The psychology of public shaming, relational aggression, and collective punishment

Online communities are built on trust, shared interests, and cooperation. At their best, they allow people with different backgrounds and perspectives to work together toward something meaningful.

The same relationships can also become channels for exclusion, reputational damage, and collective hostility. Sometimes someone deliberately encourages this. Sometimes nobody is coordinating anything.

A community can collectively harm an individual even when every participant believes they are acting morally. Understanding how that happens is essential to building healthy communities.

How does a community cross the line from legitimate accountability into collective punishment?

1. From disagreement to moral judgment

Consider a hypothetical situation. A community member makes an inappropriate joke. Someone objects. Others join the discussion.

Initially, the disagreement concerns the joke: was it offensive, what did the author intend, and what should happen next?

But the discussion can change. Instead of evaluating the statement, participants begin evaluating the person. “Was that comment wrong?” becomes “What kind of person would say something like that?”

A single action becomes evidence of character. Ambiguity is interpreted negatively. Previous disagreements are reexamined through the same lens. The person’s explanation becomes another test of their character, rather than information that might change the group’s understanding.

Jon Ronson’s So You’ve Been Publicly Shamed explores the human consequences of public condemnation. His reporting asks us to consider the person experiencing a shaming, including the consequences that continue after the audience has moved on.

The danger is that resolving a specific problem gives way to defining a person by their perceived moral shortcomings. A community can lose sight of what would actually repair the original harm.

2. Relational aggression: using relationships as weapons

Psychologists use relational aggression to describe attempts to harm someone through their relationships or social standing. A foundational study by Nicki Crick and Jennifer Grotpeter examined this as a distinct form of aggression among schoolchildren. Applying the concept to adult online communities requires care, but it directs attention to harm that does not depend on overt insults or threats.

In an online community, this might take the form of spreading damaging rumors, recruiting allies to isolate someone, or threatening social consequences for people who continue to associate with them.

These actions can be consequential where reputation determines access, influence, and opportunities. The same relationships may connect someone to friends, collaborators, and paid work. Losing standing in one group can affect several parts of their life at once.

Exclusion alone does not establish aggression. People can end relationships, refuse collaboration, or remove someone from a space to protect themselves. The question is whether relationships are being used to harm or control someone, and what evidence supports that interpretation.

Is relational aggression more common among females?

Some early findings helped establish that association. Crick and Grotpeter’s 1995 study of 491 schoolchildren found higher relational aggression among girls. But the broader evidence is more mixed.

John Archer’s broader review found higher female indirect aggression only in later childhood and adolescence, with results varying by measurement method. That does not establish that adult females generally engage in it more than adult males.

Adult experiments have also found sex differences in particular forms of exclusion. Benenson and colleagues found that females formed more exclusionary alliances in competitive games, including when facing a threat of exclusion themselves. Other adult research finds no overall difference in relational aggression, with differences instead depending on whether aggression concerns peers or romantic partners.

3. The collective-action paradox

Collective hostility does not require a malicious organizer.

Imagine ten people who each believe they are defending their community. One publishes a critical comment. Another repeats an accusation. A third questions why anyone would defend the target. A fourth withdraws support.

Each person sees their own contribution. The target experiences the combined effect: repeated confrontation, damaged relationships, and pressure to leave.

A group can produce an outcome that few participants would consciously choose on their own.

Research offers evidence for one mechanism that can amplify this process. In a 2021 study, William Brady and colleagues found that positive feedback for outrage on Twitter predicted more subsequent outrage expression. Experiments in simulated social networks also showed that feedback and observed group norms could shape expression. The study measured expressed outrage, rather than proving what people privately felt or whether their complaints were justified.

My concern for communities is what happens when condemnation earns approval while asking for restraint becomes socially costly. Defending fair treatment may come to be interpreted as endorsing the original wrongdoing.

Alice Marwick’s research on morally motivated networked harassment describes how accusations of violating a network’s norms can mobilize harassment and reinforce group membership. Her interview-based account helps explain a pattern; it does not establish the motives of everyone in any particular dispute.

The resulting hostility can sustain itself through social rewards, without anyone directing the whole process.

Clay Shirky’s 2003 talk A Group Is Its Own Worst Enemy offers another lens. Drawing on Wilfred Bion, he describes how shared hostility can displace a group’s original purpose:

Nothing causes a group to galvanize like an external enemy.

My concern here is what happens when a member is recast as that enemy. Condemnation can become a way to demonstrate belonging. The group may feel unusually united while losing its ability to examine the accusation. That feeling of unity is not evidence that its judgment is sound.

4. When accountability becomes collective punishment

Communities need standards. People should be able to challenge harmful conduct, report misconduct, and establish boundaries. Public criticism can be necessary when those with authority have ignored private reports.

Accountability can include sanctions. What matters is whether the response remains connected to specific conduct, supported by evidence, and proportionate to the harm and ongoing risk.

