In the summer of 1969, Philip Zimbardo set up what appeared to be a straightforward electric shock experiment at New York University. Participants were told they were studying the effects of punishment on learning. Half of the participants were made to feel anonymous: they wore large lab coats and hoods that obscured their faces, were never referred to by name, and were processed in groups without individual acknowledgment. The other half were individuated: they wore name tags, were greeted by name, and felt personally visible throughout the procedure. When instructed to administer shocks to a learner — a confederate who received no actual shock but performed escalating expressions of distress — the anonymous participants delivered shocks that were twice as long as those delivered by participants who had been individuated. Zimbardo published these findings in the Journal of Personality and Social Psychology in 1969, and the concept of deindividuation entered the empirical literature with considerable force.
The results seemed to confirm something that crowd theorists, historians, and moralists had long suspected: that anonymity, submersion in a group, and the erosion of personal identity loosen the constraints that ordinarily keep human behavior within socially sanctioned limits. When people cannot be seen and identified as individuals — when they become, in effect, nobody — they may act in ways that the same people, clearly identified and personally accountable, would not. The finding was alarming not because it described monsters but because it described ordinary people whose ordinary moral restraints had been dissolved by situational conditions that stripped away the self.
What followed over the subsequent five decades was a research program of remarkable productivity and a theoretical evolution of comparable complexity. The early, intuitive model of deindividuation — anonymous crowds become impulsive, aggressive, and uninhibited — proved to be, at best, a partial account. The mechanisms were real but the interpretation was incomplete. Critiques accumulated. Alternative frameworks emerged. The most influential of these, the Social Identity model of Deindividuation Effects, argued that anonymity does not strip away norms but shifts which norms apply. By the early 2000s, deindividuation had become one of social psychology's most contested and intellectually generative concepts.
Individuated vs. Deindividuated States
The following comparison captures the core psychological differences between states of high individual salience and states of reduced individual identity as described in the literature from Festinger et al. (1952) through Diener (1980) and Prentice-Dunn and Rogers (1982).
| Dimension | Individuated State | Deindividuated State |
|---|---|---|
| Self-awareness | High — ongoing monitoring of one's own thoughts, feelings, and actions against personal standards | Reduced — attention shifts outward toward the group and immediate environment |
| Identity salience | Personal identity is salient; individual values, beliefs, and commitments are psychologically active | Personal identity recedes; group membership or collective identity becomes the operative frame |
| Behavior regulation | Internal standards, personal guilt, and anticipated individual accountability restrain behavior | Reduced internal monitoring; behavior is more impulsive, more responsive to immediate situational cues |
| Sensitivity to group norms | Moderate — group norms are one input among several | High — behavior becomes strongly guided by whichever norms are most salient in the group context |
| Sense of personal responsibility | High — outcomes are attributed to the self | Diffuse — responsibility is distributed across the group or assigned to leaders and authority figures |
| Temporal orientation | Reflective — past behavior and future consequences influence present action | Present-focused — immediate stimuli and group momentum drive behavior |
The Intellectual Lineage
The concept of deindividuation did not emerge from laboratory psychology alone. Its roots reach into nineteenth-century crowd theory, specifically into the work of the French sociologist Gustave Le Bon, whose 1895 book Psychologie des foules described the crowd as a unified organism capable of transforming rational individuals into something more primitive and dangerous. Le Bon argued that in a crowd, a person descends several rungs in the ladder of civilization: the crowd mind is impulsive, credulous, and extreme. Le Bon's framework was explicitly biological and racial in ways that have dated catastrophically, but his core observation — that group immersion changes individual psychology — proved durable.
Sigmund Freud engaged directly with Le Bon in his 1921 essay Group Psychology and the Analysis of the Ego, interpreting crowd behavior through the lens of libidinal ties and the replacement of the individual's superego with the group's leader. Freud's contribution was to psychologize what Le Bon had biologized: the crowd did not produce a different organism but activated different psychic structures. The individual's ego ideal — the internalized standard against which behavior is measured — was, Freud argued, temporarily surrendered to the collective.
