In light of the first-person shooter-style massacre at Virginia Tech, Slate columnist Amanda Schaffer cribs from the academic literature to argue that video games really do contribute to violent behavior. The point about video game violence is important, so I write about it first. After that, I’ll spar with Schaffer for apparently summarizing somebody else’s summary of the literature without giving proper credit.
Part I: The link between violent games and real-world violence
Schaffer basically gets it right: playing violent video games increases the odds (but does not guarantee) that a person will engage in real-world violence. Here is the relevant block of text from Schaffer’s Slate article (with my editorial additions in parentheses):
Three kinds of research link violent video games to increased aggression. First, there are studies that look for correlations between exposure to these games and real-world aggression. (These are cross-sectional studies that take a snapshot survey and look for relationships between variables; using them to establishing causality is tricky but not impossible.) This work suggests that kids who are more immersed in violent video games may be more likely to get into physical fights, argue with teachers, or display anger and hostility (STUDY 1: Gentile & Anderson, 2006).
Second, there is longitudinal research (measuring behavior over time) that assesses gaming habits and belligerence in a group of children. One example: A study of 430 third-, fourth-, and fifth-graders, published this year by psychologists Craig Anderson, Douglas Gentile, and Katherine Buckley [STUDY 2], found that the kids who played more violent video games “changed over the school year to become more verbally aggressive, more physically aggressive,” and less helpful to others.
Finally, experimental studies randomly assign subjects to play a violent or a nonviolent game, and then compare their levels of aggression. In work published in 2000, Anderson and Karen Dill [STUDY 3] randomly assigned 210 undergraduates to play Wolfenstein 3-D, a first-person-shooter game, or Myst, an adventure game in which players explore mazes and puzzles. Anderson and Dill found that when the students went on to play a second game, the Wolfenstein 3-D players were more likely to behave aggressively toward losing opponents. Given the chance to punish with blasts of noise, they chose to inflict significantly louder and longer blasts than the Myst kids did. [...]
If we had only one of the three kinds of studies, the findings wouldn’t mean much. But taken together, the body of research suggests a real connection.
I share this view of the literature. While the academic research on the subject is still growing in important ways, it tends to support a substantial connection between playing violent video games and an increased propensity to behave violently.
There are dissenters. In researching this story, I found a deeply flawed 2007 meta-analysis by Christopher J. Ferguson; he argues that this trend is best explained as the result of “publication bias”, or the problem that journal editors and researchers are more likely to trumpet positive results (games cause violence) than negative results. As just one indicator of how deeply flawed this study is, Ferguson claims that unpublished studies could easily disprove claims of a significant intereaction, but he never actually tries to track down unpublished results. This ignores one of the basic guidelines of meta-analysis: “All research, not just published research, should be included.”
There are also snooty chuckles from ill-informed video game apologists on sites such as ArsTechnica (a good site on some counts but not a bastion of scholarly insight) and the Entertainment Consumers Association blog GamePolitics (which has a transparent agenda). ArsTechnica cites Ferguson despite admitting that they have not read the article (let alone other studies on the subject). In short, there is some decent evidence that there is a connection, hecklers notwithstanding.
This fits well with the research on the broader research on the effects of media violence. In 2003, eight established media effects researchers provided a comprehensive analysis of the available literature (pdf) on the effects of violent media, including video games. Their conclusion:
Research on violent television and films, video games, and music reveals unequivocal evidence that media violence increases the likelihood of aggressive and violent behavior in both immediate and long-term contexts.”
Obviously, it takes more than a year of playing a first-person shooter to become a homocidal maniac. Many of my friends play FPS games and I’ll still trust them around my knives. Yet the evidence strongly suggests that these games, like tobacco, should be viewed as unhealthy and for adults only.
We at ShoutingLoudly aren’t calling for a ban. I’m not even calling for tighter regulation. It does, however, suggest the need for collective social and especially parental responsibility. I don’t think that’s too much to ask.
Part II: Did Schaffer crib a bit too much?
Disclaimer: Some may view this as nitpicking, but it’s important to point out instances of incomplete source citation and ask the resulting painful ethical questions. Schaffer has clearly done a certain degree of original writing and research to write this piece, but a substantial amount looks cribbed in important ways.
