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

Social psychological explanations argue that aggression is not simply produced by brain chemistry or genes. Instead, it is shaped by the situations people find themselves in and by their interactions with other people. This page covers two of these explanations. The frustration-aggression hypothesis explains aggression as a response to having our goals blocked. Social learning theory explains aggression as a behaviour we learn by watching other people and seeing what happens to them.

The frustration-aggression hypothesis

Key terms

• Frustration: the state that results when a person's attempt to achieve a goal is blocked by some external factor.

• Aggressive drive: the psychological tension created by frustration, which pushes a person towards aggressive behaviour.

• Catharsis: the release of built-up aggressive drive through an aggressive act, which reduces the drive.

• Displacement: redirecting aggression away from the true source of frustration onto a weaker or more available target.

• Aggressive cues: features of the environment, such as weapons, that are associated with aggression and make an aggressive response more likely (the weapons effect).

The original hypothesis

Dollard et al. (1939) proposed that frustration always leads to aggression, and that aggression is always the result of frustration. Frustration occurs when a person's attempt to achieve a goal is blocked by an external factor. For example, a student who has revised for weeks but is refused entry to an exam because their train was cancelled has had an important goal blocked.

According to Dollard et al., frustration creates a psychological tension called an aggressive drive. This drive builds up and must be released, rather like pressure building up in a hydraulic system. The drive therefore leads to aggressive behaviour, and the hypothesis presents this as an almost automatic process rather than a conscious choice.

Dollard et al. combined Freud's psychodynamic idea of an aggressive drive that needs to be released with the language of learning theory. The hypothesis predicts that frustration is stronger the more a goal is wanted and the closer a person is to achieving it, so being pushed out of a queue just before reaching the front should be more frustrating than being pushed out near the back. Harris (1974) supported this in a field experiment: when confederates pushed into queues, people reacted more aggressively when the queue-jumper pushed in front of the 2nd person in line than the 12th. Aggression is also more likely if the person believes it will help them reach the goal, and if aggression has been rewarded in the past, for example by removing the obstacle, it becomes the dominant (most likely) response to future frustration.

Aggression is cathartic and can be expressed indirectly

Dollard et al. argued that acting aggressively is cathartic: releasing the aggressive drive reduces the feelings of frustration and so reduces the drive itself. Once the tension has been released, the person is less likely to behave aggressively again in the short term.

However, aggression is not always directed at the source of the frustration. There are three main reasons for this:

• The cause may be abstract, such as the economy or "the system", so there is no person to attack.

• The cause may be too powerful to confront directly, such as a boss or a teacher, and attacking them risks punishment.

• The cause may be unavailable at the moment the person wants to act, for example because they have already left.

In these situations the aggression is displaced onto a weaker and more available target, sometimes called a scapegoat. For example, a student who receives a disappointing grade may shout at a younger sibling when they get home, because the sibling is a safer target than the teacher who gave the grade.

The weapons effect

Berkowitz (1989) revised the hypothesis. He argued that frustration does not automatically cause aggression. Instead, frustration creates a readiness for aggression, and whether that readiness becomes actual aggression depends on cues in the environment. Objects associated with aggression, such as guns, act as aggressive cues that make an aggressive response more likely in someone who is already angry or frustrated.

Key study: Berkowitz and LePage (1967)

Aim: To test whether aggressive cues in the environment increase aggression in people who have been angered.

Procedure: Male university students took part in what they were told was a study of stress. A confederate gave each participant either one electric shock (low anger) or seven electric shocks (high anger) as "feedback" on a task. Participants were then given the chance to shock the confederate in return. On a table nearby there was either a shotgun and a revolver, a pair of badminton racquets, or nothing.

Findings: Participants who had been angered gave significantly more shocks when the guns were on the table than when there were badminton racquets or nothing. For example, angered participants gave an average of about 6 shocks when the guns were said to belong to the confederate, compared with about 4.7 when there was nothing on the table.

Conclusion: Environmental cues such as weapons act as catalysts that stimulate aggressive behaviour. Anger creates a readiness to aggress, and the presence of an aggressive cue makes it more likely that this readiness is expressed as aggression. As Berkowitz put it, "the finger pulls the trigger, but the trigger may also be pulling the finger".

Key study: Geen (1968)

Aim: To investigate how different sources of frustration influence aggressive behaviour.

