aqa  Social psychological explanations of human aggression, including the frustration-aggression hypothesis, social learning theory as applied to human aggression, and de-individuation.

Social Learning Theory

Albert Bandura

According to Bandura (1965) aggressive behaviour is learned either through direct experience or by observing others.

Learning by direct experience: If a child pushes another child and as a result gets something they want, the action is reinforced and is more likely to occur in similar situations in the future.  These principles are similar to those of operant conditioning.

Learning by vicarious reinforcement: This form of observational learning occurs when a child sees a role model behaving in a particular way and reproduces that behaviour.  The child is then said to be imitating the behaviour of the model.

Media Violence & Aggression

SLT leads us to consider the various ways in which children might be exposed to aggressive models.  TV has been examined as a powerful source of imitative learning.  Huesmann (1988 cited in Cardwell et al., 2004) suggested that children may use television models as a source of ‘scripts’ that act as a guide for their own behaviour.  For example, if they see a movie hero beat up the bad guys that get in his way, this may become a script for any situation in which it might be deemed appropriate.  These scripts are stored in memory, and are strengthened and elaborated through rehearsal.

  • The relationship between observation of aggression in the media and subsequent aggressive behaviour is a complex one.  It appears to be influenced by several variables including:
  • If the observed violence is thought to be real behaviour compared to if it was considered fictional or fantasy violence.
  • If viewers identify with the aggressor in some way, they are subsequently more aggressive than if they do not identify with the aggressive model.  Heroes are therefore more powerful models than villains.
  • Observing unsuccessful aggression, in which the aggressor is punished tends to inhibit aggressive behaviour in the observer.[space=10]
Bandura, Ross and Ross (1963)

Bandura divided 66 nursery school children into three groups. All three groups watched a film where an adult model kicked and punched the Bobo doll.  There were three different conditions:

Bobo Doll

Condition 1: Children saw the adult model being rewarded by a second adult.
Condition 2: Children saw a second adult telling off the adult model for the aggressive behaviour.
Condition 3: The adult model was neither rewarded or punished.

The children were then allowed to play in the room with the Bobo doll whilst experimenters watched through a one-way mirror.


Condition 1: Children behaved the most aggressively.
Condition 2: Children behave least aggressively.

However, a distinction needs to be made between learning and performance.  All the children learnt how to behave aggressively, but those in condition two did not perform as many aggressive acts until later, when they were offered rewards to do so.  When this happened, they quickly showed that they had learned as many aggressive acts as the children in condition 1. [/message_box]

[toggle_box] [toggle_item title=”Supporting Research” active=”true”][message_box title=”” color=”blue”]

Patterson et al. (1989) offered further empirical support for Bandura’s study.  He demonstrated that role models are important in the development of anti-social behaviour (in both boys and girls) and that parents are the most important ones.  Through the use of surveys/questionnaires they found that that very aggressive children are raised in homes of high aggression, little affection, and little positive feedback.  This suggests that there is wider academic credibility for the important role played by parents when forming/developing their behaviour.

Bandura et al. (1963) provided further empirical support for SLT.  They found that viewing aggression by cartoon characters produces as much aggression as viewing live or filmed aggressive behaviour by adults.  This suggests that there is wider academic credibility for the notion that the entertainment industry has an influential role in the social development of children. [/message_box][/toggle_item] [toggle_item title=”A02/IDAs” active=”true”][message_box title=”” color=”red”]

Reliability: A second strength of the research into SLT is that it has high reliability.  The reason for this is because Bandura’s research was predominately carried out in the laboratory where he had complete control over the IV (whether there was positive/negative reinforcement) and the DV (the behaviour shown by the child).  This suggests that if the research carried out again then the same results could be achieved.

Cross-cultural support: Mead (1935) found The Arapesh is an example of a non-aggressive culture in which aggression is not admired (reinforced) or modelled by adults.  The Mundugmor show the opposite pattern, in which violence is the norm and status is determined by the amount of aggression shown.  This suggests that SLT can be applied universally. 

Ecological validity: One weakness of the SLT is that because the theory is based in research from the laboratory is lacks ecological validity.  The research was carried out in an artificial environment.  For example, the Bobo doll was designed to be hit so it is no real surprise that it was, and does not mean that children would necessarily hit a real person.  This suggests that the findings from this research could not be applied to real life situations

Reductionist: A second weakness of SLT is that it is oversimplistic.  Evidence from Flanagan (2000) suggests that testosterone has been cited as a primary cause of aggression and other genetic and neuroanatomical structures are involved.  SLT does not stress the importance of biological factors and relies on learning and the environment.  This suggests that SLT is reductionist when explaining aggressive behaviour.

Demand characteristics: A third weakness of Bandura’s research is that the children experienced demand characteristics They said that when they were in the experiment they felt they were expected to act aggressively towards the Bobo doll.  This suggests that there are methodological flaws with the research.     

