You are viewing a free preview of this lesson.
Subscribe to unlock all 10 lessons in this course and every other course on LearningBro.
Whether exposure to violent media — television, film and especially computer (video) games — increases real-world aggression has been one of psychology's most contested questions for over half a century, and one of the most socially and politically charged. This lesson examines the mechanisms proposed to explain media effects (desensitisation, disinhibition and cognitive priming, alongside observational learning), the three main kinds of research evidence — experimental, correlational and longitudinal — and a careful, balanced evaluation of a literature in which reputable researchers reach opposing conclusions. The topic is treated scientifically and clinically: we weigh effect sizes and methods rather than amplifying moral panic, and we keep firmly in view that "an effect on a laboratory aggression measure" is not the same as "a cause of serious violence." The unifying theme is that the evidence supports a small, real, but heavily moderated association, best understood within an integrative model rather than as a simple cause.
Key Definition: Media influences on aggression are the ways in which exposure to violent content (in television, film, computer games and online media) may increase aggressive cognitions, affect and behaviour in the audience, whether briefly (short-term priming/arousal) or cumulatively (long-term learning).
This lesson covers the AQA 7182 Paper 3 strands media influences on aggression, including the effects of computer games, and the explanations of media effects: desensitisation, disinhibition and cognitive priming. You must be able to describe (AO1) each mechanism and to outline experimental studies (e.g., Bartholow and Anderson), correlational studies, and longitudinal studies (e.g., Robertson et al.) of media and computer-game effects. You then convert this into AO3: the correlation-versus-causation problem, the ecological validity of laboratory aggression measures, publication bias and small effect sizes, the difficulty of controlling third variables, and the socially sensitive nature of the debate — integrating the strands via the General Aggression Model. The recurring examiner theme is methodological: the quality and convergence of the evidence, and what can legitimately be inferred from a small effect on a proxy measure, matter more than any single dramatic finding.
Several mechanisms have been proposed for how media violence could raise aggression. The specification highlights desensitisation, disinhibition and cognitive priming; observational learning (from Bandura's social learning theory) provides a fourth, longer-term route.
Cognitive priming proposes that exposure to violent media activates a network of aggression-related concepts in memory — hostile thoughts, angry feelings and aggressive behavioural scripts — which then remain temporarily accessible and bias how subsequent, often ambiguous, situations are interpreted and responded to. After violent content, a person may be quicker to read a neutral act as hostile, to recall conflict, and to respond aggressively to minor provocation.
Huesmann's script theory extends this: repeated exposure lays down and rehearses aggressive scripts — stored "if-then" routines for behaving in social situations — which become increasingly automatic and easily triggered. Priming thus has both a short-term (momentary accessibility) and a long-term (script acquisition) aspect, bridging to observational learning.
The priming idea is supported by experimental work showing that briefly exposing people to aggression-related media or cues speeds the recognition of aggressive words and increases the tendency to interpret ambiguous behaviour (such as an accidental bump) as hostile — a hostile attribution bias. This matters mechanistically because aggression often turns on how an ambiguous social act is interpreted: if violent media makes hostile interpretations more accessible, it raises the probability of an aggressive response without the person necessarily learning any new behaviour. Priming therefore offers a plausible route by which even brief media exposure can influence behaviour in the moments that follow, complementing the slower, cumulative route of script acquisition and observational learning.
Desensitisation proposes that repeated exposure to media violence progressively reduces the normal emotional and physiological response to violence. With habituation, the arousal and anxiety that violence would normally evoke diminish, and with them the empathy for victims and the inhibition against aggression that such responses support. Predicted consequences include reduced physiological arousal (e.g., lower heart rate and skin conductance) to violence, decreased empathy, greater tolerance of real-world aggression, and weakened restraint.
