The costs of being open-minded

Reasonable People #76: You probably think open-mindedness is a virtue. Be careful. Others may be willing to punish you for it.

You probably think open-mindedness is a virtue. You say you want to see more of it in others. You probably think of yourself as open-minded.

Be careful. Others may be willing to punish you for it.

Research by Hussein and Wheeler (2024) shows that the high esteem of open-mindedness might only hold in the abstract. In their experiments, they presented participants - US adults - with individuals who took concrete steps to learn about an opposing point of view. In most conditions they tested, they showed that these actions caused the person taking them to be viewed less favourably and more negatively.

In effect, participants wanted to punish those they saw as being receptive to alternative points of view. That instinct to punish could be triggered by something as innocuous as following the wrong social media account or reading the wrong article online.

Let’s look at the details, since understanding what they did will both help us gauge how strong and general this effect is, set us up for a check on the overall credibility of the finding, and help us understand the one condition where seeking out opposing views didn’t trigger a backlash.

What’s the mystery?

Hussein and Wheeler introduce their studies with a conundrum. Lots of research has shown that open-mindedness is viewed positively, with people saying they view the open-minded as more trustworthy, more intelligent and more persuasive - the kind of people, in short, that we all would like to work with, or to be.

At the same time, other research shows that people can find people of opposing views morally repugnant, especially when different positions have become polarised (Hussein and Wheeler take for granted the US context, of polarisation between Republicans and Democrats, which may be fair enough given they are both based in America).

This presents a paradox: how can listening to opposing views be a Good Thing, but holding opposing views a Bad Thing? The resolution, say Hussein and Wheeler, is that previous research focussed on the views - contrary information about your current beliefs - and not on the source - who it was that was sharing that information. Make the source salient and, as we’ll see, the abstract approval for open-mindedness turns into concrete disapproval.

11 studies

The report is of a series of studies - 6 in the main paper, 5 in the supplemental material. Each study is an evolution or variation on the previous one, testing some new aspect of the fundamental result (which they consistently find).

The first study shows the basic set up.

Participants - US adults - are recruited online and told a story about John, a member of the same political party as them (i.e. it was adjusted depending on the party the participant declared at the start).

Next, participants were “asked to imagine that while talking to John, the topic of politics and social media came up. John mentioned that Twitter (a social media platform) recently recommended a new Twitter account for him to follow, and that the recommended account belonged to an out-party member.”

Then they were shown the twitter account that was recommended to John. This was a profile picture and bio, and then recent posts. For this first study they made the profile highly stereotypical. You can see the full thing in the paper, but here’s a detail which shows just how stereotypical. I’ll leave it to you to figure out which post is from the stereotypical Democrat and which is from the stereotypical Republican:

Then they are told that John either said: “I started following him. I wanted to listen to and engage with this perspective,” OR he said “I blocked him. I did not want to listen to or engage with this perspective.”

Finally, they asked participants how they viewed John - favourably or unfavourably, and positively or negatively (on seven point scales).

The main outcome is the difference in this evaluation between the two conditions. Participants who heard John blocked the social media profile from the out-party member rated John much more favourably.

Additionally, the extent to which an individual participant viewed members of the opposing party as immoral predicted the favourability ratings, suggesting that it was this moral dimension which drove negative evaluations when John decided to find out more about someone from the other side.

This basic pattern held, across all the subsequent studies. The next study toned down the recommended profile, making it more prototypical rather than stereotypical (they had a nice method for generating judging what supports of each party thought was fairly representative). The effect was smaller, but still robust.

They showed the effect worked for deciding to attend (or not) a rally by a politician from an opposing party, rather than deciding to follow (or not) a random social media profile. They showed the effect with celebrity politicians (like Obama and Trump), with different event formats (a lecture or a rally), and across a range of issues (climate change, immigration, terrorism, military spending, boycotting the Winter Olympics in China, and regulating social media companies).

Finally, adding a note of hope to what is otherwise a depressing story, they showed that the normal effect was reversed in very specific conditions - when the out-party source was not typical of that party. Here’s part of the stimuli they used to present a high and low typicality Democrat (only):

In this final experiment, the results showed that John’s receptiveness (his decision to follow rather than block) led to more favourable evaluations - a reverse of the finding in the previous experiments.

This is shown in the results figure, right side :

Despite the reversal of the effect, the authors report a common, core, mechanism: perceptions of (im)morality. Participants didn’t view the low prototypicality party member as immoral like they did the prototypical party member. In the absence of moral condemnation, receptiveness really was a virtue.

Credibility Check: Should we believe it?

Not all research reports are reliable. What to make of this one?