The formal scientific lineage begins with Leon Festinger, Albert Pepitone, and Theodore Newcomb, who published "Some Consequences of De-individuation in a Group" in the Journal of Abnormal and Social Psychology in 1952. They were studying the conditions under which individuals felt free to express hostile feelings toward their parents — a topic of clinical significance in the post-war period — and found that groups in which members felt submerged and unidentifiable produced more anti-parental statements than groups in which members remained individually salient. Festinger and colleagues coined the term "deindividuation" to describe the state in which the individual becomes less prominent as a unit in their own perception and in the perception of others.
Zimbardo, who had been a graduate student in the tradition of social psychology that produced both Milgram's obedience studies and Festinger's cognitive dissonance work, built directly on the Festinger et al. framework when designing his 1969 shock experiments. He expanded the theoretical account considerably, arguing that deindividuation was produced by a cluster of antecedent conditions — anonymity, arousal, sensory overload, altered consciousness, and diffuse responsibility — and that it produced a cluster of consequences: reduced inhibition, impulsivity, reduced guilt and shame, a weakened connection to normative standards, and heightened sensitivity to group cues.
Revised Theories: Diener and Prentice-Dunn and Rogers
The next major theoretical revision came from Edward Diener at the University of Illinois. Writing in a 1980 chapter in Paul Paulus's edited volume Psychology of Group Influence, Diener proposed that the core of deindividuation was not anonymity per se but reduced self-awareness. Drawing on the by-then influential literature on objective self-awareness developed by Shelley Duval and Robert Wicklund, Diener argued that self-awareness is the psychological mechanism that keeps behavior aligned with internal standards. When people focus attention on themselves — when they are in front of a mirror, when they have been made personally visible, when their individual identity is salient — they compare their current behavior against their internalized standards and experience affect (guilt, shame, pride) that motivates conformity to those standards. Deindividuation, on this account, is the disruption of this self-regulatory cycle. Anonymity and group immersion matter not because they make people feel they won't be caught, but because they shift attention away from the self entirely.
Scott Prentice-Dunn and Ronald W. Rogers refined this model further in two influential papers published in 1982 and 1983. Writing in the Journal of Personality and Social Psychology, they distinguished between two types of self-awareness that contribute to deindividuation by distinct routes. Public self-awareness — awareness of how one appears to others — is reduced by anonymity. When people cannot be identified, they lose concern with their social reputation and the external sanctions that reputation enables. Private self-awareness — awareness of one's own internal states, values, and standards — is reduced by group immersion and arousal. When attention is entirely captured by the group and the immediate environment, people lose contact with their own conscience. Prentice-Dunn and Rogers argued, through experimental work, that it was the reduction of private self-awareness, not public self-awareness, that produced the most dramatic behavioral change. A person who is anonymous but privately self-aware retains their internal moral compass. A person who is public-anonymous and privately unaware loses both external accountability and internal guidance simultaneously.
Cognitive Science Section
The deindividuation literature intersects with several streams in cognitive science in ways that were not fully visible to the field's founders. The self-regulatory cycle that Diener and Prentice-Dunn and Rogers placed at the center of their models maps closely onto what cognitive neuroscientists would later call executive function — the suite of processes centered in the prefrontal cortex that support planning, impulse inhibition, and the monitoring of behavior against goals. States of high arousal, intoxication, and certain altered states demonstrably reduce prefrontal activity and diminish executive function. This neural observation provides a mechanistic account for why arousal — one of Zimbardo's original antecedent conditions — contributes to deindividuated behavior: it is not merely that aroused people feel less self-conscious but that arousal biochemically impairs the systems that regulate behavior against internalized standards.
Roy Baumeister and colleagues' work on ego depletion, published across the 1990s and 2000s in the Journal of Personality and Social Psychology and elsewhere, introduced a related concept: self-regulation is a limited resource that can be exhausted. Extended group immersion, particularly in high-arousal situations, may deplete the regulatory capacity that keeps individual behavior aligned with personal values, creating a functional deindividuation through a different route.
Claude Steele and Robert Josephs' research on alcohol myopia, published in Psychological Review in 1990, is directly relevant: they demonstrated that alcohol narrows the attentional field to the most immediate cues, reducing the capacity to consider distal consequences or internal standards. Many of the real-world situations in which deindividuated behavior is most clearly documented — riots, certain crowd events, military atrocities — involve states of arousal that may produce analogous attentional narrowing through non-pharmacological routes.
