What’s Behind the Rush to Join an Internet Pile-on?

Joining a social-media pile-on, calling for someone’s firing in an online petition: these are ways to signal to others just where we stand on social issues and win plaudits from like-minded peers.

But this kind of virtue-signaling has come under scrutiny, thanks in part to high-profile instances where people have rushed to condemn, only to later backtrack as additional facts trickle in.

Consider the Harvard professors who expressed dismay at their university’s decision to sanction a colleague for sexual harassment, only for the majority to later retract their words, admitting that they “were lacking full details about the case.” Or recall the viral video of the white Kentucky teenager who was widely derided for mocking an Indigenous elder, only for a longer video to ultimately reveal a more-nuanced situation. “I shared the clip, uncritically, on my own Twitter feed,” lamented one journalist. “We were wrong to jump the gun.”

Instances like these got Kellogg management and organizations professor Nour Kteily wondering about the psychology behind the impulse to condemn quickly. “Are people evaluating the evidence? Are they actually looking at the full picture?” he asked. “Or are people choosing to socially punish without considering all of the evidence simply because they think it looks good to do so?”

Along with a colleague, Jillian Jordan, formerly a postdoctoral fellow at Kellogg’s Dispute Resolution Research Center and now an assistant professor at Harvard Business School, Kteily set out to investigate the propensity to punish before we’ve taken the time to consider opposing perspectives—what the researchers call “punishment without looking.”

Kteily and Jordan wanted to understand just how common this phenomenon is. They also wanted to know: Do we fail to seek opposing perspectives simply because it takes extra work? Or could skipping this step even be seen as preferable—perhaps because it signals the strength of our moral position? “If I chose to punish without looking,” reasoned Kteily, “that might make me seem especially loyal.”

The researchers find that, on balance, there is no reputational reward for failing to seek out alternative perspectives. This should come as a relief for anyone hoping to live in a world with fewer reflexive pile-ons. But it’s not all good news, either, because there is a reputational reward for punishing, and some of the punishers aren’t going to bother looking.

Investigating punishment without looking

To investigate punishment without looking, the researchers designed a clever online game with two roles. One group of participants played the role of “actor”: they read a public petition chosen to resonate with their political views (e.g., Democrats read a petition calling for the firing of a police chief who made offensive comments about the progressive Black Lives Matter movement). Then they were given a choice whether to sign it. But first, actors had the opportunity to explore opposing perspectives by reading articles that defended the “other side” or by searching the web for countervailing evidence.

The second group of participants, who shared the actor’s political affiliation, played the role of “evaluator.” They observed at least some of the actor’s behavior and then chose whether to reward it with a small sum of money. Evaluators also used a 0–100 scale to rate actors on their fairness, competence, loyalty to the cause, and overall positive impression.

The reputational rewards of punishment

Researchers used a version of this online game across four experiments.

In the first two, they focused on the evaluators. How would they perceive individuals who chose to sign the petition (or not)—and would that depend on whether they took the time to consider opposing views before making their choice?

The researchers found that evaluators did reward actors’ decision to punish. “Actors who chose to sign were financially rewarded for having done so,” says Kteily.

But critically, they found no evidence that evaluators preferred actors who didn’t look. While evaluators rated non-looking actors as more loyal to the cause, this was “counteracted by an even greater hit to perceptions of the actor’s competence and fairness,” says Kteily. Overall, actors who looked at countervailing evidence before signing the petition were rewarded more generously by evaluators than those who did not.

This preference for deliberation held even in scenarios when evaluators might value loyalty more than usual—such as when a Democratic actor has a family member in the police force, casting doubt on their loyalty to a cause such Black Lives Matter.

Putting on a show

For the final two experiments, the researchers turned their focus to the actors themselves. Just how much did their decision to punish hinge on the fact that they were being observed? And what about their decision to seek out additional evidence?

Across both experiments, researchers varied how much of the actor’s behavior would be scrutinized by evaluators. For some actors, both punishing and looking were observable; for others, just punishing—and for still others, neither—was observable. The actors were also told that their evaluators were either highly or only moderately ideological.

“People are sensitive to their reputation,” says Kteily. For that reason, you’d expect actors to be more likely to punish when they believe they are being judged by an ideological audience.

This is exactly what the researchers found. Democratic actors were more likely to sign the petition when their decision to punish was being scrutinized by a highly ideological evaluator (30 percent) versus not scrutinized at all (19 percent), with similar but weaker results for scrutiny from a less-ideological evaluator. Republican actors demonstrated a similar pattern when they knew they were being scrutinized by Republican evaluators.

In other words: the pressure to punish is real. “We were actually able to influence real behavior about a real petition,” says Kteily.

But when actors knew that their search for countervailing information would be observable, they were more likely to “look” than when only their decision to punish was observable.

This suggests that actors aren’t purposefully avoiding alternative perspectives to impress a like-minded audience.

Still, when punishment was visible but looking was not, many participants did punish without looking—not to demonstrate loyalty but to reap the social rewards of punishing without doing any extra work.

This pattern of results is particularly concerning for our online ecosystems, Kteily says.

“On most social-media platforms, our audience only learns whether we punished. The process by which we come to that decision generally is not disclosed. As a result, we miss out on the opportunity to leverage the reputational pressure to ‘look.’ Instead, we end up with people who feel pressured to punish, many of whom don’t bother to look, increasing overall rates of punishment without looking.”

A way forward

Given this new understanding of why we rush to judgment, can anything be done to thwart the cycle of online outrage?

One idea would be to put more of a reputational emphasis on consideration. “Imagine a world where, when you signed a petition, you had to give your reasons,” says Kteily.

It sounds far-fetched, but other mechanisms designed to encourage deliberation have already been tested on online platforms. Kteily points to one tested on Twitter (now known as X). Before posting a link to an article that users had not themselves accessed via the platform, they would receive a quick prompt: “Hey, did you read the article? Do you still want to post it?” says Kteily.

Adding a mechanism that would make a user’s consideration or fact-finding more visible to the public would not be easy, he says, but if we can pull it off, there’s reason to believe it could help.

“The very act of asking people to give their reasoning, or to put the best version of the other side’s perspective into sharper relief, could start to separate the wheat from the chaff,” says Kteily. “If I struggle to come up with reasons for the other side, then that’s a signal I haven’t really engaged with the other side at all.”