With the latest news as reported by El Reg that a US congressman has proposed a ‘health warning’ to be added to video games about sexual and violent content, I ended up thinking about it (a curse I know). The article quotes him as citing “the proven links between violent video games and violent behavior”.
Why is it that there seems to be an even number researchers finding a link, and finding no evidence of one ? El Reg suggested elsewhere that perhaps it’s because the brain is a complex thing which doesn’t have any kind of single guaranteed response to a given stimulus and this is probably true, but I think it’s not all of the story.
The first is a problem that plagues psychology as a science and is one of the main reasons why it advances very slowly and all it’s results should be taken with several bags of salt (remember, it’s a slow-advancing science and just so happens to be one of the youngest sciences too). Causal relationships are among the hardest things in science to prove at the best of times.
A classic example I remember from my university philosophy class (specifically in the module on logic and critical thinking) cited a set of statistics that indicated that sexual education in US schools increased by 80% during the 50′s. In the same period the incidence of teenage sex was shown to have increased by about 45%. The original author tried to use this as proof that sex-ed encourages promiscuity.
In fact, it had made no such proof. The data shown could have one of three possible links with one another:
1) It is possible that, as the data suggested, sex-ed in general led to more teenagers having sex.
2) It is possible that other factors led to more teenagers having sex, and this led to the increase in sex-ed (in other words the causal relationship could just as easily have existed in the opposite direction)
3) The two may not have been related at all, just two things which happened at the same time and didn’t actually influence one another, possibly due to some unknown other factor.
To actually get an idea of any valid conclusions of the data, we need to know what other factors may have been present. Well the 50′s was the age when birth control was invented for one, and female education reached hitherto unknown heights. It’s widely believed that these things led to the sexual liberation of women as we know it today. It thus makes a lot of sense that promiscuity among teenagers most likely increased because the biggest risk in the activity (unplanned pregnancy) had become mitigated. It is also likely that this increase may have prompted increased sex-ed in schools (but this is by no means proven). But it does seem that the original arguers conclusions were fatally flawed and probably the least likely of the possible scenarios suggested by known data.
This is a large part of why video-game violence studies keep coming up with conflicting results, it’s almost impossible to prove causal relationships to begin with, proving them with something as varied as personality responses to stimulous borders on the impossible – people just don’t react the same way to the same things, thinking teenagers are more likely to have some standard reaction is just stupid.
But shouldn’t there be at least a trend, a marked amount of people showing at least some of the same reaction ? Well no, there doesn’t have to be. Some people are widely turned on by a simple picture of a naked person of their preferred sex, some are disgusted by it, some are completely unaroused (but may respond well to the sight of the same person in some sexual act they personally fantasize about). The site of a nicely done piece of steak will make some of us drool with hunger and others throw up in disgust. If even our most basic drives are not consistent in how they reflect into our behavior and personalities, why should high-level concepts like fictionalized violence have any consistency whatsoever ?
What is more, there is a big difference between immediate response and general personality. Immediate response is not always consistent with how a person would mostly respond to the same stimulus. I love steak and normally I would be in the ‘drool with hunger’ group, but during my recent illness the sight of food held no appeal to me and steak with it’s fat and juices would probably have made me throw up. A temporary unmeasured factor led to my response to the stimulus being completely out of sync with my normal behavior. It certainly didn’t mean that when I got better one of the first things on my list to do was no longer to “get a nice steak”.
If video games just make you feel a little aggressive while you play them… well who cares ? The question is whether this immediate response will lead to a long term change in behavior. Personally I think we see such varying results because playing games probably does increase adrenaline levels leading to a short-term immediate rise in aggression among many players, but the moment the stimulus is removed the aggression disappears and there is no reason to believe it will have any additional influence on behavior or ever return without the same stimulus being there.
Personally I think the studies that say ‘yes’ are looking at short-term response only, the ones that look for long term influence on personality find that whatever instance of it exists is so small they pretty much never get an example in their focus group.
In short – I guess the inconsistency of human behavior and the gap between stimulus response and personality development is so huge, that researchers will always find what they are looking for by just looking at what suits them.
This of course, suggests that those researchers studying the question are consistent about one thing: very bad science. A good scientist will never try to prove a theory, he tries to disprove his theory. If good science was their goal, they would take the theory, ‘video games can lead to violent behavior’ – and try to prove it is false, if they were consistently unable to succeed – that alone would give the theory any merit. We trust scientific theories not because of the experiments that seem to prove them, but because of the experiments that fail to prove them wrong. This is a basic element of the scientific method.
Sadly, in my experience, psychologists are generally not very good scientists – not least because their chosen field is one where scientific rigor can guarantee you one thing for sure, slow results. Studying human behavior is a field where the data set is so huge and varying that getting any reliable trends whatsoever out of it will never take massive amounts of effort and huge amounts of time by thousands of people working at the same single problem – and that is not the kind of science that gets grants, sadly it’s the reason why pretty much every theory in psychology when applied to any given individual has proven about as useful and accurate as astrology.
One thing I am certain of, justifying such radical measures as censorship (and yes, it is perhaps the most radical measure a government can take against it’s own citizens short of outright massacre) using a science so young while it’s faced with a challenge so huge is outright stupid – it’s an excuse for politicians to get more power, it is definitely not a case of sensible legislative response to scientific advance.