Emir Efendić

I’ll just gpt code it to something fancy 😀

homer simpson from the simpsons is standing in front of a grassy field . ALT: homer simpson from the simpsons is standing in front of a grassy field .

Looking at all these fancy people here making jokes about R and pipes and tibbles while I'm just doing basic Qualtrics experiments and my code to clean the data has been the same for the last 4 years.

Online Studies Psychological Science requires that authors who use samples from online data collection include a statement in the Method section explicitly addressing their approach to preventing and detecting automated or AI-generated responses. Rationale As large language models and other generative AI tools become more accessible, the risk of data contamination by non-human respondents has increased dramatically in research. Psychological science (and the social sciences generally) is particularly susceptible to this issue given its growing reliance on online data collection. Preventing automated responses during data collection and detecting them afterward often involve methodological trade-offs. For instance, technical barriers that aim to prevent LLM use (e.g., blocking copy-pasting functionalities) may eliminate behavioral indicators needed for detection (e.g., pasting rather than typing). This policy aims to enhance transparency and reproducibility of reported results by requiring authors to articulate their approach across both prevention and detection dimensions, enabling readers and reviewers to assess the likelihood of reported data being influenced by automated responses. Scope This policy applies to any submission with at least one study that includes data collected online without direct human supervision (e.g., via crowdsourcing platforms, student participants who complete the study online, online recruitment ads, or remote survey distribution tools). Required Reporting Authors must include in the Methods section either: A statement confirming that procedures were in place to prevent and/or detect and exclude automated or AI-generated responses, including a description of those procedures (e.g., explicit participant instructions against LLM use, disabled copy–paste functionality, CAPTCHA use, IP filtering, consistency checks, attention checks, adversarial prompting) as well as the types of automated responses that these procedures are suitable …

Maybe of interest: The submission guidelines of Psychological Science now demand an explicit statement on measures taken to reduce the risk of AI-generated responses for all online studies!

www.psychologicalscience.org/publications...

Many are appropriately outraged by Altman’s comments here implying that raising a human child is akin to “training” an AI model.

This is part of a broader pattern where AI industry leaders use language that collapses the boundary between human and machine.

🧵/

Built out in the last couple of years. We added mounds upon mounds.

Spent the day skiing. On actual snow.
I see people are talking about pipes here while I’m skiing this half pipe. (Note: no pipes or half pipes were harmed.)

This is fascinating. I always used these chills and goosebumps to tell me what I liked but also when I play music to figure out which tones work with each other even before knowing theory e.g. scales.

Could be interesting to: @dgrand.bsky.social @gordpennycook.bsky.social @tomcostello.bsky.social

There's a lot of nuances to this finding. Some improvement was also observed in the agreeing LLM condition, and conversations weren't uniformly beneficial.

In fact, when the first prediction was pretty accurate, we saw a slight decrease in accuracy.

Comments are welcome!!!

Now for the cool stuff: when people talked to a disagreeing LLM they revised their predictions more and were more likely to revise them in the right direction (upper panels). This improved accuracy (lower panels) and the improvement occurred much more when initial predictions were inaccurate.

We had people make predictions and either converse with an agreeing LLM or a disagreeing LLM.

They had to explain their reasoning behind the prediction and after the conversation, they could take another shot at the prediction.

Disagreement (having one's views challenged) is a really good way to improve decisions. But, people avoid it because it's uncomfortable (among other things).

But LLMs are really good at conversation so we thought why not leverage this to deliver disagreement without the social consequences.

We have a new pre-print! 📝🖨️

We find that conversing with a disagreeing LLM helped improve people's inaccurate predictions!

osf.io/preprints/ps...

Let me tell you all about it:

📣 Applications for the 23rd Summer Institute on Bounded Rationality are now open!

✨Join us in Berlin @arc-mpib.bsky.social June 08–16, 2026, to explore the topic of “Decision Making in the Age of AI”.

✏️ More details + application form (deadline: March 16): www.mpib-berlin.mpg.de/research/res...

Sir Ian McKellen performing a monologue from Shakespeare’s Sir Thomas More on the Stephen Colbert show. Never have I heard this monologue performed with such a keen sense of prescience. Nor have I ever been in this exact historical moment.TY Sir Ian, for reaching us once again.
#Pinks #ProudBlue

New paper (forthcoming in Cognition): Context-dependent effects of branches in decisions under risk authors.elsevier.com/a/1mXL%7E2Hx...
Key finding: when people choose between risky options, they’re more likely to pick the one with more distinct probabilistic outcomes (“more pathways to winning”).

Here’s Duke behavioral economist Dan Ariely asking Jeffrey Epstein for “the name and email of the redhead that was here with you.” This is four years after Epstein’s 2008 conviction for procuring a child for prostitution.

Introducing Causion: A web app for playing with DAGs | Peder M. Isager Personal website of Dr. Peder M. Isager

New blog post introducing Causion - a web app for causal inference teaching and learning: pedermisager.org/blog/causion....