Ethics: Final weakness of of Bandura’s research are ethical problems.  The experiment conducted was unethical and morally wrong because the children were encouraged to be aggressive.  Thus, the children did not leave the experiment as they entered it.  This suggests that reliable, well controlled research may lead to the exploitation of children.



People usually refrain from acting in an aggressive manner because they are easily identifiable and because they belong to societies that have strong norms against aggressive and other forms of antisocial behaviour. Therefore, people are usually in an individuated state (where they believe they are under scrutiny by others and so must act in a socially acceptable manner). However, in certain situations, such as large crowds, these restraints on aggressive behaviour can become relaxed, so people are more likely to engage in aggressive and antisocial behaviour. When people become faceless and anonymous – as in large crowds – they may enter a deindividuated state, where their behaviour is based on more primitive urges and does not conform to society’s norm.

Diener (1980) said that deindividuation occurs when self awareness is blocked by environmental events.

Critical factors include :

  • Strong feelings of group membership
  • Increased levels of arousal
  • Focus on external events
  • Feeling of anonymity


Prentice-Dunn and Rogers (1982) modified Diener’s theory to distinguish between:

  • Public self awareness – concern over the impression of yourself you are presenting to others when you are aware of being judged.
  • Private self awareness – your sense of self, consisting of thoughts, feelings, values and internal standards of behaviour.
Supporting Research

Zimbardo (1969) explored deindividuation in female undergraduates. Group 1 dressed in white lab coats with hoods over their faces. Group 2 wore large name tags. All pps observed a woman being interviewed and evaluated her performance by administering electric shocks. Condition 1 – pleasant interviewee, condition 2 – obnoxious. Group 2 shocked the obnoxious interviewee more than the pleasant one. Group 1 (deindividuated) shocked both interviewees equally. Zimbardo concluded that deindividuation increased aggression, making it indiscriminate and not at all influenced by individual characteristics.

Ellison et al. (2005) Driving simulation experiment  with 289 psych student participants. Measured aggressive driving (speed, jumping red lights, collisions etc.) in tops up / tops down conditions. More aggression shown in tops up (anonymous) condition.

Rehm et al. (1987) looked at aggression in handball. Deindividuation  was created by giving one team orange shirts, whilst other team wore own clothes. In boy teams, uniformed teams were more aggressive than non-uniform. In girl teams, no differences found. Researchers concluded that uniform > loss of individuality > deindividuation.
Watson (1973) tested the hypothesis that warriors who singificantly changed their appearance (i.e. entered a deindividuated state) when going to war would be more likely to torture, mutilate and kill their victims, compared to warriors who did not change their appearance. Watson looked at 23 societies. The results of the analysis showed that, of all 13 societies that showed high levels of aggression towards their victims (that is, they were more likely to kill, torture or mutilate them), all but one of these significantly changed their appearance prior to battle. Of the 10 societies who showed low levels of aggression towards their victims, seven of these did not significantly change their appearance (that is, they were less deindividuated prior to battle).
Issues & Debates

Lack of research support: A meta-analysis of 60 studies of deindividuation (Postmes and Spears, 1998) found insufficient support for the main claim that antisocial and aggressive behaviour is more common in large groups or in anonymous settings.

Gender differences: Research suggests that males and females may not respond in the same way when deindividuated. For example, Cannavale et al. (1970) found that males and females tend to respond differently when under conditions of anonymity. An increase in aggression under such conditions was evident in the behaviour of males, but not in females. Thus, evidence suggests that males may be more prone to losing inhibitions concerning aggressive behaviour when in a deindividuated state than females.

Real life application: Describing football violence as the product of mass deindividuation appears to be an oversimplification. Marsh et al. (1978) found that what might appear to be an undisciplined mob can actually consist of several different groups of supporters, each with their own status. They discovered that by serving an ‘apprenticeship’ of organised aggression over time, young supporters can be ‘promoted’ into a higher group and can thus continue a ‘career’ of football violence. In most cases this behaviour is highly ritualised, rather than physically violent.

Application to violence in Northern Ireland: Silke (2003) examined the relationship between anonymity (a key feature that contributes to deindividuation) and aggression in violent assaults that occurred in Northern Ireland over a 30-month period. Information was gathered through media reports, newspapers and from a victim support group.  Of the 5000 violent attacks analysed as part of this study, 206 were carried out by anonymous individuals who wore disguises to mask their true identities (e.g. balaclavas or paramilitary uniforms). There was a significant positive relationship between the use of disguises and different measures of aggression. Disguised offenders were more likely to inflict more serious physical injuries, attack more people at the scene, engage in more acts of vandalism, and were more likely to threaten victims after the attacks.


Extension Resources

Here are some