Four Named Case Studies
Case Study 1: Zimbardo's Electric Shock Study (1969)
Philip Zimbardo's 1969 experiment at New York University remains the foundational laboratory demonstration of deindividuation effects. The design manipulated anonymity through costume: half of participants wore hoods and oversized lab coats and were never addressed by name; the other half wore name tags and were individually introduced. Both groups were instructed to administer electric shocks to a "learner" (a confederate) as part of a learning study. The anonymous group delivered shocks of nearly twice the duration as the individuated group, and they did so regardless of whether the confederate had been previously described as a pleasant, kind person or as an obnoxious one. Individuated participants modulated their behavior based on the confederate's described personality — they gave the pleasant confederate shorter shocks. The anonymous participants did not discriminate: they were equally punitive toward both.
Zimbardo interpreted this finding as evidence that anonymity removes the feedback loop between individual identity and behavior. When participants couldn't be seen, they couldn't be judged, and the inhibitory function of anticipated social evaluation was eliminated. The study was published in the Journal of Personality and Social Psychology (1969) and has been cited thousands of times, though its reliance on a single laboratory context and the specific manipulation of hood-and-lab-coat anonymity has attracted methodological scrutiny in subsequent decades.
Case Study 2: The Stanford Prison Experiment (1971)
The most famous application of deindividuation principles to an extended real-world-simulating context was conducted by Craig Haney, Curtis Banks, and Philip Zimbardo at Stanford University in 1971. Twenty-four male college students, pre-screened for psychological stability, were randomly assigned to roles as guards or prisoners in a mock prison constructed in the basement of Stanford's psychology building. The experiment was intended to run for two weeks. It was halted after six days.
Within the first twenty-four hours, the guards had begun to abuse their authority in ways that escalated steadily: forcing prisoners into stress positions, depriving them of sleep, subjecting them to arbitrary and humiliating punishments. Prisoners, for their part, began to exhibit acute emotional distress, submissive behavior, and in several cases complete psychological breakdown. The deindividuation elements were structural: guards wore mirrored sunglasses and khaki uniforms, which obscured their individual identities and merged them into a role. Prisoners wore smocks and were referred to only by their assigned numbers. The elimination of personal identity on both sides, combined with a clear authority hierarchy, a physically enclosed environment, and escalating arousal, produced behavior that the participants themselves later found incomprehensible.
The Stanford Prison Experiment became a landmark in public understanding of deindividuation. It also became, decades later, a major site of methodological and ethical critique. Haslam and Reicher, in a 2012 paper published in PLoS ONE, and in Reicher and Haslam's own BBC Prison Study conducted in 2001 and published in the British Journal of Social Psychology in 2006, challenged the determinism of Zimbardo's account. They argued that the guards in Zimbardo's study were not simply overwhelmed by situational forces but were actively encouraged toward brutality by Zimbardo himself in his role as prison superintendent. The guards who did not participate in abuse — and there were some — were systematically ignored in the original theoretical account, which Haslam and Reicher argued produced a misleading picture of how roles and group identity actually shape behavior.
Case Study 3: Mann's Baiting Crowds (1981)
Leon Mann at Flinders University in Australia published a 1981 analysis in the Journal of Personality and Social Psychology that took a deliberately non-laboratory approach to deindividuation. Mann systematically reviewed newspaper accounts of 21 incidents in which crowds had gathered at sites where a person was threatening to jump from a building or bridge. In 10 of the 21 cases, crowd members had baited the potential jumper — shouted encouragement to jump, mocked, jeered. Mann then examined the conditions under which baiting occurred and found that it was significantly more likely when the crowd was large, when the incident occurred at night (reducing individual identifiability), and when the crowd was physically distant from the jumper (reducing the personal immediacy of the situation). Each of these conditions maps directly onto Zimbardo and Diener's accounts of what produces deindividuation: large crowds diffuse individual identity; night reduces personal visibility; distance reduces individual salience of the consequences.