Evidence for the physiological aspect is reasonably direct. Studies measuring autonomic responses (heart rate and skin conductance) have found that people who have just played or who habitually consume violent media show blunted arousal when subsequently shown scenes of real violence, compared with controls — the predicted habituation. There is also evidence that this physiological blunting is accompanied by slower or weaker helping responses to a staged emergency, suggesting that reduced arousal translates into reduced prosocial responding. The mechanism is therefore not merely about attitudes but about a measurable dampening of the body's normal alarm response to violence, which connects desensitisation to the autonomic arousal systems studied in biopsychology.
Disinhibition refers to the weakening of the social and moral restraints that normally inhibit aggression. Where media repeatedly portrays violence as justified (used by the hero), effective (it achieves goals) and without lasting consequences (victims are not shown suffering), it can shift the viewer's social norms, so that aggression comes to seem a more normative and acceptable way to resolve conflict. Disinhibition concerns beliefs about the acceptability of aggression rather than the mere accessibility of aggressive thoughts (priming) or reduced emotional response (desensitisation).
The disinhibition account predicts that the context in which media violence is presented should matter: violence depicted as rewarded, morally justified and painless should disinhibit aggression more than the same violent act shown as punished, unjustified and causing visible suffering, because only the former conveys the norm that aggression is acceptable and effective. This prediction connects disinhibition to social learning theory's vicarious reinforcement (rewarded models are imitated more) and explains why "sanitised," consequence-free screen violence is regarded as more concerning than violence whose harmful consequences are made salient — a point with clear implications for how media is classified and depicted.
flowchart TD
A[Exposure to\nviolent media] --> B[Cognitive priming\nactivates aggressive\nthoughts + scripts]
A --> C[Desensitisation\nreduced arousal +\nempathy over time]
A --> D[Disinhibition\naggression seen as\nnormative + justified]
A --> E[Observational learning\nimitation of rewarded\nmodels]
B --> F[Increased likelihood\nof aggressive response]
C --> F
D --> F
E --> F
| Mechanism | Core process | Time course | Distinctive feature |
|---|---|---|---|
| Cognitive priming | Activates aggressive concepts/scripts | Short-term + long-term | Biases interpretation of ambiguous cues |
| Desensitisation | Reduces emotional/physiological response | Cumulative | Lowered arousal and empathy |
| Disinhibition | Shifts norms about acceptability | Cumulative | Aggression seen as justified/normative |
| Observational learning | Imitation of reinforced models | Long-term | Acquisition of new specific acts |
A strength of this field is that it draws on three converging methods, each with complementary strengths and weaknesses: experiments (which establish causation but in artificial settings), correlational studies (which use real-world measures but cannot show causation), and longitudinal studies (which track effects over time).
Bartholow and Anderson (2002) tested the short-term causal effect of violent computer games. Aim: to establish whether playing a violent game increases immediate aggression. Method: participants played either a violent game (Mortal Kombat) or a non-violent game (PGA Tournament Golf) for a short period, then completed the Taylor Competitive Reaction Time Task, in which they could set the intensity of an aversive noise blast delivered to an "opponent" — a standard laboratory proxy for aggression. Findings: participants who had played the violent game subsequently selected higher noise-blast intensities, and the effect was somewhat larger for men. Conclusion: in a controlled experiment, violent game play causally increased a laboratory measure of aggression in the short term — though the artificiality of the measure constrains real-world inference.
Such experiments are valuable because the random allocation and manipulated independent variable license a causal conclusion. Their key limitation is ecological validity: pressing a button to deliver noise is not interpersonal violence, and the effects are necessarily short-term.
A further experimental contribution comes from Fischer and Greitemeyer (2006), whose work is often cited in connection with the cognitive mechanisms (priming) and with the importance of media content rather than the medium as such. Aim: to test whether the specific content of media shapes aggression-related cognition and behaviour. Method: in studies of songs with aggressive (and, in related work, woman-demeaning) lyrics versus neutral lyrics, participants were exposed to one type and then measured on aggression-related thoughts, hostile perceptions and aggressive behaviour towards a target. Findings: exposure to the aggressive-content media increased aggression-related cognitions and behaviour relative to neutral content. Conclusion: it is the aggressive content that primes aggressive thoughts and feelings, supporting the cognitive-priming mechanism and indicating that media effects are content-specific rather than a blanket property of "media." Importantly, the same research programme showed the mirror-image effect for prosocial media — prosocial lyrics increased prosocial thoughts and helping — which strengthens the causal interpretation by demonstrating that content drives cognition in both directions.
