Notes on Media Violence by Barrie Gunter (2008)
GUNTER, B. 2008. Media Violence. American Behavioral Scientist 51, 1061 –1122.
Some overall points
• Gunter paper is a review of behavioural research on media effects on violence - extensive and useful, goes over the different types of research, etc
• The most fundamental difficulty is that everyone is different, media effects get filtered through individual cognitive and cultural screens, "certain forms of media violence can exert certain kinds of effects on some media consumers some of the time." (p1113)
– Real or fiction: people more likely to respond and be affected by real violence (or violence said to be real, e.g. in experiments)
– Also legitimacy: when violence is shown as legitimate, people more likely to repeat/be violent
– The lab experiments (mostly with college students) tend to show effects (violence analogue, electric shock, 'angered' subjects) - aslo the Bandura one - tend to support media violence causality, but these are always artificial situation (altho eg when children left to play after the experiment has 'finished' then arguably natural, socially complex, environment) (see also p1111)
– More effects on younger children - hardly surprising
– Tend to be more males apparently affected, but this leads to on bias present in many - people who are already violent probably prefer violent media content and therefore there is not necessarily a causal link, only an associational one
– Literacy: positive results when children taught to interpret the texts - this implies that without such training, there is more likely to have negative effects
– Statistical significance: what does it mean exactly? Correlation /= correlation... (pp1109)
– Many studies done on violence, and on youth - why? (Ruddock's political point). Somewhere in Gunter he speaks of this bias also, and one (meta analysis I think) showed that depending on initial assumption (violence, neutral, or non violent) the outcome tended to support the assumption (see also p1111)
– "A simplistic, unidirectional model of media effects therefore may hamper the achievement of a comprehensive understanding of why individuals respond to media violence the way they do." (p1112)
A Prezi which summarises the main points
Some overall points
• Gunter paper is a review of behavioural research on media effects on violence - extensive and useful, goes over the different types of research, etc
• The most fundamental difficulty is that everyone is different, media effects get filtered through individual cognitive and cultural screens, "certain forms of media violence can exert certain kinds of effects on some media consumers some of the time." (p1113)
– Real or fiction: people more likely to respond and be affected by real violence (or violence said to be real, e.g. in experiments)
– Also legitimacy: when violence is shown as legitimate, people more likely to repeat/be violent
– The lab experiments (mostly with college students) tend to show effects (violence analogue, electric shock, 'angered' subjects) - aslo the Bandura one - tend to support media violence causality, but these are always artificial situation (altho eg when children left to play after the experiment has 'finished' then arguably natural, socially complex, environment) (see also p1111)
– More effects on younger children - hardly surprising
– Tend to be more males apparently affected, but this leads to on bias present in many - people who are already violent probably prefer violent media content and therefore there is not necessarily a causal link, only an associational one
– Literacy: positive results when children taught to interpret the texts - this implies that without such training, there is more likely to have negative effects
– Statistical significance: what does it mean exactly? Correlation /= correlation... (pp1109)
– Many studies done on violence, and on youth - why? (Ruddock's political point). Somewhere in Gunter he speaks of this bias also, and one (meta analysis I think) showed that depending on initial assumption (violence, neutral, or non violent) the outcome tended to support the assumption (see also p1111)
– "A simplistic, unidirectional model of media effects therefore may hamper the achievement of a comprehensive understanding of why individuals respond to media violence the way they do." (p1112)
A Prezi which summarises the main points
So I guess my senility started early 
Or, more likely, it was because it was a home computer test (on a ZX Spectrum, for the grey hairs out there) and the test was therefore not reliable. Something like this one is likely to be as good as it gets.