Mann's study is significant for several reasons. It demonstrated that deindividuation could produce antisocial behavior of a kind that was not merely laboratory-measured discomfort but real, documented cruelty toward a person in genuine danger. It also demonstrated the research methodology of archival analysis as a complement to laboratory work — a way of examining whether laboratory findings hold in naturalistic settings. The finding that baiting was more common at night, in large crowds, and from a distance has proven robust to subsequent analysis.
Case Study 4: Silke and Political Violence in Northern Ireland (2003)
Andrew Silke, a forensic psychologist then at the University of Leicester, published a study in 2003 in the Howard Journal of Crime and Justice that examined the role of disguise in violent offenses during the Northern Ireland conflict. Analyzing 500 violent incidents drawn from police and court records, Silke found that attackers who were disguised — wearing masks, balaclavas, or other concealing clothing — committed significantly more severe violence than attackers who were not disguised. Disguised attackers were more likely to assault multiple victims, more likely to use weapons, more likely to inflict serious injuries, and more likely to continue attacks even when victims had ceased to resist.
Silke's study is particularly important because it extends deindividuation research into a domain — organized political violence — where other explanations (ideological commitment, group pressure, military discipline) might seem more parsimonious. The finding that physical disguise predicted severity of violence even within a population of people already committed to using violence suggests that deindividuation adds an independent contribution to behavioral disinhibition beyond ideology or group pressure alone. The study also connects to Prentice-Dunn and Rogers' emphasis on public self-awareness: disguise specifically eliminates the capacity for others to identify the individual, which is precisely the manipulation that reduces public self-awareness and its associated inhibitory function.
Empirical Research Section
Diener et al.'s Halloween Study (1976)
Edward Diener and colleagues published a deceptively simple field experiment in the Journal of Personality and Social Psychology in 1976 that has become one of the most cited demonstrations of deindividuation in naturalistic conditions. On Halloween, research assistants positioned a bowl of candy in a home with instructions to children that they should take one piece each. Some children arrived alone and were asked their name and address; others arrived in groups and were not asked for identifying information. Children who were anonymous took significantly more candy than children who were identified, and children in groups took more than children alone. The combination of anonymity and group membership — the two core antecedent conditions of deindividuation — produced the most pronounced violation of the stated norm.
The study's elegance lies in its mundane stakes. The candy experiment does not involve violence or extreme behavior; it involves ordinary children violating an ordinary social norm under conditions of low personal accountability. This minimality is precisely what makes it compelling: it suggests that deindividuation effects do not require extreme situations or disturbed personalities but operate across the full range of human social settings.
Postmes and Spears 1998 Meta-Analysis
The most important empirical challenge to the classical deindividuation account was published by Tom Postmes and Russell Spears in Psychological Bulletin in 1998. Postmes and Spears conducted a meta-analysis of 60 deindividuation studies and found that the evidence for the classical model was far weaker than the literature's prominence might suggest. The meta-analysis found no consistent relationship between deindividuation manipulations and increased aggression or antisocial behavior taken as a general class. Instead, the pattern of findings was substantially more complex: behavior under deindividuated conditions was not uniformly more impulsive or antisocial but rather more responsive to the prevailing norms of the group. When the group context emphasized antisocial norms, deindividuation increased antisocial behavior. When the group context emphasized prosocial norms, deindividuation increased prosocial behavior.
This finding — that deindividuation increases normative sensitivity rather than normative absence — was the empirical foundation of the SIDE model, and it fundamentally revised the theoretical landscape.
The SIDE Model
Stephen Reicher, Russell Spears, and Tom Postmes published the Social Identity model of Deindividuation Effects in 1995 in the European Review of Social Psychology, and it represents the most significant theoretical development in the deindividuation literature since Zimbardo's original formulation. The SIDE model begins from a critique of the classical account: it is not true, the model's authors argued, that anonymity and group immersion eliminate normative behavior. What they eliminate is the salience of personal identity. And when personal identity becomes less salient, social identity — identity as a member of a group — becomes more salient. This shift does not produce normlessness; it produces a shift in which norms apply.
