One of the reasons we loved Mickey as grad students is because he would share all his opinions and he welcomed disagreement. He was fun to debate and never got defensive. This made science discussions with him a lot of fun. He now writes the “Speak Now, Regret Later” substack where he shares his controversial takes on the field of social psychology and often responds to his critics.

Speak Now Regret Later

An unfiltered exploration of psychology, culture, and the world through the eyes of an impulsive yet remorseful professor

By Michael Inzlicht

We invited him to share his controversial column on the downfall of Stereotype Threat. When he started at the University of Toronto, Mickey was most known for his research on stereotype threat and that line of research was among the most prominent ideas in the field. We remember walking through SPSP during graduate school and seeing countless posters on implicit attitudes, ego depletion, and stereotype threat—it seemed like everyone was studying at least one of those topics.

In the last decade, all of those topics have come under a great deal of scrutiny. A mix of failed replications and methodological criticisms has called into question these bodies of research. (Jay has a chapter on the “replication crisis” in social psychology that you can read here). And, while there is still ongoing debate about the scope and scale of these ideas, we wanted to share Mickey’s perspective on the topic since it touches on the core themes of our newsletter (i.e., identity and group dynamics).

Another day, another idol falls.

This one has been teetering for years, so the collapse didn’t come as a shock. But that doesn’t make it any less painful.

I’m talking about stereotype threat, a once-revolutionary idea that shaped how social psychologists thought about identity, achievement, and inequality. For decades, it inspired research, drove interventions, and promised insights into the invisible forces that constrain human potential.

I still remember seeing its most eloquent advocate, Stanford University’s Claude Steele, deliver a keynote address in 1999 at the annual convention of what was then called the American Psychological Society. It was my first ever conference, my first trip to Denver, and Steele was nothing short of magnetic. Charismatic and at the height of his powers, he commanded the stage like no academic I had ever seen. He delivered his message with the kind of confidence that makes you believe science can change the world. Professor Steele was a rock star, and I was as giddy seeing him on stage as I was seeing Kurt Cobain on stage a few years earlier.

What is Stereotype Threat?

The concept of stereotype threat, first proposed by Claude Steele in the early 1990s, posited that individuals who are part of a negatively stereotyped group can, in certain situations, experience anxiety about confirming those stereotypes, leading paradoxically to underperformance, thus confirming the disparaging stereotype. The initial research was groundbreaking.

In 1995, Steele and his student Joshua Aronson—who went on to become my postdoc supervisor years later—demonstrated that the notorious Black-white gap in academic performance could be partially closed when negative stereotypes impugning Black people’s intelligence were made irrelevant. When Black students at Stanford University were told that a test was diagnostic of intellectual ability, they performed worse than their white counterparts. However, when this stereotype threat was ostensibly removed—by simply framing the test as a measure of problem-solving rather than intelligence—the performance gap between Black and white students nearly vanished.

Suddenly, here was an explanation for why certain groups didn’t perform as well in academic settings. And it wasn’t just race; follow-up studies looked at women in math and science. Women, who dominate men in most academic disciplines, underperform in STEM fields because they were regularly, albeit subtly, reminded of the stereotype that women aren’t good at math, or so the story goes. The idea felt revolutionary, hopeful even, because it suggested that these vexing performance gaps could be addressed by changing people’s immediate environments rather than accepting them as fixed outcomes, inherent to the groups themselves

These findings were exhilarating. Before long, stereotype threat was not only the darling of social psychology, but it also became the darling of the political left who now had an answer to prevailing views of group differences held by the political right. This is partly because shortly before stereotype threat took its turn in the spotlight, Charles Murray and Richard Herrnstein published The Bell Curve, which resulted in a media firestorm that has had repercussions to this day. Not only did the book discuss racial differences in intelligence as real and consequential—and not mere products of culturally biased IQ tests—it suggested that a non-negligible factor in this gap was due to biological differences. This thesis was so toxic that the octogenarian Murray is still considered a pariah, shouted down and deplatformed from talks he tries to deliver at respectable colleges to this day.

Stereotype threat, in contrast, was a breath of fresh air. It promised that group differences were malleable, not fixed. They could be explained as momentary apprehension, akin to the nerves that might cause an elite athlete to choke on competition day. Yes, these group differences still have consequences, but now we have a remedy—change the situation so that stereotypes are less likely to be in the air and watch as all the Black students and female mathematicians rise to the top.