If you compare the meat of Schaffer’s discussion about the effects of violent video games (the block quote above) with Gentile and Anderson (Study 1), you’ll notice some striking similarities. While Schaffer link-cites Study 1 in a way that implies she is relying upon it for a very specific claim (correlational studies demonstrate a cause-and-effect relationship), Study 1 actually makes the exact same over-all claim as Schaffer:
The three major types of studies—experimental, correlational, and longitudinal—have different strengths and weaknesses. (p. 228) […]
A strong case for a real effect arises if the same results are found no matter what way one studies it. […] Although more research is needed, all of these types of studies have been conducted with similar results: playing violent video games can indeed cause increases in aggressive thoughts, feelings, and behaviors. (p. 229)
In short, Gentile and Anderson have already written this part of the Slate story. All Schaffer needed to do was to distill it into easily-digested form. That is a fine and commendable function of journalism, but one must be honest about who has done the heavy lifting.
This alone may be chalked up to sloppier writing standards on the internet (e.g., giving 25% credit where 75% is due), but there’s a quote out of place that shifts Schaffer’s problem into a higher gear. As quoted above:
A study of 430 third-, fourth-, and fifth-graders, published this year by psychologists Craig Anderson, Douglas Gentile, and Katherine Buckley [STUDY 2], found that the kids who played more violent video games “changed over the school year to become more verbally aggressive, more physically aggressive,” and less helpful to others.
The link is to a publisher’s website describing the book, so I am unable to examine Study 2 myself. That is not necessary to find the source of the quote, however, which is Study 1, above. On page 231, the authors write:
430 third, fourth, and fifth graders; their teachers; and their peers were surveyed at two points in the school year. […] Indeed, children who had high exposure to violent video games changed over the school year to become more verbally aggressive, more physically aggressive, and less prosocial. [emphasis added]
This is Study 1’s summary of Study 2, Anderson, Gentile, and Buckley—which Schaffer represents as a separate study she herself is summarizing. She implies that the quotation is in Study 2, but it is clearly in Study 1, a piece she has already cribbed from. I must hold out the possibility that this exact quote is also in Study 2 and Schaffer discovered this in the library, but the odds seem slim.
Study 3 is also likely to have been discovered via Study 1, which cites it. To give credit where due, Schaffer clearly does download and peruse the PDF of Study 3. She knows the games used (Wolfenstein 3-D and Myst), details not included in Study 1. But the study’s results and its significance within the literature are both spelled out explicitly in Study 1.
Schaffer’s description of the research is, to my knowledge, spot-on. I thus find it hard to imply that something smells fishy. But it does.
If I were Schaffer’s professor, I would insist on a face-to-face meeting and enter it with a healthy dose of skepticism.
What do you think? Is this plagiarism? Is it more or less acceptable from a professional writer than from a student or hobbyist?
P.S. This post is a great time to come clean about something. The Slate story, and several other stories I have written on ShoutingLoudly, have come from the Temple University Mass Media and Communication listserv, run by Associate Professor Matthew Lombard. Lombard is also the heaviest contributor, using the listserv like a private news aggregator blog, and he posts a lot of timely stuff that is of interest to mass comm scholars.
I’ve not disclosed this before for several reasons. First, it’s a listserv, so link credit is no simple task.
Second, when I’m using it this way, it’s generally serving the same role as a news aggregator. Come clean, fellow bloggers: have you really credited EVERY story you found on Slashdot, BoingBoing, or Digg? (Do you even always remember where you found the link?) Do Google News and Technorati deserve link love—consistently, occasionally, or never? Especially when compounded with the lack of links, this basically leaves me crediting somebody who has chosen not to run a blog and rarely includes more than a sentence of preface. Now that I think about it, this is probably still good practice, but I dare say it hardly counts as plagiarism until and unless one is relying on such a source largely to the exclusion of gathering news oneself.
Third, it’s a small-audience list (~150) that Lombard staffs all by himself. I’d hate to make his workload unmanageable by drawing the attention of our whopping 3.5 readers.
It feels dirty to nail Schaffer without admitting that I count on others (my wife, Tina Collins, is also very helpful) to help me find interesting content for this blog. But rest assured, dear reader, that the half-informed opinions and dim-witted humor are absolutely mine. Except all those quips I stole from who knows.
For a particularly excellent article playing with the ideas of plagiarism, authorship, and copyright, see Jonathan Lethem’s article in Harper’s, “The ecstasy of influence: A plagiarism.”
