The limbic system, serotonin and testosterone

• The limbic system — the amygdala, hippocampus and hypothalamus. The amygdala is central: it evaluates the emotional importance of sensory information, and how it responds determines whether a threat is met with aggression. Gospic et al. (2011) put participants in an fMRI scanner while they played the Ultimatum Game; when participants rejected an unfair offer (a mild aggressive reaction), there was a heightened response in the amygdala. A benzodiazepine given beforehand halved the number of rejections and reduced amygdala activity — supporting a causal link.
• Serotonin — normally has an inhibitory effect, calming neuronal firing in the orbitofrontal cortex and so exerting self-control. Low serotonin removes this brake, so impulsive, aggressive behaviour becomes more likely. Virkkunen et al. (1994) compared levels of a serotonin breakdown product (5-HIAA) in the cerebrospinal fluid of violent impulsive and non-impulsive offenders and found levels were significantly lower in the impulsive group.
• Testosterone — an androgen that produces male characteristics, including aggression. It acts on brain areas implicated in aggression from early adulthood onwards. Dabbs et al. (1987) measured salivary testosterone in prison inmates and found those with the highest levels had a history of primarily violent crime, while those with the lowest had committed non-violent crimes.

Evaluation: the evidence for a serotonin link is strengthened by drug studies — drugs that increase serotonin also reduce levels of aggressive behaviour, which supports a causal, not merely correlational, role. The testosterone research is far less secure: Carré and Mehta's dual-hormone hypothesis proposes that testosterone only leads to aggression when cortisol is low, and that high cortisol blocks the effect — so the relationship is not simple. Much of this research is also correlational and much of it is done on non-human animals, whose aggression is far less mediated by cognition. Finally, the biological approach is reductionist: it cannot explain why the same person is aggressive in one context and not another.

Genetic factors and the MAOA gene

Twin and adoption studies point to a genetic component. Adoption studies find that aggressive behaviour correlates more strongly with biological than adoptive relatives.

• The MAOA gene — codes for the enzyme monoamine oxidase A, which breaks down neurotransmitters, including serotonin, after they have transmitted their impulse. A low-activity variant (MAOA-L, sometimes sensationally called the 'warrior gene') produces less of the enzyme, so serotonin metabolism is disrupted.
• Brunner et al. (1993) studied 28 men from a large Dutch family who had a history of impulsive, violent criminal behaviour including rape, attempted murder and arson. All had abnormally low levels of MAOA and carried the low-activity variant of the gene.
• Gene-environment interaction — Caspi et al. (2002) studied over 500 male children from a longitudinal cohort and found that those with the low-activity MAOA variant were significantly more likely to show antisocial behaviour only if they had also been maltreated as children. Those with the low-activity gene who had not been maltreated, and those with the high-activity variant who had been maltreated, did not show heightened antisocial behaviour. It is the interaction that matters.

Evaluation: Caspi's finding is the single most important thing to quote — it kills the simplistic 'gene for aggression' idea and provides a textbook diathesis-stress account. Genetic research is also complicated by the problem of measuring aggression: some studies use convictions (which capture only aggression that has been detected and prosecuted), some use self-report, and these produce different results — which is why concordance figures vary so widely between studies. Aggression is polygenic, and MAOA can only be a small part of the story.

Interpret Caspi

Ethological and evolutionary explanations

The ethological explanation studies animal behaviour in its natural setting, and assumes aggression is adaptive: it spreads members of a species out over a wider area, reducing competition for resources, and it establishes dominance hierarchies which give the dominant male access to mates.

• Ritualistic aggression — most aggressive encounters between members of the same species consist mainly of threat displays (baring teeth, displaying claws) and end with an appeasement display from the loser, which inhibits further attack. This prevents the death of a conspecific, which would be evolutionarily wasteful. Lorenz observed that fights between animals of the same species produced very little actual physical damage.
• Innate releasing mechanisms (IRMs) and fixed action patterns (FAPs). An IRM is an inbuilt neural network which, when triggered by a specific sign stimulus, releases a FAP — a stereotyped, universal, ballistic (once started, it runs to completion), unaffected-by-learning sequence of behaviours. Tinbergen (1951) found that a male stickleback would attack any wooden model with a red underside — including a completely unrealistic shape — but would not attack a realistic model that lacked the red spot. Once triggered, the aggressive FAP always ran its full course.

Evolutionary explanations of human aggression — aggression is adaptive where it solves an adaptive problem.