I too was swept up by this mania. I studied stereotype threat as a PhD student and published some of the first papers on the topic. My dissertation and very first publication suggested that subtle aspects of a room—like how many men and women were in a math classroom—could be enough to evoke stereotype threat and undermine performance. Because my field became captivated by stereotype threat, this meant that I was quickly offered jobs, grants, tenure, and acclaim. I edited a book on stereotype threat and was asked to add my name and research to briefs delivered to the US Supreme Court. My career benefitted immensely.

The Replicators are Coming

Then things started going sideways. And not just for stereotype threat.

It’s all very complicated. Lots of strands in old Duder’s head. But here’s the skinny. In the early 2010s, psychology started looking inward, asking hard questions about the robustness of our most cherished findings. This happened because in the early 2000s, our best journals regularly included studies that were ludicrous and hard to believe. For example, researchers made dubious claims about the role of blood glucose and self-control, and found positive evidence for these claims, despite their biological impossibility. A paper was published in social psychology’s most prestigious journal claiming evidence for clairvoyance, essentially offering a ringer for a ringer. Ludicrous. If these impossible ideas were generating support with the standard methods of social psychology, maybe our methods are not what we thought they were.

A small cadre of reformers then started raising awareness that all was not right in how we conducted our science: we did not bother replicating important studies, we were misusing and abusing our statistical tools, and we did not publish all our studies—particularly the failed ones. And when some brave scientists decided to audit the field by closely replicating many studies, only about a quarter from social psychology could be successfully replicated. Since these dark days, the field has changed immensely, and we’re slowly producing more respectable science today.

Nonetheless, the entire field’s evidentiary basis was now suspect. After all, they were produced by methods that we now consider questionable. Stereotype threat was no different. I would love to say that stereotype threat was an exception, that it survived replication attempts and other audits, and that a beloved idea can still be used to counter damaging claims about group differences. But new data now reveal what many of us suspected for at least ten years: stereotype threat does not replicate, and it does not undermine academic performance in the ways we thought.

This new data emerged from what is called a Registered Replication Report. This was no ordinary replication study; it used the gold standard of scientific rigor. Conducted by multiple labs across the U.S. and Europe, and led by Andrea Stoevenbelt this study (still a preprint) was preregistered (meaning all methods and analyses were specified before the data were collected) and involved over 1,500 participants. It replicated the exact procedures of a well-known stereotype threat study published in 2005 by Mike Johns, Toni Schmader, and Andy Martens—all colleagues and friends I deeply respect.

The original study had found that women performed worse on math tests when reminded of gender stereotypes but performed on par with men when they were instead taught about stereotype threat. The idea was that awareness of the phenomenon of stereotype threat helped mitigate its effects, which was why this original paper was so influential: it offered a simple intervention to close the gender-gap in math performance. The replication was designed to be thorough, with consistent methodology across sites and a sample size large enough to detect even small effects.

Despite following these procedures to the letter, the replication found no effect. Women who were ostensibly in a threat condition didn’t perform any worse than those who were instead taught about threat. And the difference between men and women’s math performance remained consistent across the board, regardless of how the test was framed. The stereotype threat effect, once thought to be so robust, just wasn’t there.

What Does This Mean for Stereotype Threat?

Does one failed replication debunk the entire theory of stereotype threat? No, of course not. But it’s not just one study. There are now multiple failed replications, large-sample studies that found no effect, and at least one bias-corrected meta-analysis pointing to the same conclusion: if stereotype threat exists, it is far weaker and more inconsistent than we originally believed. I no longer believe it is real, but you can make up your own mind.

I have seen some people online suggest the reason this failed to replicate is that women are no longer stereotyped as not being good at math. While I do not disagree that cultural stereotypes about women in STEM might have changed since 2005 when the original paper was first published, I’m skeptical this is the main culprit behind this non-replication. First, women remain heavily outnumbered in STEM fields. The latest statistics indicate that women comprise only 25% of STEM workers in Canada and 27% in the US. And, depending on what is counted as STEM—I have heard some argue that psychology should be included—this number might be a lot lower. So, the stereotype about what is and what is not a female job might still be around, as much as we’d like it not to be.

Second, for years, many of us have suspected that something wasn’t right. There were warning signs: tiny sample sizes, flexible analyses, and implausibly large effect sizes given the relatively modest interventions being tested. In some cases, stereotype threat effects were found only in very specific handpicked samples—another red flag. It turns out that many of the original studies were conducted at a time when researchers—and I count myself here—were less stringent about methodological rigor.