The design is strong: it’s a series of studies, it tests variations and discovers a boundary condition (low prototypicallity sources), with a plausible mechanism (perceptions of morality). It has various credibility signals; some of the studies are preregistered, and all studies share the raw data, the materials and analysis code. That doesn’t mean there’s nothing wrong with the studies, but opening up your factory for anyone to review how the sausage is made is a strong sign you’ve got nothing to hide.

Because of this, my view is that the limitations of the study are about the interpretation, not about the robustness of the reported effects.

My main hesitation is that all studies used the same outcome measure - simple survey measures of favourable-unfavourable and positive-negative views. What people say they feel may not reflect what they really feel, or translate across to other behaviours.

We could worry that the participants are all recruited online. The experiments actually used three different online recruitment platforms (Prolific, Lucid and AMT), which might mitigate the peculiarities of each, but still leaves the general problem that online crowd workers can be more involved in earning money as quickly as possible rather than sincerely engaging with your experiment.

A related issue is that of Demand Effects. Maybe participants said that they disapproved of John when he was receptive because that seemed like the right kind of thing to say, given what the researchers asked and the situation seemed to suggest?

For future studies it would be nice to see the effect demonstrated for “thicker” real world behaviours, including things which actually cost the participant some effort. It is easy to say you disapprove of someone, but how do we really know what that means until you decide to do (or not do something) that costs you time effort or opportunity?

There are always further questions. Overall, I take this report as good evidence that the phenomenon it is reporting is real.

(for more along the lines of how to assess the credibility of studies, please see my experiment design checklist ).

What it means.

Overall, these are well crafted studies which credibly describe an interesting result: open-mindedness can be punished when it is directed at the “wrong sort” of people. The studies also suggest the mechanisms behind this: it occurs in polarised environments when we regard the prototypical outgroup member not just as wrong, but as dangerously immoral (such as the contemporary US).

The result matters. As the researchers put it “A well-functioning democracy demands from its citizens a willingness to engage with ideas and people they disagree with”. If we impose costs on those who listen to the other side - even if it is as small as giving them the cold shoulder - it will reinforce partisanship and reduce our collective intelligence.

The reaction to people’s information consumption habits also suggest another way in which our current moment is strongly flavoured by the world of online social media. Hussein and Wheeler: “Historically, other people’s information consumption was inconspicuous; with the proliferation of social media, however, information consumption is increasingly salient and easy to observe”, with their findings raising the possibility “that group members are motivated to observe and judge the information consumption of other group members”.

Individual freedom of thought and conscience has long been a cornerstone of liberal democracy. But thought has never entirely been a property of individuals. Our thinking increasingly happens in public, with the ideas we are exposed to open to audit by others in our social network, exposing this surface to greater risk of moralising censure.

What to do about it?

The authors thoroughly report their results, and their reasons for thinking they are robust. Perhaps rightly, they have less to say about what to do about them. If polarisation is a problem, and receptiveness is a true virtue, then it is bad if those who might listen to the other side get viewed negatively.

One chink of light is suggested by the final experiment, in which listening to an outgroup member is viewed positively if they are perceived as non-typical.

For our own behaviour, this suggests there is a benefit in cultivating non-typicality. If our views are diverse, not just borrowed wholesale from a party, then we can’t but help than be less typical than the stereotype, and so others may be more inclined to listen to us.

Despite the label, I actually think being non-typical may be more common that it appears. The first experiments in this study pick stereotypical or prototypical members of each party, or politicians (who have to represent the party). Imagine the situation of the participants for a moment. Almost all they know about the fictional source of information is their party membership. If all you know about someone is that they are a Democrat or Republican, then it is natural to mentally fill in the rest with stereotypes. In reality, the more you find out about someone the more aspects you will find which don’t fit a stereotype. The Democrat or Republican voter you speak to may also be an aunt, or a colleague, or a sailing enthusiast. These extra details will reduce the salience of their party membership, so make it more acceptable to listen to them.

The tragedy of social media is that it restricts the information you have about people to a few lines, and then encourages us all to play-act the most easily-typecast version of ourselves for clicks and engagement from the in-crowd.

So, maybe, however robust this result, it may weaken in more complex situations, especially those offline and where more is known about people. Certainly, one counter-measure we could all take is to find out more about people, and to interact with people in multiple contexts, before making judgements about whether to listen to them or not.

Fundamentally, we need to be aware of our shared and deep attraction to tribalism. It’s a facet of human nature that isn’t going to go away, but perhaps a little more faith in the reasonableness of others is warranted, a bit more trust that our fellow citizens need to hear views we disagree with and make up their own minds.