The Iowa Gambling Task is an extreme example of Jingle Fallacy and schmeasurement.

In 100 articles we found 244 different ways of scoring it, 177 were never reused. Correlations between them range -.99 to .99.

At the same time, we show meta-analyses combine these results as if they’re equivalent.

RegCheck RegCheck is an AI tool to compare preregistrations with papers instantly.

Comparing registrations to published papers is essential to research integrity - and almost no one does it routinely because it's slow, messy, and time-demanding.

RegCheck was built to help make this process easier.

Today, we launch RegCheck V2.

🧵

regcheck.app

Soooo is there an explanation about Macron’s shades or are we all just cool with it?

This feels like one of those things we’re gonna look back to in a generation whilst gasping in shock.

Yes can confirm this just happened to me today. I know it's an OSF pre-print but I could only find it through a link on a third party website.

Behavioral Science What do we do? We conduct large-scale behavioral science studies to improve the daily decisions, behavior, and experience of vulnerable individuals (e.g., the well-being of citizens and families / ed...

Do you want to work with me?:) Please spread the word! We are looking for talented Post-doc candidates for a 10-month Junior Fellowship at the Behavioral Science Center, hosted by the Corvinus Institute for Advanced Studies (CIAS). 1/6

www.the-bs-lab.com

I’ve spent this morning morbidly watching the MTV channels close down and the final songs played were:
MTV Music - Video Killed the Radio Star (Buggles)
Club MTV - Don’t Stop the Music (Rihanna)
MTV 90s - Goodbye (Spice Girls)
MTV 80s - Together in Electric Dreams (Oakey and Moroder)

These types of papers are really nice to see out there.
An adversarial collaboration sounds like an oxymoron but seems to be producing good science.

Text reads: About synthetic panels Recruiting the right participants for a study can be difficult. You may not get the exact demographics you need, and the shorter the deadline, the less sure you can be that everyone will answer on time. One possible solution can be to use synthetic panels. Synthetic panels are powered by a first party proprietary AI model developed here at Qualtrics. Our synthetic panel is trained on thousands of responses from a variety of demographic backgrounds in order to more accurately predict how certain populations would respond to a survey. Our synthetic panel is based on the United States General Population, and is only available in English. This panel comes with ready-made quotas and target breakouts in order to represent your chosen population and make it easy to launch your survey right away.

Text reads: Question-writing best practices To get the most reliable and actionable results from synthetic audiences, consider these question-writing best practices: Ask forward-looking and attitudinal questions. Synthetic panels perform best with perceptions, preferences, and intent-based questions. For example, “How likely are you to try…?” Synthetic panels are less applicable for studies on past behaviors, detailed recall, brand recall, or awareness questions. For example, “When did you last visit…?”

Text reads: Discussion The current study aimed to conduct a meta-analysis of the TPB when applied to health behaviours which addressed the limitations of previous reviews by including only prospective tests of behaviour, applying RE meta-analytic procedures, correcting correlations for sampling and measurement error, and hierarchically analysing the effect of behaviour type and sample and methodological moderators. Some 237 tests were identified which examined relations amongst model components. Overall the analysis indicated that the TPB could explain 19.3% of the variance in behaviour and 44.3% of the variance in intention across studies. This level of prediction of behaviour is slightly lower than that of previous meta-analytic reviews which have found between 27% (Armitage & Conner, 2001; Hagger et al., 2002) and 36% (Trafimow et al., 2002) of the variance in behaviour to be explained by intention and PBC.

Did you know that from tomorrow, Qualtrics is offering synthetic panels (AI-generated participants)?

Follow me down a rabbit hole I'm calling "doing science is tough and I'm so busy, can't we just make up participants?"

12 December 2025 Thoughtfully Shaping Our Digital Future To the parties forming the government in the Senate and House of Representatives, as well as the outgoing administration, We are writing to you in recognition of your crucial responsibility for shaping current and future AI policy, overseeing digitization, and upholding public values. We are a coalition of scientists, experts, and representatives of civil society organizations. We believe it is essential to address these matters together. This letter has two objectives: a) Provide context for current plans, including the AI Delta Plan, the AIC4NL position paper, and the Invest-NL AI Deep Dive. These investment proposals often rely on assumptions that lack scientific evidence and do not fully reflect public values. b) To offer a constructive, well-substantiated alternative approach to digital futures based on people, nature, and democracy. We believe a collaborative process should guide decisions on the needs, scope, and nature of investments by bringing together scientists, civil society organizations, and stakeholders.

📝 OPEN LETTER 📝

Are you based in NL 🇳🇱 ? Do you also want government to thoughtfully shape our digital future, with care for people and nature? Please share and sign 🖊️ this letter addressed to parties forming the new Dutch government and outgoing administration.

📝 openletter.earth/zorgvuldig-a...

This is nuts. I urge my fellow academics to take this into account, for example when organizing conferences and the like

Not only does this makes sense and is encouraging but also, any primacy that the US had in terms of conference locations for academics should be stopped.

The south of (insert any European country) is better.