Let’s be honest: that last sentence was far too generous. Many of us engaged in practices that, in hindsight, were borderline dishonest. We abused experimenter degrees of freedom, engaged in questionable research practices, p-hacked, massaged our data—you pick the euphemism. In contrast, this new replication study followed the most up-to-date best practices in psychological science, eliminating room for flexibility in analysis or results interpretation.

In my opinion—one that I have shared widely over the years—studies like this do more than demonstrate that stereotype threat is not replicable. They raise unsettling questions about the broader field of social psychology. If stereotype threat is not real, not robust, what else was I taught in my introduction to psychology classes that is also suspect? Despite all our improvements that help us in the present and future, we still have a massive backlog of studies from the past that we need to reckon with. Yes, the future looks bright, but we need to have the courage to put our most cherished findings under the spotlight.

The bill of reckoning for social psychology is past due.

A Reckoning…and a Path Forward

The fall of stereotype threat is not just about one theory collapsing: it’s a moment of reckoning for the entire field of social psychology. Stereotype threat was more than an idea — it was a promise, a way to understand inequality and to imagine solutions. Its failure forces us to confront uncomfortable questions about how science is done and what happens when beloved ideas turn out to be wrong.

But this reckoning, painful as it is, should not lead us to despair. The scientific process thrives on self-correction, on challenging old paradigms and building stronger ones in their place. What we’re experiencing now is science doing what it’s supposed to do: correcting itself. If we care about understanding the human mind and addressing real-world inequalities, we need to keep asking hard questions and demanding better evidence—not just for stereotype threat but for every cherished finding.

For me, letting go of stereotype threat has been both humbling and liberating. It has forced me to recalibrate how I think about research, advocacy, and the stories we tell about human potential. It’s a reminder that science, at its best, is about progress, not protecting idols.

We also had our own experience with Stereotype Threat research as graduate students and thought we’d share it for the first time:

Read more

You start to wonder: Am I out of touch? Is this what people really think?

Now imagine that wave wasn’t a wave of people at all.

That’s one of the central risks we outline in our new Science Policy Forum article on malicious AI swarms—coordinated fleets of AI agents that can imitate authentic social opinions and actions at scale.

Why is this dangerous for democracy? No democracy can guarantee perfect truth, but democratic deliberation depends on something more fragile: the independence of voices. The “wisdom of crowds” works only if the crowd is made of distinct individuals; when one actor can speak through thousands of masks—and create the illusion of grassroots agreement—that independence collapses into synthetic consensus.

This is no longer just theory. In July 2024, the U.S. Department of Justice announced it disrupted a Russia-linked, AI-enhanced bot operation involving 968 accounts impersonating Americans. And while estimates vary by platform and method, online discourse about major global events often involves substantial automated activity—on the order of ~20% in one global comparison.

At the same time, the market is beginning to industrialize coordinated synthetic activity: reporting in 2025 described Doublespeed, backed by Andreessen Horowitz, as marketing tools to orchestrate actions across thousands of social accounts, including attempts to mimic “natural” interaction via physical devices.

Concrete signs of industrialization are also emerging: the Vanderbilt Institute of National Security released an archive of documents describing “GoLaxy” as an AI-driven influence machine built around data harvesting, profiling, and AI personas for large-scale operations. And campaign-tech vendors market supporter-mobilization platforms and have begun promoting AI-powered analytics—an example of how “human-in-the-loop” coordination could be scaled even before fully autonomous swarms arrive.

Not all automation is malicious. But the capability jump due to developments in AI matters. What changes when automation can coordinate, infiltrate, adapt, and build trust over time?

What is a “malicious AI swarm”?

We use “malicious AI swarm” to describe a set of AI-controlled agents that can maintain persistent identities and memory, coordinate toward shared objectives while varying tone and content, adapt in real time to feedback, operate with minimal human oversight, and deploy across platforms.

That combination distinguishes swarms from yesterday’s botnets. Older systems often relied on rigid scripts, repetition, and obvious synchronization. AI Swarms can now generate unique, context-aware content while still moving together as a coherent organism. They are less like a megaphone and more like a coordinated social system—and may therefore be much harder to detect and prevent.

Several shifts make swarms especially potent. Coordination can become fluid and hive-like, rather than purely “central command.” Swarms can infiltrate communities by mapping network structure and local cues at scale. They can evade detectors tuned for copy-paste behavior by producing human-level variation. They can also self-optimize through rapid experimentation—testing alternatives, learning what works, and propagating winning strategies quickly. Most importantly, they can persist over long periods, embedding in communities for weeks, months, or longer and shaping discourse gradually.