Correlational designs measure naturally occurring media exposure and aggression and assess their association in real life. Large reviews — for example Anderson and Dill (2000) (whose correlational study linked self-reported violent-game play to real-life aggressive and delinquent behaviour) and the broader meta-analytic work of Anderson, Bushman and colleagues — report small but consistent positive correlations (typically around r = 0.15–0.25) between media-violence exposure and aggressive behaviour, thoughts and affect, and small negative correlations with empathy. Some authors have argued that an effect of this size, while small for an individual, is non-trivial at the population level because the number of people exposed is so large, and have controversially compared its magnitude to other accepted public-health associations. Critics reject this comparison (see Evaluation), but it illustrates why a "small" correlation is not automatically dismissible. These designs use real-world measures of aggression, but they cannot establish direction: aggressive individuals may simply choose more violent media (a selection effect), and unmeasured third variables (family violence, poverty, peer influence) may drive both.
Longitudinal designs follow the same individuals over time, which helps disentangle direction and capture cumulative effects.
Robertson, McAnally and Hancox (2013) used data from the Dunedin birth cohort, measuring television viewing time in childhood/adolescence and antisocial outcomes in adulthood. Findings: more time spent watching television in youth was associated with a greater likelihood of antisocial personality traits and criminal convictions in adulthood, an association that persisted after controlling for several confounds. Conclusion: consistent with a long-term link between heavy media exposure and later antisocial behaviour — though, as the authors note, amount of viewing is not the same as violent content, and residual confounding cannot be excluded.
Huesmann et al. (2003) followed children over roughly 15 years and found that childhood exposure to TV violence predicted adult aggression (including aggression in relationships and against others), even after controlling for socioeconomic status, ability and initial aggression. Longitudinal designs are powerful because the predictor is measured before the outcome; their limitations are practical (attrition, cost) and interpretive (confounds that change over time).
A useful refinement is the "downward spiral" model, which proposes that the relationship is reciprocal: aggressive children both seek out more violent media and are affected by it, so that media exposure and aggression reinforce one another over development. This matters for evaluation because it partly reconciles the "selection" criticism (aggressive people choose violent media) with the "media-effects" claim — both can be true simultaneously, with each strengthening the other. It also explains why longitudinal associations persist even when initial aggression is controlled: the effect of media is additional to, and interacts with, an existing disposition rather than being wholly explained away by it.
Exam Tip: Structure media-violence evaluation around the three methods: experiments give causation but lack ecological validity; correlations have real-world validity but cannot show causation; longitudinal studies bridge time but face confounds. Showing you understand this trade-off — and that the methods converge on a small effect — is worth more than listing studies.
The convergence of three methods on a small positive effect is the strongest argument that media influence is real. Experiments (Bartholow and Anderson, 2002) show short-term causation, correlational studies show real-world association, and longitudinal studies (Huesmann et al., 2003; Robertson et al., 2013) show prediction over time. This matters because each method's weakness is another's strength, so their agreement is harder to dismiss than any single study. The implication is that outright denial of any media effect is difficult to sustain; the serious debate is not whether an effect exists but how large, how durable and how practically important it is — which is a more defensible framing than either alarmism or dismissal.
Causation cannot be inferred from the (numerous) correlational studies, which is a fundamental limitation. Much of the real-world evidence is correlational, so a link between violent-media use and aggression is open to reverse causation — already-aggressive people choosing violent media (a selection effect) — and to third variables such as family violence, poverty and peer influence that could drive both. This matters because the headline real-world associations cannot, by themselves, show that media causes aggression. The implication is that the causal weight must rest on the experimental and longitudinal evidence, and even there confounds and artificiality temper the conclusion — so confident causal claims about serious violence outrun the data.
Subscribe to continue reading
Get full access to this lesson and all 10 lessons in this course.