Conformity, Obedience and Social Proof: Why We Follow the Crowd
Why can a crowd, norm or authority alter our judgment? A synthesis of informational and normative influence, social proof, pluralistic ignorance, obedience and the conditions for independent dissent.
If many people do something, that does not by itself prove the action is right. If a majority believes something, that does not by itself prove the claim is true. Yet it does not follow that following others is always irrational. Other people are often important sources of information, norms, coordination and safety—and for precisely that reason they can also become sources of systematic error.
This article connects four processes that everyday language often blends together: conformity, obedience, social proof, and pluralistic ignorance. The first concerns adjustment toward a group, the second a response to a request or command from authority, the third the use of other people's behavior as information about what is likely correct or normal, and the fourth a situation in which many privately doubt a norm while mistakenly believing that others accept it.
The central lesson is not 'never follow the crowd.' A more useful question is: do the others provide independent evidence, or are they merely repeating the same signal—and am I seeing their genuine judgment or only their public accommodation?
Social influence is not a defect in the human mind
Humans are social creatures. We learn most language, behavioral rules, practical skills and warnings about danger from other people. If everyone in an unfamiliar building suddenly leaves, it may be entirely rational to check whether they know something we do not. Social influence is therefore not a psychological malfunction but part of normal learning and coordination.
In their review of social influence, Cialdini and Goldstein emphasize three broad goals that often shape our responses to others: we want to understand reality accurately, preserve meaningful relationships, and maintain a favorable self-concept. The same person may conform in one situation because others are informative and in another because exclusion is costly.
The problem therefore does not begin with influence itself but with its quality. We need to ask where the signal comes from, how independent the sources are, what disagreement costs, and whether someone occupies a position from which obedience can be demanded.
Conformity is a shift in behavior, judgment or expressed opinion toward a group or perceived norm. Obedience is narrower: it involves a more direct request or command from a person or institution experienced as an authority. Public accommodation does not necessarily imply private belief.
In 1958, Kelman distinguished compliance, identification and internalization: people may outwardly comply because of consequences or expectations, adjust because of a valued relationship or identification, or genuinely integrate a new position into their belief system. This is not a complete taxonomy of social influence, but it is an important reminder that identical behavior can reflect different internal processes.
A room full of raised hands therefore does not prove that everyone thinks the same thing. And the fact that someone carried out an order does not by itself reveal whether they agreed, feared consequences, trusted expertise or simply stayed within an assigned role.
Deutsch and Gerard classically distinguished informational from normative social influence. Under informational influence, other people become evidence about reality: when a situation is ambiguous, it can be reasonable to infer that a group may be detecting something we have missed.
This strategy can work extremely well. If several people formed their estimates independently, possess different information and do not share the same source of error, their combined signal may contain more knowledge than one judgment alone. The same logic fails when ten people merely repeat the first person, all rely on the same misleading source, or no one wants to be the first to voice doubt.
So one of the most important questions behind 'everyone says so' is: how many genuinely independent observations are behind that consensus? The number of voices and the number of independent evidential pathways are not the same quantity.
Normative influence: the price of standing out
Normative influence works differently. A person may know or suspect that the group is wrong yet publicly accommodate it because they want acceptance, wish to avoid ridicule or do not want conflict. Sometimes the cost of difference is minor; in other settings it can threaten position, friendship, employment or status within a community.
This helps explain why public consensus is not always a reliable map of private opinion. When disagreement carries a high social penalty, an apparently unified group can contain many privately uncertain members. Their silence then becomes new evidence for everyone else that 'apparently everybody believes this.'
Why Do We Defend Our Beliefs Even When We Are Wrong? showed how identity shapes belief defense. This article adds the social layer: sometimes we are protecting not only an idea but our position within a network of relationships attached to that idea.
Asch's line-judgment experiments became famous because the task was deliberately simple and the confederate majority was unanimously wrong on critical trials. The result was not that 'all people blindly follow the crowd.' Many responses remained independent, but the unanimous majority produced a clear and repeatable pressure toward accommodation.