Why this is dangerous for democracy (it’s not just “misinformation”)

Democracy doesn’t require perfect truth—but it does require something more fragile: independent voices. The “wisdom of crowds” depends on independence between judgments. If a single actor can speak through thousands of inauthentic accounts, the apparent consensus of the crowd stops being informative.

One pathway of harm is synthetic consensus: creating the illusion of a majority opinion. Swarms can seed narratives across niches and amplify them via coordinated liking, replying, and cross-posting until it looks like broad support. People update beliefs partly through social evidence—what seems normal, common, or widely endorsed. Synthetic consensus exploits that cognitive shortcut.

A second pathway is segmented realities. Because swarms can mimic local language, emotion, and identity cues, they can tailor narratives community-by-community, reinforcing polarization and making cross-group cooperation and consensus harder.

A third pathway is poisoning the information substrate for future AI—sometimes described as “LLM grooming.” A long-term strategy is to flood the web with machine-targeted content so future models ingest and reproduce distorted narratives.

A fourth is synthetic harassment: scaling intimidation and smear campaigns until people self-censor. Swarms can generate relentless, tailored abuse that looks like a spontaneous public pile-on, pushing journalists, academics, dissidents, and officials out of the public sphere before defenders can even classify the campaign.

Finally, there is epistemic vertigo: if everything might be fake, why trust anything? Ironically, awareness of manipulation can deepen cynicism. If people believe large portions of discourse are synthetic, trust collapses—and the public sphere can shrink into gated channels and closed groups.

So what do we do? Focus on coordination, provenance, and incentives

We argue for defenses that make manipulation costly and risky—without turning platforms into a “Ministry of Truth.” Instead of trying to adjudicate content (“who decides truth?”), we should prioritize coordination signals, provenance, and incentives.

One priority is always-on detection for statistically unlikely coordination. Rather than episodic cleanups after campaigns go viral, platforms (and regulators) should push for continuous monitoring for anomalous patterns—signals that genuine crowds struggle to reproduce consistently. To reduce misuse, these systems should be paired with transparency measures and independent audits.

A second priority is stress-testing defenses with simulations. Defenders will always lag if they only react to yesterday’s tactics. Agent-based simulations—running swarms in synthetic networks—can help test detectors, uncover failure modes, and evaluate interventions before real attackers exploit them.

A third priority is strengthening provenance without killing anonymity. Real-ID policies can harm dissidents and whistleblowers. The goal is “verified-yet-anonymous” options—strong proof signals that raise the cost of mass impersonation while preserving privacy where it matters.

A fourth priority is building an ecosystem for shared situational awareness—an “AI Influence Observatory” model where researchers, NGOs, and institutions standardize evidence and publish verified incident reporting without centralized censorship.

And finally, we have to change the economics. Even the best detectors will be imperfect. So we should also target the commercial market for manipulation: discount synthetic engagement, enforce no-revenue policies for manipulation campaigns, and publish audited bot-traffic metrics. The goal is to make large-scale manipulation less profitable to sustain.

Bottom line

The point isn’t that AI makes democracy impossible. The point is that democracy becomes brittle when it’s cheap to counterfeit social proof—when it costs little to run a fake crowd and minutes to manufacture “public opinion.”

The mission is straightforward: make large-scale impersonation and coordination harder to run, easier to detect, and less profitable to sustain. If we get that right, the public square does not need a central authority to decide what is true. It needs conditions where authentic human participation is visible—and where engineered consensus collapses the moment it tries to scale.

Key takeaways

• The next wave of influence operations may not look like obvious copy-paste bots. It may look like communities: thousands of AI personas with memory, social identities, distinct styles, and coordinated goals.

• The most dangerous outcome is not a single viral lie—it is synthetic consensus: the illusion that “everyone is saying this,” which can quietly bend beliefs and norms.

• This is already moving from theory to reality. In July 2024, the U.S. DOJ announced it disrupted a Russia-linked AI-enhanced bot operation involving nearly 1,000 accounts impersonating Americans.

• Defenses should not hinge on policing content. They should focus on coordination and provenance: detecting statistically unlikely patterns, stress-testing defenses with simulations, strengthening identity/proof signals, and shifting platform incentives.