Procedure: Male university students were asked to complete a jigsaw puzzle. Frustration was manipulated in one of three ways. In the first condition the puzzle was impossible to solve. In the second, participants ran out of time because a confederate kept interfering. In the third, participants solved the puzzle but were then insulted by the confederate. A control group was not frustrated at all. All participants were then given the opportunity to deliver electric shocks to the confederate.

Findings: Insulted participants gave the strongest shocks. All three frustrated groups gave more intense shocks than the non-frustrated control group.

Conclusion: Frustration increases aggression, and interpersonal provocation (being insulted by another person) increases it most.

Evaluation

A key strength is research support for displaced aggression. Marcus-Newhall et al. (2000) carried out a meta-analysis of studies of displaced aggression. They found that participants who were provoked but unable to retaliate against the true source of their frustration were significantly more likely to aggress against an innocent person than participants who had not been provoked, and concluded that displaced aggression is a reliable phenomenon. This supports the hypothesis because displacement is one of its central claims: frustrated people redirect their aggression onto a weaker or more available target when they cannot attack the real cause. Because the conclusion is based on many studies rather than one, it is unlikely to be a chance finding or the product of one unusual sample. This suggests that the displacement element of the hypothesis is well supported, although it does not show that every part of the explanation is correct.

A major limitation is that aggression may not be cathartic. Bushman (2002) angered participants and then asked some of them to hit a punchbag while thinking about the person who had angered them. Rather than calming down, these participants became more aggressive afterwards than participants who did nothing. Bushman concluded that venting anger is like "using gasoline to put out a fire" (gasoline is the American word for petrol). This contradicts the cathartic element of the hypothesis, which claims that releasing aggression reduces the aggressive drive. If acting aggressively increases rather than reduces the drive, the hydraulic idea at the heart of Dollard et al.'s explanation appears to be mistaken. This casts doubt on the validity of one of the hypothesis's central assumptions, and it also suggests that advice to "let off steam" by hitting cushions or punchbags may be counter-productive.

A further limitation is that the link between frustration and aggression is more complex than the original hypothesis claims. Frustration does not always lead to aggression, and aggression can occur without frustration. A frustrated person may instead become helpless and give up, or become more determined to succeed. Equally, someone may behave aggressively for reasons that have nothing to do with frustration, such as a planned robbery carried out for financial gain. How a person interprets the frustration also matters. Dill and Anderson (1995) found that participants who experienced unjustified frustration were more aggressive than those who experienced justified frustration, although justified frustration still produced more aggression than none. This suggests that cognitive appraisal, meaning how the person evaluates the situation, plays a crucial role. The original "always/always" hypothesis therefore appears inadequate, because it can explain aggression in some people and situations but not in others.

A strength is that the original hypothesis was reformulated to fit the evidence. In response to findings that frustration does not always lead to aggression, Berkowitz (1989) argued that frustration is just one of many aversive (unpleasant) stimuli that create negative feelings. Aggression is triggered by negative feelings in general rather than by frustration specifically, and the outcome of frustration can be a range of responses, of which aggression is only one. This is a strength because it shows the flexibility of the explanation and reflects how science should operate: when evidence appears that a theory cannot explain, the theory is adapted rather than abandoned. The revised version can account for evidence such as Dill and Anderson (1995) and Berkowitz and LePage (1967), which suggests that the frustration-aggression hypothesis is best judged in its revised form rather than in its original, more rigid form.

A further strength is that the research has real-life applications. Berkowitz's argument that "the trigger may be pulling the finger" has featured in the gun-control debate in the USA. Some states allow "open carry", where a gun does not have to be concealed. According to the weapons effect, the visible presence of a gun acts as an aggressive cue, which may make its use more likely in a person who is already angry or frustrated. Berkowitz and LePage (1967) showed this effect under controlled conditions. This suggests that removing aggressive cues from the environment could help to reduce gun-related violence. However, the original evidence comes from a laboratory study using electric shocks rather than real violence, so it would be too strong to conclude that open carry directly causes more violence. The application is best seen as a plausible implication of the research rather than a proven effect.

Social learning theory

Key terms

• Observational learning: learning a behaviour by watching another person (a model) perform it.

• Vicarious reinforcement: learning indirectly from the consequences of another person's behaviour; seeing a model rewarded for aggression makes imitation more likely.

• Vicarious punishment: seeing a model punished for aggression, which makes imitation less likely.

