When you collect data online, are the results from humans or AI? In a project led by Booth PhD student Grace Zhang, we estimate the prevalence of AI agents on commonly used survey platforms:
osf.io/preprints/ps...
🧵
Thanks!! I just ordered it.
Bots have made their way to Prolific experiments. Our lab has stopped online testing of adults entirely now for this reason - we want to know if what we study is real. Probably data collected 2-3 years ago are ok, but moving forward we just can't know. www.pnas.org/doi/10.1073/...
"Machine learning research is not serious research & therefore hallucinated references are not necessarily a big deal, agrees prestigious group of ML researchers"
I ❤️ when the titles write themselves
But seriously, I don't think there's much to debate statmodeling.stat.columbia.edu/2026/01/26/m...
My editorial approach has always been pretty simple:
1. Is this a good research question?
2. Has an appropriate method been used to answer it?
And that is about it.
Oh Canada!
The results of the RCT didn't support our hypothesis ➡️ The RCT "failed".🤦♂️
This is not how science works! #ResearchIntegrity
👍Kudos to Charlotte Strøm for the review
www.linkedin.com/pulse/review...
Made me laugh out loud. Thank you!
Madness! This must be why some people some people are passing pronoun laws. I get it now.
Experimentology cover: title and curves for distributions.
Experimentology is out today!!! A group of us wrote a free online textbook for experimental methods, available at experimentology.io - the idea was to integrate open science into all aspects of the experimental workflow from planning to design, analysis, and writing.
The poll numbers that matter: www.vancouverisawesome.com/local-news/p...
@carlbergstrom.com
Great study! Really interesting that the previous findings are now reversed. We need more replication (& extension) studies like this in the social sciences!
Seems the White House is using gen AI to write their scientific reports, leading to the hallucinations and mistakes you'd expect. Laughing/crying right now.
www.nytimes.com/2025/05/29/w...
Exclusive: National Science Foundation staff were told today that the agency’s 37 divisions—across all eight directorates—are being abolished and the number of programs within those divisions will be drastically reduced.
Also check out this excellent Data Colada post if you haven't already:
datacolada.org/124
@urisohn.bsky.social reanalyzed the AER paper data and concluded that the key results were mainly driven by confused participants.
Climate Terminology Does Not Matter
Across tens of thousands of participants in two large-scale experiments, we found that labeling climate change in different ways had no effect on their stated willingness to act.
jayvanbavellab.substack.com/p/climate-te...
via @dgoldwert.bsky.social
Chegg, notorious resource for student cheating, is mad that Google is copying its answers.
Are you running AB testing studies on Facebook or Google? Or reviewing papers using them?
Check out our open access paper, On the Persistent Mischaracterization of Google and Facebook A/B Tests: How to Conduct and Report Online Platform Studies at doi.org/10.1016/j.ij...
@boegershausen.bsky.social
IMPORTANT: If you live in the United States or Canada, you’ve probably heard Trump‘s constant claim that the United States is “subsidizing” Canada. It’s wrong of course, but here’s the best explanation I’ve seen about how & why it’s wrong. PLEASE WATCH & SHARE IT A LOT. Kudos to Andrew Chang & CBC.
Underreported: Elon just picked the head of NASA—a massive conflict-of-interest that will damage a government entity that he partners with, depends on, and in some cases competes with. The nominated NASA director is a high-school dropout billionaire who is pals with Musk. He has no business at NASA.
Great write-up! Thank you!
JDM pub:
Intertemporal choice questions are usually like "Receive $10 now, or $15 next month?" Yet real-life decisions usually involve a mix of negatives and positives. Do mixed-valence questions better predict real-life behaviours? No! go.shr.lc/3OG36rz
@yeli.bsky.social @sjdm-tweets.bsky.social
A graph depicting the global monthly mean carbon dioxide (CO2) levels from 1980 to 2024, showing a steady increase in CO2 mole fraction measured in parts per million (ppm), with data points represented by red dots and a trend line.
Black Friday Deal Alert!
Enjoy our current atmospheric CO₂ because we'll never see it this low again!
Introducing RegCheck: a tool which uses Large Language Models to automatically compare preregistered protocols with their corresponding published papers and highlights deviations.
@malte.the100.ci @ianhussey.bsky.social @ruben.the100.ci @bjoernhommel.bsky.social
regcheck.app
the fallout from the Francesca Gino research misconduct scandal isn't over. @engber.bsky.social writes about how an effort by some of her colleagues to self-audit their work with her ended up raising even more questions — for them personally and their whole field www.theatlantic.com/magazine/arc...
The superendowment effect—research by Winegar and Sunstein finds users would demand $80 to share personal data, but pay only $5 to keep them private. This raises questions over the reliability of WTP/WTA to gauge the welfare significance of giving up or maintaining privacy: https://buff.ly/3Oa5jvr
Prof Kai Chan and I are recruiting a PhD student to start at UBC starting in Fall 2025, to work on Green Death Care and Relationships with Nature. Looking for applicants with a background in survey and experimental design and analysis. Dec 15 application. deadline. Details: www.grad.ubc.ca/ad/58334
For any new @bsky.app users who are interested in #ResearchIntegrity matters, I've got a #StarterPack for you to help build your #Bluesky feed:
go.bsky.app/5NJ9Z4N
I like this method for finding new people to follow and interact with. And do let me know if you'd like to be added to the list
Hi there. We're Retraction Watch -- retractionwatch.com -- and we've joined Bluesky.