This post was drafted by Daniel Thilo Schroeder and Jonas Kunst, with edits from Jay Van Bavel. You can read our full paper here:

• Daniel Thilo Schroeder et al. (2026). How malicious AI swarms can threaten democracy. Science, 391, 354-357. DOI:10.1126/science.adz1697

Thanks for reading The Power of Us! This post is public so feel free to share it.

Share

News and Updates

Check out our new Ask Me Anything sessions for the new year! Paid subscribers can join us for our monthly live Q&A with Jay or Dom where you can ask us anything from workshopping research questions, career advice to opinions and recommendations on pop culture happenings — for paid subscribers only. Upgrade your subscription using the button below. Invites to RSVP have been sent via email from powerofusbook@gmail.com

• February 20 @ 2:00 EST with Dom

• March 4th @ 2:00 EST with Jay

• April 17 @ 2:00 EST with Dom

• May 6th @ 2:00 EST with Jay

Catch up on the last one…

Last week, we discuss how aggressive policing can backfire—by fueling more intense protests.

After the invention of the printing press in the 1400s, the bestselling books were religious extremist texts and witch-hunting manuals. In other words, what went “viral” in the Middle Ages wasn’t so different from the salacious conspiracy theories you see flooding social media today.

While studies find that moral outrage and negativity about political opponents go “viral” on social media, this is also true of the offline world. Gossip is common in everyday conversation, mostly negative, and often about people we dislike.

Just as “cancellations” go viral on social media, gossip spread widely in hunter-gather societies, and was similarly centered around other people’s antisocial behavior.

In a new paper called “The Psychology of Virality,” recently published in Trends in Cognitive Sciences, we argue that, just like some viruses are more contagious than others, certain types of information (such as negativity or outrage) are naturally more contagious than others—going “viral” in all kinds of settings.

Using virality as a metaphor is controversial, but useful, because it can scaffold on our understanding of superspreaders (eg., some people spread more information that others), interventions (eg., how to stop the spread of certain types of information), environments (eg., why some platforms facilitate the spread of information), etc. See our figure for more examples.

Certain information goes viral because it appeals to our evolved psychology. For instance, negative and intense emotions are more likely to attract our attention and be remembered, which may explain their viral spread. This “negativity bias” likely evolved because it was important to pay attention to information that could threaten our survival.

Similarly, gossip is thought to have evolved to help small-scale societies cooperate, exposing immoral actors and incentivizing people to act cooperatively to avoid becoming its target. Technology may have changed dramatically across centuries, but the underlying psychology behind the information we pay attention to and share hasn’t.

Understanding the online world requires understanding this underlying psychology. While we often assume that the anonymity of social media makes us more hostile, one series of studies reveals that this assumption is incorrect or overstated: most people who are hostile online are also hostile offline. People are surprisingly similar in their online and offline behavior. Social media may simply feel more hostile because hostile actors are more visible.

Just as a few “superspreaders” disproportionately contribute to epidemic spread, only a small handful of “superspreaders” create most of the hostile and false information online. As an example, one estimate shows that around 0.1% of fake news is shared by 80% of Twitter users. Similarly, 3% of toxic users contribute to 33% of all comments on Reddit. This can create the perception that the world is far more hostile or misinformed than it really is.

These “superspreaders” of toxic content have vastly different tastes than most people. For instance, one study found that, while most people disapprove of divisive social media from politicians, the small handful of people who frequently engage with politicians actually enjoy divisive posts. This may explain in part why our research has found that the most widely shared content online is, paradoxically, not widely liked.

But there is a key difference between the online and offline world: while gossiping a little bit predicts more friendships, those who gossip the most offline have fewer friends. By contrast, extremists who frequently share divisive content online tend to have the most followers and highest engagement—in part because of recommendation algorithms designed to promote the most attention-grabbing content.

Outrage may have always gone viral. But just as viruses spread faster in a crowded room, social media is a context that allows superspreaders to spread hostile content faster and farther than they ever could have before.

So, what is the solution?

Not letting “virality” rule our lives. Our research has found that the vast majority of people across the political spectrum say that they want social media algorithms to promote positive, educational, and nuanced content. Yet these platforms promote toxic content that nonetheless grabs our attention. We can change social media algorithms so that they amplify what we want to see, rather than what we can’t look away from.

Just like we can slow the spread of viruses (through masks, distancing, vaccinations), we can also slow the spread of viral information. Doing so can have positive consequences: for example, WhatsApp found that putting limits on how much a message could be forwarded helped curb the spread of viral misinformation.