• Identification: being influenced by a model because the observer wants to be like them, for example because the model is similar, high in status or likeable.

• Self-efficacy: the extent to which a person believes their actions will achieve a desired goal.

Direct and indirect learning

Bandura argued that aggression is learned rather than being an innate drive, and that it can be learned in two ways.

• Direct learning happens through operant conditioning. If a child behaves aggressively and is rewarded, for example by getting the toy they wanted, the behaviour is reinforced and is more likely to be repeated. If the aggression is punished, it becomes less likely.

• Indirect learning happens through observing other people. Bandura argued that this indirect route accounts for most of the aggressive behaviour people learn, because children see far more aggression than they are ever directly rewarded for.

Observational learning and vicarious reinforcement

Children acquire specific aggressive behaviours by observing aggressive models, such as parents, siblings, peers and characters in the media. They learn not only how to aggress, but also what happens when people do. If a model's aggression is rewarded, or simply goes unpunished, the observing child learns that aggression can be effective. This is vicarious reinforcement, and it makes the child more likely to imitate the behaviour. If the model is punished, the child is less likely to imitate it because of vicarious punishment.

A child is more likely to imitate a model they identify with: someone similar to them in sex or age, or someone who is high in status, powerful, attractive or likeable. Over time the child may internalise the model's aggressive behaviour, adopting it as part of their own way of behaving. For example, a boy who watches his popular older brother push to the front of a queue at a theme park without being told off has observed a similar, high-status model being rewarded. He may later push to the front of the lunch queue at school.

Cognitive conditions for learning

Bandura argued that learning is not automatic. Four cognitive (mediational) processes must take place between observing a behaviour and imitating it:

• Attention: the observer must pay attention to the model's aggressive actions. A child who is distracted will not learn the behaviour.

• Retention: the observer must form a symbolic mental representation of the behaviour and store it in memory, so that it can be recalled later.

• Reproduction: the observer must be able to translate the mental representation into actual physical action. This involves mentally appraising their own ability: a small child may remember a punch but judge that they are not strong enough to use it.

• Motivation: the observer must expect that behaving aggressively in a particular situation will be rewarding. This expectation, often learned through vicarious reinforcement, provides the motivation to imitate.

Because of these processes, learning and performance are separate. A child can learn an aggressive behaviour by observation without showing it straight away; the behaviour is stored and may only be performed later, when the child expects it to be rewarded.

Self-efficacy

Self-efficacy is the extent to which a person believes their actions will achieve a desired goal. A child who uses aggression and succeeds, for example by snatching another child's pens and keeping them, learns both that aggression works and that they are good at it. With each successful outcome their confidence in their ability to aggress grows, so they become more likely to use aggression again. A child who tries aggression and fails develops low self-efficacy for aggression and is less likely to use it.

Key study: Bandura et al. (1961)

Aim: To find out whether children would imitate aggressive behaviour they had observed in an adult model, even when the model was no longer present.

Procedure: 72 children (36 boys and 36 girls) aged between about three and almost six (37 to 69 months), from the Stanford University nursery school, took part. One group watched an adult model behave aggressively towards a large inflatable Bobo doll, for example hitting it with a mallet, kicking it and shouting phrases such as "sock him in the nose". A second group watched a non-aggressive model play quietly with other toys, and a control group saw no model. Half of the children saw a model of the same sex and half saw a model of the opposite sex. To create a degree of frustration, each child was then allowed to play with some attractive toys for about two minutes before being told that these toys were reserved for other children. Finally, each child was taken to a room containing a Bobo doll and other toys, and their behaviour was observed through a one-way mirror.

Findings: Children who had watched the aggressive model reproduced many of the same physical and verbal aggressive acts, often using the same words. Children who saw the non-aggressive model or no model showed very little imitative aggression. Boys imitated more physical aggression than girls, but there was little difference between boys and girls in imitating verbal aggression. Boys were also more likely than girls to imitate a male model.

Conclusion: Children learn aggressive behaviour by observing and imitating models, and identification with a same-sex model increases imitation.