• Sexual jealousy and paternity uncertainty. A male can never be certain of paternity, and risks the enormous cost of investing in another man's child (cuckoldry). Aggression is therefore an anti-cuckoldry strategy. Wilson and Daly (1996) identified mate retention strategies including direct guarding (checking on a partner's whereabouts, monitoring their behaviour) and negative inducements (threats of violence, 'I'll kill myself if you leave'). Women who reported these behaviours in their partners were twice as likely to have suffered violence at their hands, and 73% of those had required medical attention.
• Bullying — Volk et al. argue that bullying characteristics (dominance, strength) are attractive to the opposite sex and deter rivals, so bullying is naturally selected.

Evaluation: Tinbergen's work is elegant and much replicated, but generalising from sticklebacks to humans is a serious stretch: human aggression is flexible, learned and cognitively mediated, not a ballistic FAP — which is why even Lorenz's claim that aggression is ritualistic and rarely lethal is contradicted by the evidence that humans routinely kill members of their own species. Nisbett's 'culture of honour' research found that white males from the southern USA were far more likely to become aggressive after an insult than northern males — a cultural difference the ethological account cannot explain, but which shows learning can override any innate mechanism.

What is a FAP?

The frustration-aggression hypothesis

Dollard et al. (1939) — 'aggression is always a consequence of frustration, and frustration always leads to aggression'. When our attempt to achieve a goal is blocked, we experience frustration, which creates an aggressive drive. Aggression is cathartic: expressing it relieves the drive.

• But the source of the frustration may be unavailable (too powerful, or abstract — you cannot attack the tax office), so aggression is displaced onto a substitute target that is available and cannot retaliate.
• Environmental cues. Berkowitz revised the theory: frustration creates only a readiness for aggression, and the presence of aggressive environmental cues makes aggression far more likely. In Berkowitz and LePage (1967), students were given electric shocks by a confederate, creating anger, and then had the opportunity to shock him back. The number of shocks given was higher when there were guns on the table next to the shock machine than when there were badminton racquets or no objects at all — the 'weapons effect'.
• Geen (1968) had participants complete a jigsaw. Frustration was manipulated: the puzzle was impossible; a confederate interfered; or the confederate insulted them. All three groups delivered more intense shocks than a control group, and the insulted group gave the most.

Evaluation: the strong original claim is simply wrong: frustration does not always lead to aggression (it may lead to despair, determination or withdrawal), and aggression does not always follow frustration. The idea of catharsis is also poorly supported — Bushman (2002) found participants who vented their anger by hitting a punchbag actually became more angry and aggressive afterwards, not less; venting was 'like using gasoline to put out a fire'. Berkowitz's reformulation — that frustration is only one of many aversive stimuli that produce negative affect, and that negative affect is what causes aggression — is a much better account.

Social learning theory and de-individuation

Social learning theory (Bandura). Aggression is learned by observing models and by vicarious reinforcement — seeing aggression rewarded. Learning requires the mediational processes: attention, retention, motor reproduction and motivation. Self-efficacy matters too: a child whose aggression has worked in the past develops confidence in their ability to be aggressive successfully, which makes them more likely to be aggressive again.

The Bobo doll studies. In Bandura et al. (1961), children who watched an adult attack a Bobo doll subsequently imitated the specific aggressive acts they had seen; children in a non-aggressive model condition showed almost none. In Bandura (1965), children saw the model rewarded, punished, or receive no consequence. Those who saw the model punished imitated least — but when all the children were later offered a reward for reproducing the model's behaviour, they all did so equally well. This is the crucial distinction between learning and performance: everyone had learned the aggression; the consequences only affected whether they chose to perform it.

De-individuation — a psychological state in which we lose our sense of individual identity and personal responsibility, typically in a crowd or when anonymous. The normal restraints on aggressive behaviour — guilt, shame, fear of being judged — are removed. Zimbardo (1969) distinguished individuated behaviour (rational, in line with personal norms) from de-individuated behaviour (emotional, impulsive, irrational). De-individuation is increased by anonymity (uniforms, masks, darkness, online usernames) and by reduced private self-awareness.

• Dodd (1985) asked 229 undergraduates 'What would you do if you could be totally invisible and had no fear of being caught?' Independent raters classified the responses: 36% were antisocial, and 26% were actually criminal — with 'rob a bank' the most common answer. Only 9% were prosocial. Anonymity, it seems, releases antisocial impulses.
• Douglas and McGarty (2001) found a strong correlation between anonymity and 'flaming' (posting hostile messages) in online chatrooms — the most aggressive messages came from those who chose to hide their real identity.

Evaluation: SLT explains why aggression varies so much between cultures and between contexts, which biological explanations cannot. But the Bobo doll studies have low ecological validity — a Bobo doll is designed to be hit and cannot retaliate, so hitting it is not really aggression — and there is a strong risk of demand characteristics. On de-individuation, the theory is not always supported: de-individuation can produce prosocial behaviour too (a crowd at a religious rally), and Gergen et al. found that people put in a darkened room with strangers began to touch and kiss — de-individuated, but not aggressive. It is therefore the norms of the group, not the loss of identity as such, that determine what a de-individuated person does.

