In a crowd, there are intragroup processes (e.g. leaders and followers). There are also intergroup processes (e.g. police vs. protesters).

Common-bond groups refer to social groups based on attachment between group members (e.g. friends). The attachment is based on similarity (1), likability (2) and familiarity (3). There is more attachment to individual group members than to the group identity. Common-identity groups refer to social groups based on attachments to group identity (e.g. sports team). There is more attachment to group identity than to individual group members.

People in crowds often have no schema of behaviour or clear expectations about what is going to happen. However, there is a need that something needs to be done to address a common concern. Social contagion states that cognition (i.e. attitudes), affect (i.e. emotions) and behaviour spreads quickly among crowd members. Non-normative behaviour is more likely in crowds.

The traditional view of crowds is that crowds are outbursts of madness (1), crowds are irrational (2), crowds are a threat (3) and individual crowd members sacrifice their individuality (4). However, crowds are often non-violent (1), selective in their targets (2) and at the basis of social change (3). Crowds can be instrumental to individuals reaching their ideals. Meaningful crowd behaviour can be seen as collective action.

The mass society theory states that physical and social divisions between classes makes social order difficult to maintain as people might become vulnerable to manipulation and especially powerful in a single group when freed from hierarchical order.

Traditional research on crowd behaviour focuses on the individual. Zimbardo’s deindividuation approach states that anonymity leads to reduced self-observation and self-evaluation which, in turn, leads to fewer feelings of guilt and shame which lowers the thresholds for antisocial behaviour, making antisocial behaviour more likely. This has a strong focus on the individual and there is little attention to context.

Later deindividuation theories focus more on intragroup dynamics and underlying mechanisms that could explain why individuals show non-normative behaviour.

Public self-awareness refers to the concern about appearance and impressions made. Behaviour is monitored through other people’s standards (e.g. social norms). Private self-awareness refers to monitoring own behaviour regarding own standards (e.g. values).

Differential self-awareness theory of deindividuation states that a low public self-awareness will lead to the conception that there is no expected punishment by others for behaviour and that low private self-awareness will lead to deindividuation (i.e. no feelings of guilt). Low public self-awareness and low private self-awareness will lead to non-normative behaviour in crowds.

Lower accountability leads to low public self-awareness which leads to the notion that aggressive behaviour goes unpunished which leads to aggression in crowds. External attentional cues lead to low private self-awareness which leads to deindividuation which leads to aggression in crowds.

However, there is no strong support for deindividuation and non-normative behaviour although lower accountability does predict non-normative behaviour.

There are four reasons to study crowds:

1. Studying crowds reveal political influences.
2. Studying crowds provides lessons for research methods (i.e. hypothesis generation and validation).
3. Studying crowds provides lessons in theory (i.e. the relation between individual and society)
4. Studying crowds provides lessons on society (i.e. crowds reproduce social relations which constitute society).

The methodologies of studying crowds mostly rely on internal referentiality; the question of how an experiment relates to previous experiments. It should rely more on external referentiality; the question of how an experiment relates to the external world. The focus of studying crowds should be more on generating and validating hypotheses than testing hypotheses.

The social identity model of crowds states that crowd members shift from personal identity to social identity and do not lose their identity in a crowd. The elaborated social identity model of crowds (ESIM) states that conflict escalates because the outgroup treats everyone in the ingroup the same. This leads to the notion in the ingroup that the outgroup is a threat, which will lead to the ingroup feeling more connected to each other.

Enacting identity refers to the phenomenon that participants in crowd events make the identity more tangible and relevant by being a model of the identity. This gives a sense of identity (i.e. personal coherence) and provides a sense of continuity and agency (i.e. meaning in solidarity). People are mostly influenced by sources who share identity, values and perspective.

The impact of crowd events on the watching public depends on the way crowd members are categorized, making categorization central to what a crowd can achieve. The understandings and the organization of society can be reframed by crowds, depending on the categorical representation. However, it is not clear how a crowd can influence a non-crowd member.

The dialogue approach of dealing with crowds consists of understanding of aims and intentions of crowd members (1), helping the crowd achieve legitimate aims (2) and graded interventions targeting those causing disorder without denying majority rights.

Norm theory states that crowd events are shaped by collective norms emerging in the situation itself. This occurs in a period of milling which precedes action but this does not explain situations in which crowds act quickly without time for milling. Self-categorization theory states that defining oneself as a member of a social category is the precondition for group behaviour. This involves an attempt by the individual to conform to the standards of the group. The classic agitator theory states that a small group of anti-social individuals can take advantage of the fact that crowd members have lost their judgement in order to turn the mass to anti-social ends.

Conflict only occurs if violence is seen as legitimate (1) and when conflict is seen as efficacious (2).