Dennis McCauley here, editor of GamePolitics. So, what’s the “transparent agenda”?
Yes, we are a pro-game consumer (not pro-game industry) site, but we routinely report research both pro and con on game issues. Had you been a regular reader or taken the time to search our “video game research” category ( http://gamepolitics.com/category/video-game-research/ ), you would have known this.
At GP we’ve taken the game industry on over a number of issues over the years when warranted, and our record more than bears that out.
In Amanda Schaffer’s case, we accurately pointed out the APA’s 2005 conclusion that violent games are statistically linked to increased aggression, not actual violence.
As a fellow contributor to Shouting Loudly and a video game researcher, it’s probably my responsibility to chime in now to say that I don’t agree with some of the conclusions you come to here.
I don’t really have as much an opinion on “Part II” regarding plagiarism. These seems to be some lazy writing going on here, but we as teachers have enough verbatim lifting from Wikipedia to deal with that it’s not really worth it to call someone in to the office who can’t keep her sources straight. Mark down a letter grade, write in the margins in red pen, and move on.
As for “Part I,” however, while I am all for “collective social and especially parental responsibility,” I don’t agree that video games require this any more than other media. Video games are at the center of a moral panic, and their impact has been grossly blown out of proportion. I only know of one study on Bible passages’ effects on aggression, for example, but it also found a positive correlation.
I can’t blame you for describing the VA Tech massacre “first-person shooter-style,” considering that that seems to be what everyone is calling it; consider, though, that the shooter owned no video games, and hadn’t even played Counterstrike since high school (he was a senior in college). While columnists and politicians are still latching onto this as a video game issue, responsible news outlets who checked their sources have noticeably distanced themselves from such claims. (I’ve written about this on my other blog, as well.) Law reviews and magazine articles discussing video game violence typically start with some invocation of statistically unusual school shootings in which police and judges have consistently dismissed video game play as a causal factor. See, too, Dmitri Williams’s “The Video Game Lightning Rod,” which demonstrates how journalistic coverage of games has followed a fairly typical pattern associated with earlier moral panics surrounding popular media. In other words, I think this has more to do with more lazy writing and following trends than with the actual state of current research on game violence, or even current social problems actually linked to games for that matter.
In fact, Schaffer’s approach to the literature is reductive and strangely reminiscent of the way the defendants in video game violence legislation cases described this body of research. Schaffer focuses on literature by Craig Anderson and a small group of others who have transitioned from working on TV violence research. I cover a lot of this ground in an earlier post here, but just to pull out a couple details, first consider Judge Kennelly’s statement in overturning Illinois’s doomed game regulation law. The defense seriously weakened its argument by only presenting research that agreed with its position, rather than addressing any of the research that comes to other conclusions.
There’s more of this literature published than you might expect, too: consider Williams and Skoric’s short-term longitudinal study of an MMO (which found no effect) and John Sherry’s meta-analysis of video game violence research (which found a positive effect, but less than other studies, and put into perspective). Meanwhile, the British Board of Film Classification is considering rating games less harshly than film (the reverse being case as of now in the U.S., at least), as a recent study suggested that games are actually less emotionally engaging than other media. (I know there are more examples, but I wasn’t planning on blogging today, so hopefully I can be forgiven for glossing over this a bit.)
Also consider Kennelly’s (and Sherry’s, and others’, and my) contention that “aggression” as operationalized in much of the violence literature is still a far cry from real-world “violence.” Honking a horn at someone really loud or writing swears on a piece of paper do not equal real world violence. Moreover, what if the “aggression” these studies are picking up is a good thing for many or most players? To quote from my earlier post: “The ‘aggression’ encouraged by playing games might not necessarily be a bad thing. A member of the Quake Grrlz movement (as quoted by Henry Jenkins in his Senate testimony) once suggested, ‘Maybe it’s a problem…that little girls DON’T like to play games that slaughter entire planets. Maybe that’s why we are still underpaid, still struggling, still fighting for our rights. Maybe if we had the mettle to take on an entire planet, we could fight some of the smaller battles we face everyday.'” Seeing as how I’m currently working on a dissertation on “geek culture,” it might also be worth noting in passing that middle and high school ethnographies have indicated that non-aggressive boys are particularly likely to be labeled “geeks” or “nerds,” and these titles (perhaps not coincidentally) have also long been applied to gamers.