More important than one famous percentage are the conditions. Asch showed that unanimity, the presence of an ally and the way a response is given all matter. Bond and Smith's meta-analysis of Asch-type studies also found that conformity varied across time and cultural settings.
Why Conformity Often Overrides the Facts remains the owner article for conformity in detail. The broader point here is that a crowd does not need physical coercion to alter public behavior; perceived group unity alone can change the psychological cost of an independent answer.
A single crack in consensus can change the situation
One of the most practically useful findings from Asch's research program is that pressure from a unanimous majority decreases when unanimity breaks. An ally does not prove that the minority position is correct, but it changes the social geometry: a person is no longer the only visible dissenter.
This reveals a crucial distinction between evidence and social permission to doubt. The first dissenting voice may allow others to say what they already believed. In other cases a consistent minority can trigger genuine reconsideration over time; meta-analytic work on minority influence shows that influence is not reserved exclusively for numerical majorities.
The practical implication is strong: systems that want high-quality decisions need safe channels for dissent. If everyone sees only the final consensus, they cannot see the process by which that consensus was produced.
Cialdini, Reno and Kallgren emphasized the distinction between a descriptive norm—what people actually do—and an injunctive norm—what people approve or believe ought to be done. The two can align, but they can also point in opposite directions.
If a place is covered in litter, the descriptive norm signals that people litter there even if a sign says it is forbidden. If everyone privately condemns a practice but publicly performs it, behavior can create a false appearance of approval. To understand a norm, then, ask separately: what do people do, what do they privately approve, and what do they think other people approve?
This distinction prevents quick inferences. A common practice is not necessarily morally endorsed, a common opinion is not necessarily true, and public silence is not necessarily private agreement.
Social proof: a useful shortcut that can become a dangerous counter
Social proof refers to using other people's behavior or choices as a cue for one's own judgment. Under uncertainty this is intuitive: a busy restaurant, a product with many ratings, or a message shared by many people can signal that something deserves attention.
But digital counters can also help create the signal they appear merely to measure. In a large randomized online experiment, Muchnik, Aral and Taylor found that an artificially added initial positive vote increased the likelihood of later positive ratings and produced a cumulative effect. This does not mean every like is manipulation; it means visible popularity can be not only a measurement of preference but one of its causes.
View counts, ratings and shares therefore do not equal independent reality checks. Social proof can be a useful attention signal, but it is not a substitute for evaluating the claim itself.
Pluralistic ignorance occurs when individuals privately reject or doubt a norm yet mistakenly estimate that most others accept it. Prentice and Miller demonstrated this in research on college students' attitudes toward campus drinking practices: participants perceived themselves as more uncomfortable with the norm than they believed their peers were.
The mechanism can reinforce itself. Each person sees the public behavior of others but not their private doubts. Because nobody wants to stand out, many people perform a role they do not endorse as strongly as it appears from outside. A mistaken belief about the group's belief can therefore become a real force that maintains the group pattern.
The best remedy is not always louder persuasion. Sometimes it is enough to make private preferences measurable in a safe form—anonymous responses, for example—and discover that the supposed consensus was never as solid as it looked.
Other people's silence is not reliable evidence that there is no danger
A related problem appears when people watch others to decide whether a situation is serious. Latané and Darley's bystander research showed that the presence of other people can inhibit intervention through several processes, including diffusion of responsibility and mutual monitoring.
Yet the popular story is again too simple. Fischer and colleagues' meta-analysis supported an overall bystander effect while also identifying important moderators; in some dangerous emergencies inhibition can shrink or reverse, especially when other bystanders provide potential physical support. 'More people means nobody helps' is therefore not a universal law.
For this article the more important epistemic point is this: if everyone uses everyone else's calmness as evidence that 'apparently nothing is wrong,' a group can generate a circular signal without independent verification.