The SIDE model distinguishes between cognitive and strategic effects. Cognitively, anonymity increases identification with the group and increases the influence of group norms on behavior. Strategically, anonymity reduces the power of external authorities to identify and sanction individual members, which frees individuals to express group norms they might have suppressed when personally visible and individually vulnerable.
The SIDE model made predictions that differed from the classical model in important ways. It predicted that deindividuated crowds would not behave randomly or impulsively but would behave consistently with the norms of their group, whether those norms were violent or orderly. It predicted that online anonymity would increase adherence to group norms rather than producing lawless behavior. And it predicted that the outcomes of deindividuation would vary systematically with the character of the group rather than being uniformly antisocial.
Reicher and colleagues tested the SIDE model in studies of crowd behavior, computer-mediated communication, and political protest, consistently finding that anonymity shifted behavior toward group norms rather than toward uninhibited impulsivity. The model has achieved substantial acceptance within European social psychology while remaining more contested in American research traditions that maintained closer ties to the classical Zimbardian account.
Online Disinhibition as Modern Extension
John Suler, a clinical psychologist at Rider University, published "The Online Disinhibition Effect" in CyberPsychology and Behavior in 2004, identifying a cluster of phenomena in computer-mediated communication that parallel the classic deindividuation account while adding features specific to the digital environment. Suler distinguished between benign and toxic disinhibition. Benign disinhibition — sharing personal vulnerabilities, offering generosity, exploring emotions with unusual openness — is prosocial in character and is facilitated by the privacy of online contexts. Toxic disinhibition — cruelty, threats, extreme self-indulgence, antisocial expression — is the online analogue of the aggression that classical deindividuation theory predicted.
Suler identified six factors that produce online disinhibition: dissociative anonymity (the sense that the online self is separate from the real self), invisibility (the absence of physical presence and its associated social monitoring), asynchronicity (freedom from real-time consequences), solipsistic introjection (the sense that online communication is an internal rather than external event), dissociative imagination (treating online space as fictional rather than real), and minimization of authority (the flattening of hierarchy that online anonymity enables). Of these, dissociative anonymity maps most directly onto the classical deindividuation account. Minimization of authority maps onto Prentice-Dunn and Rogers' public self-awareness dimension. Solipsistic introjection and dissociative imagination introduce genuinely new features that have no direct parallel in the pre-digital literature.
The SIDE model's prediction — that online anonymity will shift behavior toward group norms rather than eliminating norms — has found support in the online context. Suler himself noted that toxic disinhibition tends to occur in contexts where aggressive or antisocial group norms already prevail. Platforms where cruelty is normalized produce cruel behavior under anonymous conditions; platforms where cooperation is the prevailing norm produce cooperative behavior under anonymous conditions. The classical prediction of uniform disinhibition is, as in offline research, not consistently supported.
Limits, Critiques, and Nuances
The classical deindividuation account has faced three major categories of critique.
The first and most empirically developed is the Postmes and Spears meta-analytic challenge. The claim that deindividuation reliably produces impulsive, antisocial behavior is not supported by the broader experimental record. The actual finding is that deindividuation increases behavioral sensitivity to the norms most salient in the immediate situation, which may be prosocial or antisocial depending on context. This finding does not eliminate the relevance of deindividuation as a construct but it substantially revises what the construct predicts.
The second critique targets methodology. Many of the foundational experiments in the deindividuation literature — including Zimbardo's 1969 shock study — used small samples, single-session designs, demand characteristics that may have communicated experimental expectations to participants, and manipulations that confounded multiple variables (Zimbardo's hoods, for instance, simultaneously produced anonymity, reduced individual identity cues, and created a visually unusual experience). The replication of specific deindividuation effects has been inconsistent across laboratories and measurement approaches.
The third critique, most forcefully developed by Haslam and Reicher, targets the theoretical framing rather than the empirical findings. Haslam and Reicher argued that the classical account presents deindividuation as a process of psychological loss — loss of self-awareness, loss of norms, loss of inhibition — when the actual psychological process is better understood as a transformation of identity rather than its elimination. The person in a crowd is not nobody; they are a member of a group, with a group identity that carries its own norms, loyalties, and standards. The Stanford Prison Experiment is, on this reading, not about the dissolution of individual identity but about the adoption of a role identity — guard or prisoner — that carried with it a set of group norms that Zimbardo's experimental design actively encouraged participants to enact. The methodological critique that Zimbardo, as prison superintendent, directly instructed guards to be tough, to harass prisoners, and to suppress rebellion, fundamentally undermines the claim that the guards' behavior was a product of situational forces alone rather than explicit instruction from an authority figure.
