Jay Van Bavel, PhD

Are There Ideological Differences in Science Denial? Debates about issues such anthropomorphic climate change have raised alarm that there may be systematic differences in liberals’ and conservatives’ tr…

"In countries with multiparty systems or less polarized media ecosystems than the U.S., ideological and partisan gaps in self‐reported trust in science are typically substantially weaker than in the U.S., if they exist at all... not a stable characteristic of conservativism"

They are less likely to judge outgroup members more harshly than ingroup member. Instead, they treat all transgressors equally because they see people as individuals:

People who have more complex social identities view political identity as a less reliable source of information about others. Therefore they rely less on group-based stereotypes when judging outgroup members.

www.sciencedirect.com/science/article/pii/S0022103125000915

In a new paper, we find that sycophantic #AI chatbots make people more extreme--operating like an echo chamber

Yet, people prefer sycophantic chatbots and see them as less biased

Only open-minded people prefer disagreeable chatbots: osf.io/preprints/ps...

Led by @steverathje.bsky.social

Close-up of three people using hand gestures while engaged in a serious conversation indoors.

Do people expect to be judged by their own moral standards or someone else's? New #SPPS research reports that we actually expect out-group members to judge us by their values, not ours.

📖 Read more: ow.ly/1tnF50X5yNP

Other coauthors are @MerylYe @rmpillai.bsky.social @LauraGlobig and @VictoriaOldemburgodeMello

In a new paper, we find that sycophantic #AI chatbots make people more extreme--operating like an echo chamber

Yet, people prefer sycophantic chatbots and see them as less biased

Only open-minded people prefer disagreeable chatbots: osf.io/preprints/ps...

Led by @steverathje.bsky.social

I’m so excited about this—it’s great for APS to host their main conference in another country Barcelona is such an incredible city!

I’m so excited about this—it’s great for APS to host their main conference in another country Barcelona is such an incredible city!

Larger scientific teams produce more impactful breakthroughs + patents

I wrote a paper with 4 principles to optimize large scientific collaborations based on the science of cooperation & collective intelligence: osf.io/preprints/ps...

And a substack summary: powerofus.substack.com/p/four-princ...

Abstract and results summary

🚨 New preprint 🚨

Across 3 experiments (n = 3,285), we found that interacting with sycophantic (or overly agreeable) AI chatbots entrenched attitudes and led to inflated self-perceptions.

Yet, people preferred sycophantic chatbots and viewed them as unbiased!

osf.io/preprints/ps...

Thread 🧵

Larger scientific teams produce more impactful breakthroughs + patents

I wrote a paper with 4 principles to optimize large scientific collaborations based on the science of cooperation & collective intelligence: osf.io/preprints/ps...

And a substack summary: powerofus.substack.com/p/four-princ...

One of the favorite parts of my job is working with the next generation of scientists to understand the science of conflict & cooperation.

Every 2 years I host a conference with Tessa West where we invite our former PhD students & postdocs to share new research on these topics and reconnect.

Our new paper finds that AI can overcome partisan #bias

We find that AI sources are preferred over ingroup and outgroup sources--even when people know both are equally accurate (N = 1,600+): osf.io/preprints/ps...

NEW: We filed a brief in support of a lawsuit against the Trump administration’s unprecedented attacks on academic freedom at Columbia University.

The First Amendment doesn't allow the president to force a university to censor its teaching, research, and faculty to align with his ideology.

You’ll have to ask Steve and my other postdocs typical that is. i Think take out lunch is extremely common, dinner out is somewhat common and cocktail bars are less common

We have a paper on how that happens too...watch this space in a couple of days.

This video describes a day in the life of a researcher at New York University

My Postdoc @steverathje.bsky.social made this video and it provides a sneak peak into what it looks like to work at the Center for Conflict and Cooperation ww.centerconflictcooperation.com

Let us know what you think!

Agree 100%. That's a point we directly raise in our general discussion.

Our data suggest that AI’s perceived neutrality could help reduce “echo chambers” by broadening people’s exposure to diverse information--so long as AI systems remain trusted as impartial and prioritize accuracy. osf.io/preprints/ps...

People trusted ingroup over outgroup sources, but...
- In a political fact-checking task, people preferred AI over outgroup and ingroup sources
- We found a similar pattern for a task that was completely unrelated to identity
- This was not driven by priors, but by how people accumulate evidence

Do fact-checks really work against misinformation? A recent study shows that fact-checking reduces false belief in misinformation

Study shows that fact-checking reduces false belief in misinformation.

A 2023 @bjpols.bsky.social article by Alexander Coppock, Kimberly Gross, @ethanvporter.bsky.social, @emilythorson.bsky.social & @thomasjwood.bsky.social has been referenced by @3streamsblog.bsky.social.

https://cup.org/3WdASby

Our new paper finds that AI can overcome partisan #bias

We find that AI sources are preferred over ingroup and outgroup sources--even when people know both are equally accurate (N = 1,600+): osf.io/preprints/ps...

White House considers funding advantage for colleges that align with Trump policies The proposal has the potential to transform the government’s vast research funding operation, which has long awarded university grants based on scientific merit.

Exclusive: The White House is developing a plan that could change how universities are awarded research grants, giving a competitive advantage to schools that pledge to adhere to the values and policies of the Trump administration.

New PSA paper led by Tom Heyman!

We present a tutorial on crowdsourced MULTIVERSE ANALYSES—a transparent way to test how robust findings are across many analytic paths. Experts suggest, then validate possible paths. This helps quantify uncertainty and clarify when/where scientific conclusions hold.

Study: Most Americans Do Not Support Partisan Violence | Dartmouth

“Democrats think 45% of Republicans support partisan murder, 20 times larger than what data show. Similarly, Republicans think 42% of Democrats support partisan murder, which is 25 times larger than reality. Americans have a fundamentally incorrect understanding of support for partisan violence.”

Good Conversations Don’t Require Everybody to Agree, Neuroscience Shows Brain imaging is illuminating the patterns linked to productive, positive dialogue, and those insights could help people connect with others

@falklab.bsky.social wrote a nice summary of our hyperscanning work on how people form social connection and find agreement through conversation: www.scientificamerican.com/article/good...

🚨New preprint🚨

osf.io/preprints/ps...

In a sample of ~2 billion comments, social media discourse becomes more negative over time

Archival and experimental findings suggest this is a byproduct of people trying to differentiate themselves

Led by @hongkai1.bsky.social in his 1st year (!) of his PhD

Why #misinformation must not be ignored - A new, open access article by @ulliecker.bsky.social and colleagues, published in American Psychologist bit.ly/4pJ20fY

Social media is one the least meaningful things we do; spending time with family is one of the most.

Yet we are spending more time on social media and less with our kids.