We live in a world where the main way we organize and filter our online information is via algorithms that show us the most viral posts. But, we know, from psychology and history, that the most viral content is often the most harmful. The printing press made information spread faster, but after this, people needed to develop institutions to help organize, filter, and make sense of the flood of information. Using “virality” as the main way to decide the information people see every day will (like actual viruses) make us sick.

News and Updates

Check out our last Ask Me Anything session for the fall! Paid Subscribers can join us for our monthly live Q&A with Jay or Dom where you can ask us anything from workshopping research questions, career advice to opinions and recommendations on pop culture happenings—for paid subscribers only. Upgrade your subscription using the button below. We will be posting more sessions in the new year.

• January 21st @ 2:00 EST with Jay

• March 4th @ 2:00 EST with Jay

• May 6th @ 2:00 EST with Jay

Catch up on the last one…

Check out where you score in intellectual humility in our scale posted last week. It turns out that intellectual humility is a great trait to have. People who score high on these measures are less likely to spread conspiracy theories or misinformation, express out-group prejudice, and feel polarized.

In the past decade, there has been an explosion of interest on the topic of intellectual humility(see the figure below). This had led to the development of at least 10 new scales to measure intellectual humility and hundreds of new studies on the topic. If you agree with statements like “I welcome different ways of thinking about important topics” and disagree with statements like “I believe my ideas are superior to others’ ideas” then you would be considered an intellectual humble individual.

It turns out that intellectual humility is a great trait to have. People who score high on these measures are less likely to spread conspiracy theories or misinformation, express out-group prejudice, and feel polarized. They are more likely to follow public health guidelines, reason carefully about political issues, and update their beliefs in light of novel information.

However, the variety of different measures makes it difficult to know which scale to use or how to interpret different results produced by competing scales. Thus, creates the worrisome possibility that scholars in this area are committing the jingle fallacy (using different labels for the same thing) or the jangle fallacy (using the same label for different things) in this research.

Therefore, Jay conducted a series of ten studies (including over 5,900 participants) with Philip Parnamets and Mark Alfano to generate a single unified measure of intellectual humility. We called it the “Collected Intergroup Intergroup Humility Scale”. You can download our paper here if you want to read the details from our research.

One key feature of our measure is that we created a subscale to measure reasoning in the context of intergroup conflict. Prior measured focused on individual reasoning in a social vaccuum or merely in the context of dyadic disagreements. This is a remarkable omission because the biggest challenge for many people is responding with humility in the face of criticism from out-group members. Therefore, we created a number of items to capture the ability to question or challenge their own group and remain open to ideas from outsiders

Overall, we found five key factors of intellectual humility:

1. Open-mindedness: being open to new perspectives and evidence, especially when they are not consistent with one’s existing ideas. People who score high on this scale agreed with items like “I like finding out new information that differs from what I already think is true”.

2. Intellectual defensiveness: reacting emotionally or negatively to disagreement and the possibility of one’s own cognitive limitations. People who score high on this scale disagree with items like “I feel uncomfortable when someone points out one of my intellectual shortcomings".

3. Intellectual arrogance: indifference to truth or the possibility of being wrong. People who score high on this scale disagree with items like “I don’t really enjoy gaining new knowledge”.

4. Sense of intellectual superiority: believing you or your group is smarter or better at reasoning than other individuals or groups. People who score high on this scale agree with items like “My ideas are usually better than other people’s ideas”.

5. Ingroup criticism: actively seeking out and correcting errors in reasoning by one’s in-group. People who score high on this scale agreed with items like “To avoid group-think, it’s important to be extra critical of your own group’s ideas”.

We include all the items below so you can measure your own intellectual humility of give it to your friends and colleagues!

Although these factors were all correlated with one another, they seemed to capture distinct sub-scales (see figure below). We also found that intelligence, numeracy, and reasoning skills were only weakly correlated with these five factors (all correlations were below r = .16). This suggests that intellectual humility is not simply reducible to basic cognitive abilities. We also suspect that intellectual humility might be easier to cultivate than basic intelligence, which is highly stable over time within individuals.

You can see how you score on each of these items to assess your own degree of intellectual humility. You are also free to share our measure with your friends, family and coworkers and discuss their scores. Have them answer each question below on a 7-point scale their level of agreement with each item, with 1 =“strongly disagree” to 7 = “strongly agree.”

Read more

In a new pre-print, we review how and why politics can shape health outcomes, especially in a highly polarized environment. We argue that people’s political identities ultimately shape the health decisions they make.