I won’t claim that video game violence research is without merit or fails to show some sort of influence, but I will insist that video game violence is not the problem it’s being made out to be. There are too many stumbling blocks along the way to making a conclusion that this medium deserves any more attention than other media. As Henry Jenkins points out in a recent essay on media violence, the current debate is more about pointing fingers and assigning blame than about actual “effects.” A more holistic view of violence in culture opens up opportunities to talk about what might be attractive about violent content in the first place, and how to address what people do with it. Focusing on one medium disproportionately, however, just makes it easier for game publishers to make waves with controversial content. Or, on the other end of the spectrum, this puts even the publishers of non-violent games at risk of a backlash like that seen against comics in the 1950s, from which creators and publishers have struggled to recover for decades.
Moral panics come and go, but the policy that results can have a lasting impact. I know you’re not calling for regulation, Bill, but let’s consider where most arguments of this sort end up going.
To Dennis, I will agree to moderate and modify my claims about GamePolitics. (Perils of overnight blogging. Apologies.) Here’s a clarification: I still view the post trumpeting Ferguson as negligently failing to contextualize the article within a studied consideration of the debate from within the media effects literature.
There’s nothing wrong with having an agenda, of course, and it was lazy writing on my part to phrase it this way (which is, regrettably, an ad hom attack–no, really, my apologies). But this post doesn’t engage the media effects literature in a meaningful way. Perhaps it’s too high a standard to expect something like Jason’s comment (he IS a media studies scholar, after all), but it’s problematic to trumpet an article that is not available to most readers without even considering its weaknesses.
All that said, I must concede that it is more thoughtful than the fistful of other blog posts I found citing the article.
Regarding Jason’s comment: good work. I agree (and even pointed out) that video game effects research should be considered as part and parcel with media effects research generally. In light of the meta-analysis on violent effects of violent media more generally (Anderson et al., 2003), this would actually support the claim of effects–albeit with the caveat that the early evidence suggests video game effects may not be quite as strong.
That said, Jason does catch me failing to properly contextualize the debate. As we’ve discussed before in the office (read: Jason dropping knowledge, me listening), the moral panic paradigm is an excellent explanation of how the debate is playing out in the public sphere. In short, nice catch.
Yet the literature suggests violent media generally, including video games, are at least potentially dangerous. As a society, I hope we can water down the genre-specific video games by spreading that outrage to violent role modelling generally.
As far as media policy prescriptions go, this is really the South Park movie paradigm: why do we censor profanity on TV (on the assumption that kids will parrot curse words) but allow impossibly graphic violence during prime time? If we’re worried about the effects of media on kids–and I think we should be–we really ought to keep kids away from all violent media.
I don’t think we should censor games or TV or movies, but as parents, neighbors, uncles, etc., we should keep kids in media environments that encourage the right behavior. Industry-led ratings are an appropriate addition to this solution. FCC fines for naked boobies? Not so much.
If we want well-adjusted kids, maybe we shouldn’t let them watch hours of HBO every week. FPS games just go on that list as far as I’m concerned. After all, as expert Pat Brown argues–recounted on GamePolitics:
“VIOLENT video games can be a part of this picture as they lend to the loss of empathy that is a hallmark of psychopathy and young children viewing repetitive violence and participating in “killing” via video games are living in an unhealthy psychological environment. Furthermore, teenagers who are already psychopathic and then spend a great deal of time with violent video games are being inspired to act out their psychopathy in a similarly violent manner.”
Rather than attacking those who are concerned about video game violence based on uncritical citation to flawed academic research (as on ArsTechnica), the reasonable response is to enter into a broader dialogue that points the fingers at movies and other violent role models. As Marilyn Manson rightly notes, the President’s violent foreign policy is probably an even bigger problem.
Sadly, the single strongest cause of violent behavior among youth is almost certainly the hardest to solve: violent adults in households.
Bill, actually, we wrote about the Ferguson study without comment. We merely stated that there was such a study and quoted from his findings.
I do not feel sufficiently qualified from a research standpoint to comment on Ferguson or any other researcher’s conclusions.
I do comment on others who draw conclusions from such research, as I did in regard to Amanda Schaffer’s article, which, as I noted, had a sensationalized title and made heavy use of qualifiers.