Video Games Don't Kill People. . .Women Kill People

When the Columbine shootings happened, people blamed violent video games for "turning children into killers." In the words of Time Magazine [May 8,2000] "These games teach kids to connect gore and glory . . .[where] the most vicious killers are the winners." I would have let this die down, but it's becoming accepted as an axiom. The reason was a two part study that allegedly proved that violent video games produce aggressive teenagers, mentioning the Columbine shooters by name. I read that study, and it proves that two things cause real violence: video games, and being female.

The article everyone references is here: Anderson and Dill. Or, if you want the full name, "Video Games and Aggressive Thoughts, Feelings, and Behavior in the Laboratory and in Life."

Now, the first thing I noticed is that, as with many soft sciences, polysyllabic jargon is used extensively to mask essentially simple ideas, and by complicating the issues, to avoid equally valid interpretations. For instance, they write "We deliberately chose to use a pair of violent and nonviolent games that are equally well liked for Study 2. In effect, this choice closes off this particular route to aggression, allowing a cleaner test of the more critical hypothesis that violent content itself can increase aggression. "

What they say they are doing is comparing violent to nonviolent; what they are also doing, though they deny it, is comparing active to inactive. A roller coaster is more exciting than "It's a Small World"', though both are train rides. However, by talking about the "General Affective Aggressive Model" (conveniently developed by the two authors), you put the core arguement in a context which supports it.

Right off the bat, you can tell I'm not down with their theory.

So, let's look at the actual study. There were two parts: one looking at students' previous history with video games and their current aggression levels, and one which exposed students to violent and nonviolent video games and measured their aggression levels afterwards.

Part 1: Tell Me About Your Misspent Youth

The first part of their study is some long-term statistical stuff proving that increased aggression correlates with increased time playing violent video games. They refer to this as "the predicted main effect of violent video games on aggression." Note the phrasing: it implies that they have proven that video game violence is the cause, and the aggression is the effect. The statistics do not prove this; the phrase statisticians like is "correlation does not imply causation." To use a brutal example, ice cream sales correlate with incidents of rape; that does not mean that ice cream causes rape. [In that instance, both numbers go up in warm weather.]

It is also worth noticing that they tested three secondary predictions of their model:

None of those predictions were borne out; violent video games did not produce the expected worldview. The crime and personal safety issues correlated mainly with gender, with females seeing the world as a more dangerous place. GPA had a slight negative correlation with total time spent playing all video games. (Makes sense to me; if you spend all your time playing video games, you're not doing your homework.) So if a defensive worldview is caused by an aggressive mindset, as implied by the model, high school girls are more likely to massacre their classmates than high school boys. I didn't bring Columbine into this: Anderson and Dill did.

In the conclusion to part 1, they admit that correlation does not prove causation, but note that lack of correlation would have disproven it. Part 2 is designed to prove that exposing college students to Doom makes them into cold-blooded murderers.

Part 2: Torturing Psych Students

The second study attempted to prove causality by exposing college students to one of two games, a violent one and a nonviolent one, and then measuring their responses.

There was a pilot study, where they were looking for two games that differed only in amount of violence and were otherwise as identical as possible. They noted that students reported that Myst and Castle Wolfenstein had the same ratings for difficulty, enjoyment, frustration, and action speed. Subjects reported a difference in excitement level between the two games, but this was completely ignored by the researchers, who instead used measurements of heart rate and blood pressure to measure excitement. Let me repeat that: The subjects' report on the excitement level of the two games was ignored. Myst is, essentially, a series of static slides, whereas in Wolfenstein 3-D the point of view moves smoothly. Any film student knows that moving the camera creates a feeling of excitement, whereas static scenes are relaxing. Once they decided that Myst was Wolfenstein minus the blood, they tested two groups of college students, whom I will refer to as Wolfies and Mysties, to see if one group was more aggressive after playing half an hour of video games.

The first experiment exposed the subjects to the game for 15 minutes, then tested State Hostility . State Hostility was proved to track with irritability: subjects who had previously been tested to have higher irritability [which is considered part of "aggression" from the first study] had higher State Hostility, and women had higher state hostility than men. Also, and this is a direct quote, "Game type effect and [higher-order effects] were nonsignificant." So if irritability causes mass murder, they only proved women will do it more than men.

The second session was to see how fast the subjects responded to aggressive words as compared to control words. They did some severe data treatment, and produced some completely unscaled numbers. On one hand, they got the results they expected: Wolfies had a faster reaction to aggressive words than to control words, compared to Mysties' reactions. On the other hand, there was no correlation with the subjects' irritability, which their model had predicted. Their proposed explanation is that people's personalities are totally overridden by 30 minutes of video game action. They really said that.

The third session is the interesting one. Participants are given a game to play- they are told that if they win, they can blast their opponent with a loud noise, and if they lose they are blasted. After each session [i.e. directly after the noise] they get to choose the volume and duration of the next noise they will deliver. In fact there is no opponent, and the actual win/loss and the noise level they receive is randomly selected. The expected result is that after getting blasted, subjects will choose a longer, louder noise as a return volley than if they win and deliver the noise. The aim of the experiment is to see if Wolfies are more vengeful than Mysties- when they "lose" and get the loud noise, do they respond with a louder, longer noise than their peaceful brethren?

The answer is,"sort of." Again, the data was heavily treated. There were several results found:

From that last point, I conclude that receiving a loud, annoying noise in the ear actually makes people gentler.

This gives me great hope for humanity.

Noise Times
Event Women Men% Difference
Win 6.896.653.5%
Lose 6.866.594%
% Difference .05% 1%
Event Wolfies Mysties% Difference
Lose 6.816.652.4%

Back to rants


They are expecting people to program "aggressive scripts that teach and reinforce vigilance for enemies (i.e., hostile perception bias), aggressive action against others, expectations that others will behave aggressively, positive attitudes toward use of violence, and beliefs that violent solutions are effective and appropriate." Back to top.

"On April 20, 1999, Eric Harris and Dylan Klebold launched an assault on Columbine High School in Littleton, Colorado. . . One possible contributing factor is violent video games. " Back to top.

The numbers they measured were the difference between the average result and the average for the specific group they measure, so that Wolfies might be "+5.54" and Mysties might be "-6.69". 5.54 out of 7? 5.54 out of 2,000? They don't say. Back to top.

Among other things, 6% of the points were thrown out for being "outliers"- that is, higher or lower than the researcher thought were reasonable. A few spurious data points is one thing, but six percent is a lot. Back to top.

" One possibility is that playing a highly violent versus a very mellow and nonviolent game for two 15-min periods of time was sufficient to temporarily override the usual differences between people high and low in irritability in relative accessibility of aggressive thoughts." Back to top.

Anderson and Dill like the word "blast" in conneection with the noise; they use it consistently. Back to top.