Trevor with bat in GTA 5

There is a pattern in the way that studies on the effects of playing video games are presented in the news media, and it's a pattern with some serious flaws. It begins with academic publications boiling down long and detailed studies into a few short paragraphs for a press release, with a title that's tantalizing enough to attract interest from editors (bonus points if the title rhymes).

If the press release does its job then it will get picked up by a number of news outlets. Very few - if any - of the writers charged with turning the press release into a story will actually read the full study, and will instead use only the quotes and results offered up by the publication. In turn, these are spun into something that the average reader will find interesting with a headline designed to get people clicking (or buying the newspaper, if you're feeling old-fashioned).

Here's a recent example. A team of researchers from Brock University conducted a study into the effects of violent video games on children's attitudes towards violence and the maturity of their sociomoral reasoning. The results were mixed and inconclusive: on average, there was no difference in stages of sociomoral reasoning found between the children who played mostly non-violent video games and those who played mostly violent video games. However, children who played violent video games for more than three hours each day were found to score lower on the sociomoral reasoning test than those who played only one or two hours.

Since the words "mixed" and "inconclusive" do not look at all good in a headline, this study - the conclusion of which was well-balanced and recommended simply that parents and teachers better familiarize themselves with the games that children were playing - was inevitably going to get skewed to one side or the other. Somewhat predictably, it was skewed to make the effects sound clearer and more significant than they actually were. Any results that suggested video games might not have a negative effect were omitted from Routledge's press release, and were subsequently left out of news reports as well. Add in a little bit of editorial embellishment and you get headlines like "Violent video games leave teens morally immature" and "Playing too much GTA can actually make you a sociopath."

Assassin's Creed Freedom Cry screenshot

This week another study on violent video games has been making the headlines. Douglas Gentile and Craig Anderson, of Iowa State University, analyzed a set of data from a three-year longitudinal study of 3034 Singaporean youth aged 11-17. The children were asked about their gaming habits and also given questionnaires designed to measure their behavior and their attitudes towards violence. Overall the children's reported aggressiveness decreased over time, normal for teenagers whose self-control is maturing, but those who played more hours of video games were more likely to demonstrate increased aggressive attitudes. For Anderson, it was pretty clear what had happened:

"What this study does is show that it’s media violence exposure that is teaching children and adolescents to see the world in a more aggressive kind of way. It shows very strongly that repeated exposure to violent video games can increase aggression by increasing aggressive thinking."

Those who have studied sociology, psychology or statistics in any capacity are no doubt already raising their eyebrows at this bold violation of the "correlation does not imply causation" rule. It could just as easily be the case, for example, that children who were more aggressive were more likely to take an interest in playing violent video games. It's also possible that the time spent playing violent video games and the attitudes towards aggression were both caused by another factor or mix of factors. It's worth noting that another study of 11,000 children in Scotland failed to find even a correlation between time spent playing video games and behavioral problems, let alone anything to suggest a causal link.

XKCD correlation
Source: XKCD

Gentile and Anderson's peers have levied their own criticisms of the study. Psychologist Christopher Ferguson, a familiar name in such debates, told Reuters that it is simply "not a very good study" and that "this data set has been criticized before." Science Media Centre collected responses from several different academics, and none were particularly convinced. Andrew Przybylski, a University of Oxford research fellow who has previously studied attitudes towards violent video games, argued that the results themselves don't seem to show the significant effect that the study's authors claim:

"Working backwards from some of the statistics present in the figures and tables it appears that violent game play accounts for a very small amount of variability in self-reported aggressive behavior. Said differently, imagine a Venn Diagram. The circle on the left represents all the variability observed for violent game play and the one on the right represents all of the variability observed in self-reported aggression: This research suggests that the overlap between these circles is in the neighbourhood of half of one percent."

David Spiegelhalter, a statistician and Winton Professor of the Public Understanding of Risk at the University of Cambridge, criticized the study for claiming a causal link between video games and aggression without proper evidence of one:

"This study shows an association, of unclear magnitude, of violent video game-playing with subsequent aggressive behavior. It does not, and cannot, show that the association is causal. The authors assume from the start that violent game play is what is influencing aggressive thoughts — the possibility that kids with latent aggressive tendencies tend to play violent video games does not seem to have been considered."

Patrick Wolfe, a Professor of Statistics at University College London, also criticized the paper for claiming that correlation implied causation, and for treating questionnaire answers as an accurate measure of real-world behavior:

"It is important to note what was measured: answers to questions about aggressive behavior — not the behavior itself... Of course, it could be that there is an additional unseen factor which is mediating both aggressive game play and the youths’ answers to questions about aggressive behavior. There is no way to tell from this study and so the study cannot justly conclude that video game play “influences” or causes aggressive behavior, or even answers to questions about aggressive behavior."

It probably doesn't need to be said that the majority of news reports on Gentile and Anderson's study did not critique it in a similar manner or even include any of the criticisms from other academics. Sample headlines include "Violent video games makes [sic] children grow up into aggressive adults, "Violent video games teach children aggressive thought and behavior patterns," and the straight-forward, "Kids learn violence from video games."

Spec Ops the Line artwork

Science and journalism are two fields that are ostensibly dedicated to a similar pursuit: finding out facts and sharing them with anyone who will listen. When the desire for a clickbait-y headline takes priority over transparency and full disclosure, it creates a certain narrative about video games in the news media, and it's a narrative that could have a damaging effect.

It scarcely needs to be said that there's more to violent video games than just violence. Sucker Punch's latest open-world superhero/villain game inFAMOUS: Second Son is set in a version of the USA where government fear-mongering over terrorism has led to some serious sacrifices of personal liberty. Upcoming Ubisoft title Watch Dogs has strong thematic links to the recent NSA scandals. Spec Ops: The Line takes the assumption that the crusading American soldier is always the hero and brutally subverts it.

When video games carry messages as strong as those in books or films, it's unnerving to see headlines like "Little By Little, Violent Video Games Make Us More Aggressive" in the news media, because these end up being used as the fuel for video game burnings and calls for government control and censorship of the medium. All because the problem of violence - a problem that has been around for thousands of years - is rather dubiously being blamed on a medium that didn't even exist until a few decades ago.

One thing's for sure: until this broken cycle gets fixed, stock photos of children holding video game controllers and looking angry are where the money's at.