preregistered replication study - Translational Psychiatry
Translational Psychiatry - Epigenetic signatures of different types of childhood maltreatment in a high-risk sample: a preregistered replication study
doi.org
November 10, 2026 at 8:00 AM
a recent English study following over one million matched booster recipients for nearly a year. Although these are not exact replications, lack of consistent independent corroboration weakens the interpretation that Florida identified a large, persistent, broadly applicable harmful effect of Pfizer
November 8, 2026 at 2:39 AM
In 50 simulated markets where budgets follow demand, the 94% interval for TV’s marginal return covered the truth 4% of the time for a plain MMM, 22% with lift-test calibration, and 94% with the budget model and test design. Intervals are wider.
November 9, 2026 at 10:23 PM
What you've described here is copyright infringement and not theft. Furthermore, it's a poor analogy for what generative AI is generally doing, as its purpose is to create novel outputs and not replications.
November 8, 2026 at 1:06 AM
A shelf on failed replications in psychology ends with a retracted paper.

Protzko et al. 2023, Nat Hum Behav, had reported high replicability.

The curator: "The best example in the selection!"

https://readmarginalia.com/collections/the-reproducibility-crisis-in-psychology
November 11, 2026 at 10:46 AM
It also takes ordinary instruments (e.g. media-cost shocks) via instruments=, handles holdout designs, and works with incrementality, the budget optimizer.

The control-function coefficient is an exogeneity check: near zero if spend was exogenous.
November 9, 2026 at 10:23 PM