Learning vs performance

Aggression in prisons: importation and deprivation

• The dispositional (importation) model — Irwin and Cressey (1962). Prisoners are not blank slates: they bring with them — 'import' — the values, norms and attitudes of the subculture they came from. Many are young men from backgrounds where violence is a normal way of resolving disputes, so aggression inside prison is simply a continuation of aggression outside. DeLisi et al. (2011) studied 813 juvenile delinquents in Californian institutions and found that inmates who imported a set of negative dispositional features (childhood trauma, high anger, irritability, a history of substance abuse) were significantly more likely to engage in suicidal activity and serious violence than a control group of inmates with fewer of these features.
• The situational (deprivation) model — Clemmer (1958). Aggression is the product of the oppressive conditions of the institution itself. Prisoners are deprived of liberty, autonomy, goods and services, heterosexual relationships and security. These deprivations create psychological stress, which is expressed as violence. Steiner (2009) analysed inmate violence in 512 US prisons and found that inmate-on-inmate violence was more common in prisons with a higher proportion of female staff, more overcrowding, and more inmates in protective custody — all situational variables, independent of the inmates' individual characteristics.

Evaluation: both models have supporting evidence, which is itself the strongest argument for an interactionist account: the imported dispositions of the inmates and the deprivations of the regime combine. Evidence against a pure deprivation model comes from Hensley et al. (2002), who studied prisons in Mississippi that allowed conjugal visits and found that this made no difference to levels of aggressive behaviour — undermining the claim that deprivation of heterosexual relationships is a cause. The importation model has been criticised for offering little practical guidance: if aggression is imported, there is not much a prison can do about it, whereas the deprivation model suggests real reforms — reducing overcrowding, improving conditions.

Computer games, desensitisation, disinhibition and priming

The evidence on computer games is mixed and depends heavily on the method:

• Experimental studies — Bartholow and Anderson (2002) had students play either a violent (Mortal Kombat) or a non-violent (PGA Tournament Golf) game for 10 minutes, then carry out the Taylor Competitive Reaction Time Task, in which they could deliver a blast of white noise to an opponent. Those who had played the violent game selected significantly higher noise levels. Experiments establish causation but measure only trivial, short-term proxies for aggression.
• Longitudinal studies — Robertson et al. (2013) followed just over 1,000 New Zealanders from birth to age 26 and found that excessive television viewing in childhood and adolescence was associated with a greater likelihood of having a criminal conviction and of showing antisocial personality traits in adulthood. But such studies are correlational and cannot exclude confounding variables.
• Meta-analyses — Anderson et al. (2010) reviewed 136 studies and found exposure to violent video games was associated with increases in aggressive behaviour, thoughts and affect, and with decreases in empathy and prosocial behaviour, in both collectivist and individualist cultures — although the effect sizes are small, which is why the debate continues.

Three explanations of the media effect:

• Desensitisation — repeated exposure to violence reduces the normal physiological arousal (racing heart, higher blood pressure, anxiety) that violence would ordinarily produce. Once the emotional response is blunted, the usual inhibition against aggression is weakened. Repeated exposure also fosters a belief that violence is a normal way to resolve conflict.
• Disinhibition — most people hold a strong social norm against violence. Media that portrays violence as normative, justified or rewarded, and that minimises its consequences for the victim, removes those inhibitions. Cartoon violence, where no one is really hurt, is a classic disinhibitor for children.
• Cognitive priming — a violent film or game provides a script for aggressive behaviour, which is stored in memory. The script is triggered (primed) when the person later encounters cues in the environment that resemble the media content, and it guides their interpretation of the situation and their response — often automatically and without conscious awareness.

Evaluation: laboratory measures of aggression (blasting white noise, giving hot sauce) plainly lack validity — nobody is actually hurt, so we cannot say the studies measure real aggression. Field and longitudinal studies have the opposite problem: better validity, but correlational, so a third variable (an already-aggressive disposition, or a chaotic home) may cause both the media choice and the aggression. The most defensible conclusion is that media violence is one risk factor among many, with a small but real effect — and this is a socially sensitive conclusion that is routinely over-stated in the press.

Name the mechanism

Serotonin's role

Is catharsis real?

De-individuation's limit

Importation or deprivation?

Biological, social or media?

Tap a card, then the explanation it belongs to.

Study and finding

Tap an item on the left, then its partner on the right.

The big ideas to know

You have covered the whole of AQA 4.3.8 Aggression. Press Finish to see your score.