Violence in Games: A Conversation with Christopher Ferguson, Part One
By Matt Peckham
PCWorldNov 4, 2008 11:22 am PST
I don’t know much that’d pass scholarly muster when it comes to violent games and children, but since it’s pretty much the hot button in video gaming these days, I’m constantly looking for people who do.
Texas A&M psychology professor Christopher Ferguson was more direct in his formal response to Pediatrics. The journal published his letter on the study, a letter in which Ferguson accuses the research of “weak results” and “misleading conclusions.”
Ferguson teaches at Texas A&M International University, where he researches violent behavior. He lists his interests as “examining violent behavior from a multivariate format, examining the combined impact of genetics, family environment, personality, mental health, and media violence.” He adds that much of his recent research “has focused on positive and negative effects of playing violent video games.”
Game On: So the news about this study in Pediatrics broke yesterday, but the reaction in public and professional quarters seems to be a bit skeptical.
Christopher Ferguson: There’s some skepticism, but of course, as you’re probably seeing there always is a little bit of, I don’t know what the right word for it is. It gets a lot of attention. And of course there are a lot of adults who don’t play games, who don’t know much about them, and I think there’s some anxiety about games like Grand Theft Auto in society, and particularly among older adults who don’t know much about games. A lot of these studies that come out of Iowa and Michigan and other places tend to get a lot of attention, as this article obviously is, because of those underlying anxieties about games.
GO: You’ve actually responded to the study in the same journal that this study was just published. Can you outline what they’re arguing?
CF: Sure. I think there are two points here, and while I’m not endorsing what they’re saying, this is it in a nutshell.
Their argument is that they’ve got a bunch of correlational studies that they claim link violent video game playing to aggression. And we have some experimental studies that say there’s at least a short term effect for playing violent video games on aggression, and their argument is that the thing that’s been lacking so far has been these longitudinal studies which show long term effects. So basically you measure violent video game playing at time one, then you come back, in the case of this article three to six months later, and then you measure their aggressiveness at a later time period. And if there’s a significant correlation there, then basically it suggests there’s at least some sort of long term correlational relationship.
The other thing that they attempted to do with this study, and I think it reflects some of their irritation with the criticisms or counter-arguments that they’ve encountered, is this U.S., Japan comparison. People point out all the time that Japan is saturated with violent media, probably more, if anything, than the United States. They’ve got the hentai, the sexualized violence, and all that kind of stuff, and yet they’re a very low violent crime society. So the argument is if violent media causes aggressiveness, how come it’s not doing it in Japan? And of course the argument has always been about the violent crime rate.
Now what this article is doing is ignoring the violent crime rate difference, and saying hey look, we did this longitudinal study in both countries and we got more or less the same effect in both countries. Basically they’re trying to argue “this is important,” and ignore that whole violent crime rate difference, which I think is probably the less convincing part of their argument. Not that I’m very well convinced by either one of them, but those are the two arguments that they’re trying to raise. They’re saying look, the effects are the same in Japan and the U.S., and look, they’re long term, because we’re doing a longitudinal study.
GO: They’re trying to say there were some important things they controlled for like gender and prior aggressiveness. You’re responding by saying wait a minute, there are several other things that you didn’t control for and those things could potentially explain why you got the results you did.
CF: They didn’t control for very much, quite frankly. I mean certainly the things they did control for were important. You know, they controlled for gender, and that’s important to note because of course males play more violent video games, and males are more aggressive. And so not uncommonly, you find what we call a bivariate correlation [a correlation which measures both the strength and the direction of a relationship between two variables], you find a correlation between violent game playing and aggression, sure, because males play more violent video games and males are more aggressive.
If you control for gender, oftentimes controlling for gender alone drops any relationship from a moderate size to a very small size. And in fact in their study that’s exactly what happens. They also control for what we call a trait aggression at time one, so preexisting trait aggression. Aggressiveness tends to be very stable across time. However aggressive you are today, you’ll probably be about that aggressive five years from now, 10 years from now, unless you get head injuries or something of that sort.
So once they control for the stability of aggression and they control for being male or female, they then report these, what I would describe as very small remaining correlations between video game playing and aggressiveness. But they don’t control for a lot of other theoretical variables that certainly could have caused these. In their argument, they’re using correlations to argue for causation, which is a problem that even most undergraduates could tell you is an issue. They’re not controlling for family environment, they’re not controlling for peer group influences, they’re not controlling for genetic influences, which to be fair, is very difficult to control for certainly. But at minimum, my concern is that these at least ought to be acknowledged. If not controlled for, at least acknowledged as alternate potential causes for what is fact a very small relationship between violent game playing and aggression.
GO: And you have evidence for this?
CF: Sure, some of my own research that I’ve done, I’ve found that controlling for family violence exposure pretty much wipes out any relationship between violent games and aggression, so the correlation is essentially zero once you control for family violence. They didn’t do that in this study, which is a significant concern for me.
GO: Is there an agenda here? And is there anything at all, separate from the problems you have, that’s interesting about these results?
CF: I would certainly say there’s an agenda here. There’s an interesting book that just came out by some other authors who are scholars, and they actually in their book go through how scientific ideology and political ideology have really affected the entire field of media violence research.
In terms of the methodology, very commonly the methodology in media violence in general, and certainly video games being a part of that, has historically been very weak. Unfortunately there’s a huge mismatch between the methodology that’s been historically used in media research, the findings in media research, and the kind of conclusions that are drawn by the authors themselves, or at least some of the authors. This study is an excellent example. Even if you took it at face value, which I don’t, video game violence overlaps somewhere between, based on their own statistics, a half a percent to two percent, with a variance in aggression. If you woke up tomorrow and you were half a percent more aggressive than you were today, would you notice that? It’s just not much of an effect.
If the author said look, there’s a little effect here, maybe video games increase aggression a tiny bit, but it’s not going to make anyone into a serial murderer, yeah, alright, we may argue a little bit over the methodology, though I’d still say they should’ve controlled for other stuff.
But what Craig Anderson argues in his paper, he then goes into describing youth violence, talking about how serious a public concern youth violence is. He doesn’t measure youth violence in his study. He doesn’t measure anything even close to it. The aggression measure he uses is not a behavioral measure, it doesn’t measure aggressive behaviors. It doesn’t predict youth violence. So they’re engaging in hyperbole that is not warranted by the results of their study, and that to me says there’s clearly an agenda.