Reference

Hussein, M. A., & Wheeler, S. C. (2024). Reputational costs of receptiveness: When and why being receptive to opposing political views backfires. Journal of Experimental Psychology: General, 153(6), 1425–1448. https://doi.org/10.1037/xge0001579

More, the Opinion Science podcast: Showing Open-Mindedness with Mohamed Hussein (5 May 2025)

---

I’m writing more for Reasonable People while on my career break. Upgrading to a paid subscription will encourage this

Keep reading for more on what I’ve been reading on polarisation, research robustness and persuasion.

---

Elite radicalisation theory of social media

Dan Williams writes The Case Against Social Media is Weaker Than You Think. I have a lot of sympathy for this view. Just because it feels like social media is eroding democracy doesn’t mean it is, and getting it right about where it is, and how it is, is vital if you are going to plan to improve things.

Nathan Witkin responds with The Case Against Social Media is Stronger Than You Think, in which he puts forward his ‘Elite radicalisation theory of social media’

Witkin argues for a more holistic view, agreeing that it is hard to show social media’s influence on the general population, but the effects on what we think of each other, and how political actors behave are clearer.

By incentivizing the creation of disproportionately negative and sensational content, and by in turn inducing more extreme, even violent political behavior, social media has almost certainly played a major role in the destabilizing political ructions of the last fifteen years—in particular in the U.S., but probably across Europe as well.

and

The result is that Americans are coming to see each other as much more politically extreme than they in fact are—or at least were. In the words of political scientist Henry Farrell, Americans have internalized a “malformed collective understanding,” in this case a vision of their collective identity that casts them as especially angry, pessimistic, and tribal.

Casey Newton picks up these threads in Three ways of thinking about the Kirk assassination

There is no justification for violence against someone based on their speech. But if Witkin's analysis holds, it suggests that there is real danger in constantly whipping up large crowds into a fearful frenzy. Spend enough time telling a mass audience that they are locked in a life-or-death battle with their political enemies and the eventually the more disturbed among them are going to take action.

The leitmotif of this newsletter is an optimistic ‘we’re more reasonable that we think we are, and this matters’, but I guess this has a corresponding pessimism: ‘if we believe we’re less reasonable, it makes us worse.’

TALK: How common knowledge shapes the world

Highly related: Steven Pinker is lucid on the importance of “knowing that they know that I know” for many forms of social coordination (and of not knowing that for some others).

The link to the claims above is this: it is hard to show that social media changes what we know - direct effects of misinformation are weak - but it certainly changes what we know we know - the effects are on the ‘democratic publics’, in Henry Farrell’s phrasing.

Link: How common knowledge shapes the world (11 August 2025)

He’s got a book on the topic, I see: When Everyone Knows That Everyone Knows...: Common Knowledge and the Mysteries of Money, Power, and Everyday Life.

I wrote about some similar themes in RP#31: Common Knowledge

Lionel Page: Epstein files: how arguments really make people change political side

Lionel Page argues that reputation is an overlay on the collective game of rationality. The bad news is that this can make people look like they reject good arguments due to coalitional loyalty. The good, argues Page, is that this means good arguments have long term power, because being reasonable ultimately feeds into reputation/credibility. He used the case of the Epstein files, and different reactions within the MAGA coalition, to illustrate his case:

Link: Epstein files: how arguments really make people change political side (subhead: The not totally ineffective nature of political arguments)

TALK: Dorothy Bishop-How negligent publishers subvert quality control in scientific publishing

From the PCI series, which is fantastic if you are interested in scholarly publishing.

Link: Dorothy Bishop-How negligent publishers subvert quality control in scientific publishing

PODCAST: The Data Sleuth Taking on Shoddy Science

Uri Simonsohn is co-author of the seminal False-positive psychology: undisclosed flexibility in data collection and analysis allows presenting anything as significant (2011). In this podcast he discusses work on detecting fraud in research papers, including the work that got him and collaborators sued for $25 million (unsuccessfully). The podcast includes this line about the pressures of the tenure system on researchers which tells you most of what you need to know about research fraud in a few words: “If you commit fraud, you may be fired, but if you don't commit fraud, you will be fired”

Link: The Data Sleuth Taking on Shoddy Science (2 August 2025)

…And finally

Church on stilts!

via @oldenoughtosay.com

It’s All Hallows Staining tower, built 1320. They are building a massive tower block next door, and a basement underneath. You can read about the site development here.

END

Comments? Feedback? Suggested punishments for my receptiveness to ideas? I am tom@idiolect.org.uk and on Mastodon at @tomstafford@mastodon.online