To explain why this happens, we answer two questions: What shapes peoples’ health decisions? And how does social identity affect these decisions?

What shapes people’s health decisions?

When we are deciding whether to get vaccinated, start a new workout routine, or quit smoking, what determines whether we take the plunge or hold back? Psychologists argue that these decisions are based on our beliefs about the behavior. In particular, we can distinguish three kinds of relevant beliefs.

First, we can have behavioral beliefs about different properties of the behavior itself. This includes beliefs like “vaccinations will reduce my risk of getting sick” and “this workout routine will help me reach my fitness goals.” But it also includes beliefs like “vaccines can hurt a bit” and “this workout routine requires an expensive gym membership.” Overall, these beliefs shape how desirable or undesirable the behavior is to us personally.

But our individual beliefs aren’t the only thing that matters. After all, we don’t just make decisions as individuals. As social creatures, our behaviors are also shaped by normative beliefs about what others around us are doing. This can include beliefs like “nobody I know is getting vaccinated” and “everybody on my college campus is binge drinking.” So, the same exact behavior can be more or less compelling to us based on what we see other people doing or what they expect us to do.

Finally, we also have beliefs about our ability to do the behavior. If you’ve never stepped foot in a gym, you may have no idea where to start, and it may strike you as an overwhelming feat. If you live in a food desert, far from supermarkets or grocery stores, you may accurately believe that access to affordable vegetables are difficult to get. These sorts of beliefs also play an important role in determining whether our planned behaviors become a reality.

With all these different beliefs buzzing around in our heads, how do we use them to take (or avoid) action? Each of our beliefs has a weight—our sense of how important that belief is relative to others. These weights help determine whether the behavior is good enough, on average, to move us to take an action.

The amount of weight we give each belief varies from person to person and from belief to belief. For one person, the social stigma of being the only family member to wear a mask at a gathering might outweigh their desire to minimize risk of catching COVID or the flu. For another person, the importance of these different beliefs might be flipped.

Of course, people aren’t always explicitly tallying up the pros and cons of every decision to themselves in their head. These processes can go on in the back of our minds, without us having to consciously think about them. The key takeaway is that in order to understand the health decisions a person makes, we should know what beliefs may make them want or not want to take the behavior, and how important each of these beliefs are to that person.

How does social identity affect health decisions?

Now that we know a bit about how people make health decisions, we can ask how social identities—ranging from membership in a political party, to a run club, to being a university student—affect these decisions. We suggest that there are two main ways this can happen.

First, our social identities can affect the beliefs we hold about a behavior. This can be seen most strikingly in the way political identities affect beliefs. During the COVID-19 pandemic, Republicans and Democrats developed starkly different beliefs about the health complications arising from COVID and the risks of the vaccine. One explanation for this is that Republicans and Democrats inhabited different information environments, looking to different partisan news sources and following different influencers on social media. Because they saw different information, they developed different views.

Another explanation is that members of different parties have different motivations. It can be a discomforting experience to disagree with members of a social group you are a part of. As a result, partisans may have been motivated to “toe the party line” and view COVID-19 and its vaccine the same way other members of their party did.

Second, social identities can affect how important different beliefs are to us. If you are a college student feeling strongly attached to your peers, you may be more likely to value fitting in and binge drinking when they do. If you join and stay in an anti-vaccine community online, you may adopt the values of your community and start to take the demonstrated health benefits of vaccinations less seriously.

Each of these two processes—changing beliefs and changing their weight, or importance—can happen independently. Together, they make for powerful, mutually reinforcing effects of our social identities on the kinds of health behaviors we do and do not choose to do.

Encouraging people to take the right actions to improve their health will require attention to beliefs, social norms, and feelings of self-efficacy. If all of these things are aligned—people have accurate beliefs about an effective action, their group members are already taking this action, and they feel capable of doing it—then it is far more likely they will act in a way that impacts their well being.

The kinds of health behaviors we choose are consequential—not just for ourselves, but also for those around us. In order to understand how people make these decisions, it is important to consider the social context they are in. We need to understand the kinds of social groups people identify with, and how these groups shape their beliefs and values.

By understanding this aspect of group psychology, we hope that medical experts, public health professionals, and policymakers will help provide better solutions to keeping us all healthy.

This newsletter was written by Raunak Pillai and edited by Hannah and Jay. It summarizes a recent paper from our lab which you can read here. We welcome any suggestions.

