Studies of the levels of violence on television showed 7.5 violent acts per hour on US prime time television, 8.6 for German entertainment programmes, 5.8 for Dutch dramatic fiction and 2.5 for UK prime time TV.
A US child watching 2-4 hours per day (the norm) will have witnessed 8,000 homicides and 100,000 other acts of violence by age 18.
A study of the 95 most popular US video games showed 83% featured violent themes. Most games had two simple choices: kill or die. A 1993 American Psychological Association report on screen violence concluded: "In addition to increasing violent behaviours towards others, viewing violence on television ... increases viewers' appetites for becoming involved with violence."
In 1995, the extensive US Cable industry National Television Violence Study (NTVS) stated as its key finding: "there are substantial risks of harmful effects from viewing violence throughout the television environment."
The UK Gulbenkian Commission on Children and Violence, also in 1995, sat on the fence, quoting Cumberbatch of Aston University, who argued: "supposed evidence of a link between television violence and violence in real life is based on studies which are, for the most part, individually fatally flawed and collectively self-contradictory."
This article addresses two main issues:
- "is there a causal link?" and,
- "if so, how important is it?"
Is there a causal connection?
There appears now to be widespread acceptance that a correlation between screen violence and aggression is proven. Wartella (1995), after an in-depth review of studies from around the world, summarises: "More than a thousand studies in the United States and dozens within Europe have been devoted to this topic ... Distilling decades of laboratory, survey and field experimental studies, the current reviews conclude that there is a correlation between violence viewing and aggressive behaviour, a relationship that holds even when a variety of controls are imposed (e.g. age of subject, social class, education level, parental behaviour, attitudes towards aggression)..."
The National Research Council (1993), cautioned that correlations do not prove cause, and could reflect poor parental supervision causing both greater exposure to television violence and a heightened potential for violent behaviour. In 1978 one of the largest European studies, by Belson, found a relationship between viewing media violence and serious, criminal behaviour by adolescent boys. He found the evidence "very strongly supportive of the hypothesis that high exposure to TV violence increases the degree to which boys engage in serious violence".
Wartella asserts Belson has not proved a causal link, but is convinced by Eron and colleagues' longitudinal study findings (1982, 1984) that boys' viewing of television violence at age 8 predicted aggressive behaviour at age 18, and serious criminal behaviour at age 30. She found particularly convincing that similar longitudinal studies in Australia, Finland, Poland and Israel supported the conclusion that "viewing televised violence leads to aggressive behaviour and not vice versa."
Another review, by Jo and Berkowitz (1994), concluded: "there is no longer a question as to whether the portrayal of violence in the mass media can increase the chances that some people in the audience will act aggressively themselves. Such an effect can occur and often does..."
Caprara and Rutter (1995) and Geen (1994) point to the robustness of evidence from laboratory experiments, field experiments, longitudinal studies and archival studies, all supporting the conclusion that observing television violence facilitates subsequent aggression. Geen summarises "These studies have involved children, adolescents and young adults, and a wide range of constrained and unconstrained behaviours ... the large number of studies reporting the effect and the convergence of data from so many types of investigation indicates that the effect is a real one".
How important is it?
Some commentators argue that, while there is an effect, it is negligible.
Geen says: "Admittedly, effect sizes are sometimes small, as critics have pointed out." Caprara and Rutter state effects are "... fairly weak ones overall". Comstock (1990) analysed 22 different surveys and found effect sizes varied between 5 and 15 per cent.
Eysenck and Nias (1978) suggest a model which assumes pre-disposition to violence in a country is distributed in the shape of a normal curve of distribution, with 5% of people actually violent. Their model suggests a 1% shift in propensity to violence would increase the number of violent persons in the UK by 350,000, and in the USA by 1.4 million; a 5% shift effect would mean 2.25m additional violent people in the UK and 9 million in the USA.
Eysenck and Nias warn their model is over-simplified, because human behaviour is multiply determined. However, they conclude that their model demonstrates: "Even quite small effects, so small as to be hardly measurable, may have tremendous and far-reaching consequences."
Who is affected?
German and Swedish studies show children from lower socio-economic groups watch significantly more than average levels of television. Wartella's literature review comments: "Children who are more disposed to violence (such as those living in violent homes and environments) are more likely to be influenced by violent portrayals in the media."
Caprara and Rutter state that children vary in their susceptibility to media influences, with effects tending to be greatest when children already experience violence. The children most likely to be affected are the most vulnerable.
Huesmann put forward a reciprocal model in which some personal factors lead to greater TV watching and aggressiveness, and hence to a greater interest in TV violence. Viewing violent programmes then strengthens encoding of aggressive scripts, further increasing the likelihood of aggressive reactions to interpersonal conflict. This leads to greater identification with violent TV characters and strengthens the interest in TV violence, so creating a vicious spiral.
Field research confirms that children and adults who are predisposed towards violence are more likely to respond to television violence. Small average effects may not be small for the already vulnerable.
That there is a correlation between screen violence and violent behaviour in real life is now demonstrated beyond doubt. Arguments about the direction of the effect miss the point; it probably operates in both directions, in an upward reinforcing spiral including poor initial parental control and/or family violence, high exposure to screen violence, selection of violent role-models, development of violent scripts, increasing levels of violent experiences and violent practices.