A formal restriction has a defined scope. A public pile-on may have no agreed limit, decision maker, or stopping point. By collective punishment here, I mean that accumulating social penalty: many people imposing consequences whose combined severity nobody takes responsibility for.

Warning signs include:

  • Criticism shifting from conduct to personal worth.
  • Accusations growing more severe without corresponding new evidence.
  • Pressure against anyone requesting clarification or a fair process.
  • Threats or retaliation against people who refuse to join the hostility, support the targeted person, or continue associating with them.
  • Treating every apology, correction, or silence as further evidence of guilt.
  • Reputational punishment continuing indefinitely without a stated purpose.

An apology does not automatically resolve misconduct, and repeated harm can justify lasting consequences. But an opportunity to respond must be meaningful. If apologies, explanations, and silence are all treated as proof of guilt, the process offers no real chance to correct a mistaken judgment.

5. The problem of moral certainty

Moral conviction motivates people to oppose injustice and protect others. It can also make disagreement harder to resolve.

In research by Linda Skitka, Christopher Bauman, and Edward Sargis, stronger moral conviction was associated with greater distance from people holding different views and greater difficulty finding procedural solutions to disagreements, even after accounting for attitude strength.

In a community conflict, the practical risk is that commitment to a good cause begins to substitute for examination of the methods used in its name. Requesting evidence may be read as hostility toward a complainant. Questioning a sanction may look like indifference to the original harm.

Agreement can then become the safest public position. Silence looks like consensus, even when some members privately doubt the group’s response.

A healthy community needs room for someone to oppose harmful conduct while also opposing unfair treatment of the person accused. That position should not require a declaration of loyalty to either side.

6. Manipulation and spontaneous group dynamics

Sometimes people deliberately present misleading information, recruit allies, or exploit a group’s values to damage another person’s reputation. Pieter Hintjens’s The Psychopath Code offers a personal framework for thinking about manipulation and abusive relationships. It is useful here as a prompt to examine behavior, rather than as a clinical guide or a basis for diagnosing community members.

Similar outcomes can emerge without a deliberate plan. People may repeat an accusation because they trust its source, join criticism because others seem certain, or withdraw support because they fear becoming targets themselves.

The distinction matters. Deliberate manipulation involves strategic influence, often through deception. An escalating group response may emerge from many separate decisions. The two can also interact: someone can exploit hostility they did not create.

Reputational damage alone does not prove a coordinated campaign. Accusations of manipulation can themselves become a way to dismiss legitimate criticism. We should examine observable conduct and evidence before claiming to know anyone’s intentions.

We can still interrupt harassment without first solving that question. A boundary against repeated unwanted contact applies regardless of why the person contacting someone believes it is justified.

7. How communities can protect themselves

Healthier conflict requires practices that remain useful when people are angry and trust is low.

Shirky also argues that durable groups need governance, including informal norms, and that some limits on participation can protect their shared purpose. His essay is an argument about group design, rather than experimental evidence.

For conflict, I would apply this through rules agreed before a crisis: independent review, limits on repeated replies, and a clear route for appeals. Those safeguards should constrain established members and moderators too.

Separate actions from identity. Describe what someone did and its effects. Patterns matter, but they should be demonstrated rather than inferred from an unpopular person’s every action.

Substantiate accusations. Take reports seriously, investigate specific claims, and distinguish verified facts from interpretation. Receiving a report with care does not require deciding its outcome in advance.

Protect procedural fairness. Give people meaningful notice of allegations and an opportunity to respond. Protect reporters’ privacy, provide a reporting route that does not require public exposure, and review urgent protective measures once the immediate risk is contained.

Allow disagreement without demanding loyalty. Members should be able to question a process without being treated as supporters of misconduct.

Consider severity, repetition, and repair. A mistake, repeated harassment, and an ongoing threat require different responses. State the scope of a sanction and how mistakes in the decision can be challenged.

Make reconciliation possible where appropriate. Set clear conditions for return when return is possible. Reconciliation should never require an injured person to resume contact or forgive. Some conduct warrants permanent exclusion.

These practices cannot guarantee agreement. They give a community ways to act without letting popularity or outrage decide every question.

8. The real test of a community

It is easy to defend fairness when everyone agrees. The harder test comes when someone is unpopular, has said something offensive, or is accused of violating the community’s values.

Can members distinguish evidence from speculation? Can they criticize misconduct without dehumanizing the person responsible? Can they question the prevailing judgment without fear of retaliation?

As participants, we also have a smaller decision to make each time a dispute reaches us: whether our next contribution offers evidence or practical help, or simply adds another punishment.

The health of a community is measured by how consistently it applies its values when doing so becomes uncomfortable. That includes recognizing when its own response is causing harm.

A healthy community must be able to hold people accountable without turning accountability into a mechanism of collective hostility.

Further reading