These critiques do not dissolve deindividuation as a scientifically useful concept. They do refine it. The most defensible current account integrates Diener's self-awareness mechanism, Prentice-Dunn and Rogers' distinction between public and private self-awareness, and the SIDE model's insight that the shift from personal to group identity does not produce normlessness but norm-switching. Deindividuation is real, its mechanisms are partially understood, and its effects are more complex and context-dependent than the dramatic laboratory images of hooded participants delivering electric shocks first suggested.
References
Festinger, L., Pepitone, A., & Newcomb, T. (1952). Some consequences of de-individuation in a group. Journal of Abnormal and Social Psychology, 47(2, Suppl.), 382–389. https://doi.org/10.1037/h0057906
Zimbardo, P. G. (1969). The human choice: Individuation, reason, and order versus deindividuation, impulse, and chaos. In W. J. Arnold & D. Levine (Eds.), Nebraska Symposium on Motivation (Vol. 17, pp. 237–307). University of Nebraska Press. (Key findings published in Journal of Personality and Social Psychology.)
Haney, C., Banks, W. C., & Zimbardo, P. G. (1973). Interpersonal dynamics in a simulated prison. International Journal of Criminology and Penology, 1(1), 69–97.
Diener, E. (1980). Deindividuation: The absence of self-awareness and self-regulation in group members. In P. B. Paulus (Ed.), Psychology of Group Influence (pp. 209–242). Lawrence Erlbaum Associates.
Diener, E., Fraser, S. C., Beaman, A. L., & Kelem, R. T. (1976). Effects of deindividuation variables on stealing among Halloween trick-or-treaters. Journal of Personality and Social Psychology, 33(2), 178–183. https://doi.org/10.1037/0022-3514.33.2.178
Prentice-Dunn, S., & Rogers, R. W. (1982). Effects of public and private self-awareness on deindividuation and aggression. Journal of Personality and Social Psychology, 43(3), 503–513. https://doi.org/10.1037/0022-3514.43.3.503
Reicher, S. D., Spears, R., & Postmes, T. (1995). A social identity model of deindividuation phenomena. European Review of Social Psychology, 6(1), 161–198. https://doi.org/10.1080/14792779443000049
Postmes, T., & Spears, R. (1998). Deindividuation and antinormative behavior: A meta-analysis. Psychological Bulletin, 123(3), 238–259. https://doi.org/10.1037/0033-2909.123.3.238
Mann, L. (1981). The baiting crowd in episodes of threatened suicide. Journal of Personality and Social Psychology, 41(4), 703–709. https://doi.org/10.1037/0022-3514.41.4.703
Silke, A. (2003). Deindividuation, anonymity, and violence: Findings from Northern Ireland. Journal of Social Psychology, 143(4), 493–499. https://doi.org/10.1080/00224540309598458
Suler, J. (2004). The online disinhibition effect. CyberPsychology and Behavior, 7(3), 321–326. https://doi.org/10.1089/1094931041291295
Haslam, S. A., & Reicher, S. D. (2012). Contesting the "nature" of conformity: What Milgram and Zimbardo's studies really show. PLoS Biology, 10(11), e1001426. https://doi.org/10.1371/journal.pbio.1001426
Frequently Asked Questions
What is deindividuation?
Deindividuation is the psychological state in which people lose their sense of individual identity within a group or anonymous situation, resulting in diminished self-awareness, reduced adherence to personal norms, and heightened responsiveness to immediate social cues and group behavior. The term was coined by Leon Festinger, Albert Pepitone, and Theodore Newcomb in their 1952 Journal of Abnormal and Social Psychology paper, which found that group members who recalled least about what individual members said and did — indicating low individuation within the group — showed the greatest hostile comments toward parents. Philip Zimbardo substantially expanded the construct in his 1969 essay 'The Human Choice: Individuation, Reason, and Order Versus Deindividuation, Impulse, and Chaos,' proposing a list of input variables (anonymity, arousal, altered consciousness, group size, novel settings) that reduce self-observation and self-regulation, and a list of consequent behaviors (increased impulsivity, destructiveness, and disinhibition).