News and Updates

Jay is giving a public talk at NYU on “The Power of Us: How Shared Identities Shape Conflict and Cooperation” at NYU on December 4th at 5pm. RSVP here if you want to attend. Here is the abstract:

”What does it mean to belong — and how does our sense of identity unite or divide us? Group cooperation is one of the most remarkable features of human nature, enabling people to build teams, communities, and nations. Yet the same psychological forces that bind us together can also drive conflict and polarization. In this talk, I explain how our social identities — the groups we feel part of — shape our thoughts, emotions, and actions. Drawing on insights from social psychology and cognitive science, this talk reveals how understanding identity can help us understand and address some of today’s greatest challenges, including political division, misinformation, climate change, and threats to democracy. Ultimately, The Power of Us shows how we can harness our shared identities to foster cooperation, strengthen communities, and build a more connected world.”

Check out our last Ask Me Anything session for the fall! Paid Subscribers can join us for our monthly live Q&A with Jay or Dom where you can ask us anything from workshopping research questions, career advice to opinions and recommendations on pop culture happenings—for paid subscribers only. Upgrade your subscription using the button below. We will be posting more sessions in the new year.

• December 11th @ 4:00 EST with Dom

• January 21st @ 2:00 EST with Jay

• March 4th @ 2:00 EST with Jay

• May 6th @ 2:00 EST with Jay

Catch up on the last one…

Our latest paper explains why people tune their moral behavior to fit the social context. Building on the Theory of Planned Behavior, we illustrate how moral action depends on three factors: attitudes (what we believe is right and what we expect moral action will lead to), perceived control (whether we feel we can act effectively), and situational norms (what feels acceptable in that moment). The situational norms are especially influential because they signal which behaviors will be rewarded or punished, which helps explain why the s…

Read more

Jareef is a postdoc at the Department of Strategy and Management. Together with Jay Van Bavel from New York University and professors Helge Torbjørnsen and Hallgeir Sjåstad (Department of Strategy and Management), they have a new paper on group identities and dishonest behavior.

The researchers tested whether the tendency to behave dishonestly is influenced by the group identity of either the victim or the beneficiary of the cheating. This was studied in three experiments with more than 5,230 Americans in total, and with real money.

Is there a difference if the victim of your cheating is part of our “in-group” or is in the “out-group”?

Surprisingly not. For purely selfish dishonesty, people lied equally regardless of whether it came at a cost to an in-gr…

Read more

This tendency to avoid out-group sources narrows our perspectives, deepens political polarization, and can undermine trust in democratic institutions. This is part of a broader trend of political sectarianism, where the other side is seen not just as wrong, but as evil (as shown in the figure below). So, what can be done?

Our new research finds evidence for a fresh solution: Artificial Intelligence (AI) can help people bypass these deep-seated partisan biases

Nearly a billion people now use AI chatbots. And as tools like ChatGPT become more common, they are transforming how we access information. We reasoned that beca…

Read more

In the last few years, we have seen headlines like “Why Kids Aren’t Falling in Love With Reading”, “Fewer People Are Reading for Fun”, and “Books vs. Brain Rot: why it's so hard to read” dominate the news. These have cast a chill over the publishing industry and led bookworms like us to panic about the future of humanity.

Despite the gloom, booksellers in the United States saw the first annual increase of print book sales in three years in 2024. It was a reminder that even in our increasingly digital age, books remain a powerful, and popular, force—not just for entertainment, but for personal growth. Maybe the media was too fast to write the obituary about books.

Social psychologist Rémi Thériault, wh…

Read more

The Many Faces of National Pride

What it means to be proud of one’s country depends on how we think about our national identity. And national identity, as it turns out, can take many forms. One key distinction lies between healthy, secure national identity—rooted in civic values, solidarity, and genuine care for one’s fellow citizens—and a defensive, superficial one, more concerned with appearances and external recognition than with what the country stands for or how its people are actually doing. A prominent example of the latter is called national narcissism.

In Caravaggio’s famous painting of Narcissus, a hand…

Read more

What does your merch say about you?

Last week, video essayist and writer Mina Le of High Brow published a thoughtful (viral) video essay on Youtube in which she discussed the history of merch, its evolution, and how it signals different identities and statuses. She explains that the purpose of merch is to promote something—a band, an organization, etc—but it can also signal a person’s values, identities, and interests. Merch can bring people together (or drive people apart) because it visually signals our social identification to the world.

But Mina claims that the merch industry has gone too far, and it…

Read more