More on video game violence

This is an image from the video game Asheron's Call 2 (source: mmorpg.com). Does playing such a game, involving regular practicing (albeit in a virtual environment) of repetitive, violent acts, increase our general level of aggression? A recent article in New York Times says no, citing a study by "a researcher at the University of Illinois," which found, according to the article, that "violent video games have no 'long-term,' or permanent, effects on aggressive behavior." Interesting, considering the article I discussed in yesterday's post apparently found exactly the opposite.

With some difficulty, we tracked down the study, by Dmitri Williams and Marko Skoric (this would have been much easier had the Times article noted the study's authors). Let's take a closer look at the study and see if it holds up to the Times' claims. Williams and Skoric tracked 75 participants as they played Asheron's Call 2, an MMORPG (massively multiplayer online role-playing game) which calls for regular fighting with an assortment of fantastic monsters. The participants played the game for a month, averaging about 56 hours of game play each. The average age of the group was 27, which, the authors point out, corresponds roughly to the typical video game player. The game also requires cooperation with other players in an online environment, and generally human players are not allowed to attack each other—only the fictional monsters in the game.

By studying participants over a month-long period, Williams and Skoric hoped to find out if video games can actually change people's behavior and attitudes over the long term. While some studies (such as this one) have found that violent games increase aggressive behavior in the short term, the only long-term data comes from correlational studies. Williams and Skoric gave pre- and post-study questionnaires to the participants assessing their general attitudes in aggression (using the Normative Beliefs in Aggression Scale) as well as asking whether they had been in arguments with a friend or significant other in the previous month.

When the researchers compared the results for the video game group with a larger group that did not play the game during the same period, they found no difference. For both groups, there was no difference in aggressive attitudes, and no difference in the number of arguments they reported at the beginning of the month compared to the end of the month. This does indeed sound like convincing evidence, but does it show that we have nothing to worry about in terms of video game violence?

Williams and Skoric say that the primary conclusion to derive from their results is that video games and other media must be treated differently. Much past research on video games has been based on earlier media violence studies. Video games, they argue, are different, because most games are played in a social context, unlike many laboratory studies where participants play only against the computer. Their secondary conclusion is that older gamers did show a marginal increase in aggression, suggesting that there is some concern with the growing number of older people beginning to play video games. Third, they suggest that the short-term effects found in other studies may wear out over time. However, they are careful to point out that since their sample included no young children and few adolescents, results may be different for this age group.

But their final conclusion is far from the conclusion reached by the New York Times article: they point out that video games, especially games that involve social interactions with other real players, are an incredibly complex phenomenon. A single study, on a single game, can offer valuable contributions to the overall level of knowledge about video game violence, but it shouldn't be used determine general government policy on video games. So, for example, this research, studying adults playing a social fantasy game, should not be applied to children playing a nonsocial, realistic street violence simulation.

Does Williams and Scoric's data undermine that of Gentile et al., discussed yesterday? The two studies hardly intersect. The average age of Gentile et al.'s population was 13; Williams and Scoric's was 27. Gentile et al. used participants' own ratings to determine how "violent" the games they played actually were; Williams and Scoric preselected a game they had arbitrarily determined to be violent. Gentile et al.'s most significant result was a correlation of exposure to video game violence with physical fights; Williams and Scoric didn't ask their participants about physical fights at all. In fact, like Williams and Scoric, Gentile et al. did not find a significant correlation between violent game exposure and arguing.

Let's face it: human behavior is astonishingly complicated; psychologists are still working on the "easy" problems like how we track motion and color. Understanding a complex social phenomenon like video games is not going to be a simple task, and making public policy based on that understanding will be even more difficult. Perhaps the best we can hope is for policy-makers—and the general public—(not to mention science writers) to understand that we're dealing with a limited set of data, and to not put too much faith in any single study.

Williams, D. & Skoric, M. (2005) Internet fantasy violence: A test of aggression in an online game. Communication Monographs, 72, 217-233.