What did Zimbardo's 1969 electric shock study demonstrate?
In the study that established deindividuation as an experimental phenomenon, Zimbardo had New York University female undergraduates administer electric shocks to confederates as part of an ostensible learning experiment. In the individuated condition, participants wore their normal clothes, were introduced by name, and wore visible name tags. In the deindividuated condition, participants wore identical oversized lab coats and hoods that concealed their faces and were treated as a group rather than as individuals. Across multiple trials, the hooded participants held shock buttons approximately twice as long as the individuated participants — and they did so equally for a confederate described as pleasant and one described as unpleasant. The deindividuated participants did not discriminate: they were not more aggressive toward unlikeable targets (which would suggest mere frustration or dislike) but were uniformly more willing to cause pain regardless of their personal feelings about the recipient. Zimbardo interpreted this as evidence that deindividuation removes the personal values and moral constraints that ordinarily regulate harmful behavior.
What is the SIDE model and how does it revise classical deindividuation theory?
Stephen Reicher, Russell Spears, and Tom Postmes's 1995 Social Influence paper introduced the Social Identity model of Deindividuation Effects (SIDE), which fundamentally revised the classical account. Classical deindividuation theory (Zimbardo, Diener) predicted that anonymity and reduced self-awareness eliminate norms and produce antisocial, disinhibited behavior. The SIDE model proposed instead that anonymity does not eliminate normative behavior — it shifts which norms are salient. When people are anonymous as individuals but visible as group members (for instance, wearing identical costumes or operating in online groups), personal identity recedes and social identity becomes more salient. The behavior that follows is not normless aggression but heightened conformity to group norms — which may be prosocial or antisocial depending on the group's values. SIDE explains why anonymous internet users in communities with strong norms of helpfulness can be extraordinarily cooperative, while those in communities with hostile norms can be extraordinarily aggressive: the mechanism is the same, but the norms differ.
What does research on real-world deindividuation show?
Leon Mann's 1981 Journal of Personality and Social Psychology analysis examined 21 cases from newspaper reports of crowds that had baited people threatening suicide by jumping from buildings. In 10 of the 21 cases, the crowd actively baited the person. Mann found that baiting was significantly more likely when the crowd was large (over 300 people), when the incident occurred at night, and when the crowd had been waiting for an extended time — all conditions associated with reduced individuation and increased anonymity. Andrew Silke's 2003 Human Factors and Aerospace Safety analysis of 500 violent attacks during the Northern Ireland conflict found that attackers who were disguised (wearing masks, hoods, or dark clothing) inflicted significantly more severe injuries, attacked more victims per incident, and engaged in more property destruction than undisguised attackers — even after controlling for other variables. Edward Diener, Scott Fraser, Arthur Beaman, and Roger Kelem's 1976 field study of Halloween children found that children who were costumed and in groups stole significantly more candy when given the opportunity than children who were identified or alone.
Do the empirical findings actually support the classical deindividuation model?
Tom Postmes and Russell Spears's 1998 Psychological Bulletin meta-analysis of 60 deindividuation studies — the most comprehensive review of the literature — reached conclusions that substantially qualify the classical theory. Anonymity did not reliably increase disinhibited or antisocial behavior across studies; effect sizes were heterogeneous and often small. The pattern that did appear consistently was that anonymity and group immersion increased conformity to whatever norms were situationally salient — which could produce either more or less aggressive behavior depending on the context. The classic finding that deindividuation produces uninhibited aggression held only in specific experimental paradigms (particularly those using anonymous shock delivery) and did not generalize to naturalistic settings. Postmes and Spears concluded that the SIDE model better accounts for the aggregate data than classical deindividuation theory, and that the popular image of anonymous crowds as fundamentally norm-free and antisocial is an oversimplification that the empirical literature does not consistently support.