In Milgram's 1963 baseline experiment, 26 of 40 participants, under pressure from the experimenter, continued to the highest marked level on the sham shock generator. Participants did not know that the 'learner' was not actually receiving electric shocks. The finding matters, but it should not be turned into the slogan that '65 percent of people will always obey.'
Milgram's broader research program included many variations, and obedience rates changed substantially with the situation. Burger's ethically constrained partial replication decades later found that a large proportion of participants were still willing to continue past the critical 150-volt point, but the altered procedure means absolute rates cannot simply be equated with Milgram's originals.
Obey an Order or Follow Conscience? treats obedience and conscience in detail, while Does Authority Remove Personal Responsibility? focuses on responsibility. This article uses Milgram for a narrower synthesis: institutional framing, perceived legitimacy, incremental escalation and the presence of authority can substantially reshape choice.
'Just following orders' is not the only interpretation of Milgram
Traditional accounts of Milgram often emphasized an 'agentic state' in which people see themselves mainly as instruments of authority. Later research has challenged that account. Haslam, Reicher and Birney, for example, found in an experimental analogue that prompts invoking the scientific goal were more effective than the most direct order-like prompt.
This supports an alternative idea of engaged followership: people may persist not because they stop thinking, but because they identify with the authority, institution or higher purpose presented to them. The interpretation is influential, but it is not a reason to declare one new theory the final explanation of all Milgram findings.
The practical implication matters. Resisting harmful pressure is not only practice in saying 'no.' It also requires examining the story that makes continued participation feel like service to something legitimate, expert or good.
In complex systems, legitimate functional authority can be useful: a surgeon, air-traffic controller or emergency coordinator may have better information and clearly assigned responsibility. Social psychology therefore does not show that every hierarchy is harmful or that following instructions is inherently servile.
At the same time, the empirical fact that people often follow authority does not establish a moral right of authority to command. That is a separate philosophical and political question. In the project source The End of All Evil, the so-called 'authority principle' functions as a normative critique of authority; this article may use it as a philosophical frame for the legitimacy question, not as a replacement for empirical research on actual human behavior.
This is also the boundary with Obey an Order or Follow Conscience? and Does Authority Remove Personal Responsibility?: psychology helps explain why we obey; ethics and political philosophy must separately evaluate when an order is justified, when objection is a duty, and how responsibility is distributed.
Crowds can be wise—while their judgments remain sufficiently independent
A critique of conformity should not slide into the opposite dogma that an individual is always smarter than a group. Under suitable conditions, aggregating many independent estimates can cancel some individual errors. But that advantage requires at least some diversity of information and judgment.
Lorenz and colleagues showed experimentally that even relatively mild social influence can reduce the diversity of estimates without improving collective accuracy, while increasing participants' confidence after their answers converge. The finding does not prove that information exchange is always harmful; it reveals a trade-off between learning from others and preserving independent signals.
This is why collective decision systems may benefit from gathering independent judgments before discussion begins. If everyone hears the first speaker before forming an estimate, the later 'consensus' can contain far fewer independent pieces of information than the number of people in the room suggests.
Resistance to social influence does not mean automatically opposing the majority. Anticonformity is still behavior defined by the group—only with the sign reversed. The goal is more independent judgment: willingness to follow the group when it has better evidence, and willingness to depart when it does not.
Useful procedures are concrete: record an important judgment privately before discussion; separate the number of people from the number of independent sources; ask what people would say anonymously; deliberately create a protected channel for dissent; require reasons from leaders rather than status alone; in emergencies assign responsibility to a specific person; distinguish descriptive from injunctive norms; and with online metrics, hide popularity until content has been evaluated.
This article's strongest question is therefore simple: If everyone else were silent, what would I think on the evidence—and if everyone else said the opposite, what new evidence would they actually need to bring for me to rationally change my mind?
Sources and further reading
- Deutsch, M. & Gerard, H. B. (1955). A Study of Normative and Informational Social Influences Upon Individual Judgment. Journal of Abnormal and Social Psychology 51(3), 629–636 — foundational distinction between normative and informational social influence.