Evaluation

A key strength of SLT is research support from children's friendships. Poulin and Boivin (2000) found that boys aged 9 to 12 who used proactive aggression tended to choose other proactively aggressive boys as friends. Such friendships are thought to act as "training grounds" for antisocial behaviour: the boys are frequently exposed to models of aggression (each other), can observe each other successfully using aggression to get what they want, and receive rewarding approval from the rest of the group when they behave aggressively themselves. This supports SLT because it suggests that observation, vicarious reinforcement and direct reinforcement all operate in real children's lives, not just in a laboratory. However, the study shows that aggressive boys choose similar friends; it does not by itself prove that the friends caused each other's aggression, so it is best seen as consistent with SLT rather than as conclusive proof. More generally, laboratory research such as Bandura et al. (1961) can show under controlled conditions that modelling causes imitation, but cause and effect are much harder to establish for real-life aggression, and the mediational processes themselves cannot be observed directly; they can only be inferred from behaviour, so their influence is difficult to measure.

A further strength of SLT is its real-world application to reducing aggression. Huesmann and Eron (2013) argued that media portrayals of aggressive behaviour can be powerful influences on a child's acquisition of aggression, especially when the media character is rewarded for being aggressive. This is vicarious reinforcement, exactly as SLT predicts. If the same learning processes produce aggression, they can also produce non-aggression. One way to reduce aggression is therefore to provide non-aggressive models in the media who are rewarded for resolving conflicts peacefully, and to encourage children to form friendships with peers who do not routinely behave aggressively. This suggests that SLT is not only an explanation but also a source of practical strategies that parents, schools and policymakers can use, which increases its value.

A limitation is that SLT cannot explain all forms of aggression. Children who use proactive aggression tend to have high self-efficacy: they are confident and use aggression deliberately to achieve goals, such as taking a toy or gaining status. This is well explained by SLT, because the aggression has been learned and is used when a reward is expected. However, children who show reactive aggression are hostile, suspicious of others and lash out angrily in the heat of the moment in response to a perceived provocation. They are not trying to achieve anything except retribution. This impulsive, emotional behaviour is much harder to explain from a social learning perspective and may be better explained by the frustration-aggression hypothesis or by biological factors such as amygdala reactivity. SLT is therefore best seen as an explanation of how planned aggression is acquired rather than a complete account of all aggression.

A further limitation is that SLT may not fully explain cultural differences in aggression. Different cultures have different norms about which behaviours are encouraged. SLT predicts that in a culture where aggression is discouraged and not valued, such as the !Kung San people of the Kalahari, aggressive models will be rare, vicarious reinforcement of aggression will be uncommon, and aggression should therefore be very low. However, aggression still occurs in such societies. This suggests that there is more to aggressive behaviour than social learning, and that some aggression may have an instinctive or biological basis that occurs even when it is not modelled or rewarded. The cultural evidence therefore offers only partial support for SLT.

A limitation is that SLT underestimates the influence of biological factors on aggression. Bandura did recognise that biology plays a role, accepting that there may be an instinctive urge to be aggressive, so it would be inaccurate to say that SLT completely ignores biology. However, he emphasised that the specific form aggression takes is learned and is the result of nurture. In contrast, research shows that genetic, evolutionary, neural and hormonal factors can drive and regulate aggression; for example, the low-activity MAOA gene and high testosterone are both linked to aggressive behaviour. SLT barely acknowledges these physiological factors and offers no account of how they operate. Children who grow up with aggressive parents may also inherit an aggressive temperament from them, so the same evidence could support a genetic explanation. This suggests that SLT provides an incomplete explanation, and that an interactionist approach combining learning with biology is likely to be more valid.

Overall: both explanations locate the causes of aggression in the social world rather than in biology alone, but they explain different kinds of aggression. The frustration-aggression hypothesis, especially in Berkowitz's revised form, is best at explaining impulsive, reactive aggression triggered by blocked goals, negative feelings and environmental cues, although its claim that aggression is cathartic is not supported. Social learning theory is best at explaining how planned, proactive aggression is acquired through observation, vicarious reinforcement and self-efficacy, and it leads to useful practical applications. Neither explanation is complete on its own, and each is strongest where the other is weakest, which suggests that a full account of human aggression needs both, alongside biological factors.

Exam tips

• In frustration-aggression essays, use evidence as evidence. Examiners praised answers that used Geen and Harris "to good effect" and discussed Berkowitz's weapon cues as a modification.

• Avoid circular descriptions such as "frustration causes aggression because people get frustrated". Give detailed accounts of goal-blocking, catharsis and displacement.

• Keep SLT distinct from operant conditioning. Use vicarious reinforcement, identification and the mediational processes.