- Asch, S. E. (1956). Studies of Independence and Conformity: I. A Minority of One Against a Unanimous Majority. Psychological Monographs 70(9), 1–70 — classic experimental program on majority pressure, unanimity and independence.
- Bond, R. & Smith, P. B. (1996). Culture and Conformity: A Meta-Analysis of Studies Using Asch's Line Judgment Task. Psychological Bulletin 119(1), 111–137 — meta-analysis of 133 Asch-type studies across 17 countries, showing temporal and cross-cultural variation.
- Cialdini, R. B. & Goldstein, N. J. (2004). Social Influence: Compliance and Conformity. Annual Review of Psychology 55, 591–621 — review organizing social influence around accuracy, relationships and self-concept goals.
- Kelman, H. C. (1958). Compliance, Identification, and Internalization: Three Processes of Attitude Change. Journal of Conflict Resolution 2(1), 51–60 — distinguishes outward compliance, relationship-based identification and deeper internalization.
- Cialdini, R. B., Reno, R. R. & Kallgren, C. A. (1990). A Focus Theory of Normative Conduct. Journal of Personality and Social Psychology 58(6), 1015–1026 — experimental distinction between descriptive and injunctive norms.
- McDonald, R. I. & Crandall, C. S. (2015). Social Norms and Social Influence. Current Opinion in Behavioral Sciences 3, 147–151 — review of imitation, conformity and social norms as core social processes.
- Prentice, D. A. & Miller, D. T. (1993). Pluralistic Ignorance and Alcohol Use on Campus: Some Consequences of Misperceiving the Social Norm. Journal of Personality and Social Psychology 64(2), 243–256 — four studies demonstrating systematic misperception of peers' private attitudes.
- Darley, J. M. & Latané, B. (1968). Bystander Intervention in Emergencies: Diffusion of Responsibility. Journal of Personality and Social Psychology 8(4 Pt.1), 377–383 — classic experimental evidence that perceived presence of other bystanders can inhibit intervention.
- Fischer, P. et al. (2011). The Bystander-Effect: A Meta-Analytic Review on Bystander Intervention in Dangerous and Non-Dangerous Emergencies. Psychological Bulletin 137(4), 517–537 — meta-analysis showing an overall bystander effect with important situational moderators.
- Milgram, S. (1963). Behavioral Study of Obedience. Journal of Abnormal and Social Psychology 67(4), 371–378 — original baseline report in which 26 of 40 participants reached the highest marked shock level under experimental pressure.
- Burger, J. M. (2009). Replicating Milgram: Would People Still Obey Today? American Psychologist 64(1), 1–11 — ethically constrained partial replication stopping around the critical 150-volt decision point.
- Haslam, S. A., Reicher, S. D. & Birney, M. E. (2014). Nothing by Mere Authority: Evidence that in an Experimental Analogue of the Milgram Paradigm Participants Are Motivated not by Orders but by Appeals to Science. Journal of Social Issues 70(3), 473–488 — evidence supporting engaged-followership interpretations and challenging simple command-obedience accounts.
- Muchnik, L., Aral, S. & Taylor, S. J. (2013). Social Influence Bias: A Randomized Experiment. Science 341(6146), 647–651 — large randomized online experiment showing that prior visible ratings can causally bias later evaluation behavior.
- Lorenz, J., Rauhut, H., Schweitzer, F. & Helbing, D. (2011). How Social Influence Can Undermine the Wisdom of Crowd Effect. PNAS 108(22), 9020–9025 — experiment showing reduced diversity and increased confidence under social influence without corresponding gains in collective accuracy.
- Wood, W., Lundgren, S., Ouellette, J. A., Busceme, S. & Blackstone, T. (1994). Minority Influence: A Meta-Analytic Review of Social Influence Processes. Psychological Bulletin 115(3), 323–345 — meta-analytic evidence that minorities can influence majority judgment under identifiable conditions.