Cognitive Daily

Several news outlets are reporting on a study, in some cases claiming that racing video games “cause” accidents or reckless driving. But causality is difficult to demonstrate in psychology research. Do the games really cause accidents?

Many of the irate commenters on the news articles claim that the study doesn’t really show causality, or that it’s not well-designed. I took a quick look at the actual study, and came away impressed: the study does show some causal links between racing games and poor driving behavior in simulations, but it also has some limitations.

One thing I like about the study is that it offers converging evidence. The researchers first demonstrate that there’s a significant correlation between playing racing games and real-world traffic accidents in men (but not women). Next they move to the laboratory, where participants first play a racing video game, then respond to a video simulator of real-world driving situations to assess risky driving behavior. Again, men who played the racing game took more risks and were slower to react to dangerous situations then men who played another, non-racing video game. Both men and women were more aroused and aggressive after playing the games, but only men translated these emotions into bad driving behavior.

But the study does have some limitations. The most obvious one is that responding to a video isn’t the same as driving in the real world — when you’re really driving, getting into an accident has serious consequences, much more grave than responding too slowly to a video. That said, the video simulator, called the Vienna Risk-Taking Test, has been shown to be a reliable indicator of real-world driving ability: if you perform poorly on the test, you’re very likely to have a bad real-world driving record. But it’s unclear whether the effect of playing a video game for 20 minutes would carry over to the road. Contrast this to video game violence studies, where people exposed to violent games actually believe they are hurting another individual.

A second limitation is the type of game being played. The racing games studied are all the kind of game that rewards anti-social behavior: driving on sidewalks, hitting pedestrians, and so on (indeed, this is the sort of game that has been most clearly associated with real-world violence). This leaves out a vast category of racing games where any accident is bad, where leaving the track slows you down or causes a crash. As we’ve discussed before on Cognitive Daily, the rewards system of a video game is critical: Games that don’t reward anti-social behavior — that distinguish between “good guys” and “bad guys” — tend to reinforce social norms and often don’t lead to violent real-world behavior.

So to say “racing games cause bad driving” is certainly an overstatement. Maybe saying “some racing games may cause bad driving” is a better conclusion, but it’s still not a very helpful headline. People need to know which games were studied, and why some games are worse than others. And mentioning that the results only apply to men in this study might be a good idea, too.


  1. #1 John Barneson
    March 21, 2007

    Very interesting. So maybe theoretically if driving skill in a video game is a decent indicator of real world driving record, perhaps playing a racing video game which caused a gradual increase in driving skill that might translate into an improvement in the real world road.

    If the study was performed using racing games that reward staying on the track but still encouraging the maximum speeds one can sustain, would that likely promote speeding as opposed to reckless driving, or would it simply improve the drivers capacity to drive fast and relative skill, since the cost of a speeding ticket would still act as a real world inhibitor of the new behavior?

    Great article!

  2. #2 rehana
    March 21, 2007

    John: Speed isn’t everything. You could set up situations that require fast reactions that don’t (only) depend on driving speed, like people or animals running into the road, cars braking suddenly, things falling off trucks.

  3. #3 Richard
    March 21, 2007

    I think the game is the primer for these results. A computer game is played for fun and excitement, because we don’t get that when we ordinarily drive around. Then the subjects move to a simulator, after having played a computer game for fun. Is it a surprise that they take some of that driving behavior into the “other game”?

  4. #4 Chris
    March 21, 2007

    Richard, that’s what I was thinking. The study shows that playing a fast-driving video game primes fast-driving video game behavior. It wouldn’t be surprising if this carried over to real-world driving, but the study doesn’t show that. And since there is an obvious explanation for the real-world correlation (that is, reckless drivers like fast-driving video games), I’m not sure this paper is much of a help.

  5. #5 Dave Munger
    March 21, 2007

    Perhaps the most surprising result is that the result *doesn’t* apply to women. If it’s unsurprising that fast-driving games prime fast driving in men, then surely it’s surprising that it doesn’t prime it in women.

  6. #6 hibiscus
    March 21, 2007

    i dunno. the games they picked for neutral show that there’s a third group, which is aggro non-driving games. they needed that one, and another group, non-aggro driving games, as you suggested. it’s strange that they picked the racing games they did, too, because the first part of their study they asked people about some legit racing simulators.

    dunno about the boy/girl split. did they get around to asking how well the players had liked playing the games? y’know, if they were interested in playing the game again.

    while they’re doing the next round maybe they can determine if people playing racing games leads to disbelieving global warming evidence.

  7. #7 Brian
    March 22, 2007

    I’m with Richard.

    I don’t see how you can draw deep links between real-world behavior and simulator behavior. Unless you’re evaluating real-world performance, the comparison doesn’t seem, to me, to amount to much more than educated guessing.

  8. #8 llewelly
    March 23, 2007

    Perhaps the most surprising result is that the result *doesn’t* apply
    to women. If it’s unsurprising that fast-driving games prime fast
    driving in men, then surely it’s surprising that it doesn’t prime it
    in women.

    Go back and read the list of violent car games they studied. All are
    targeted at men. None targeted at women. For the male participants,
    the violent car game was a familiar activity consistent with social
    norms. For the female participants, it was likely not a familiar
    activity, and not consistent with social norms.

    To me the idea that someone participating in a familiar and socially
    normal activity would respond differently than someone participating
    in an unfamiliar and socially unusual activity is expected.

  9. #9 Doug
    March 25, 2007

    “some racing games may cause bad driving”
    isn’t a valid conclusion, either. Which games cause bad driving?

    There are still too many confounds to draw any real conclusions linking video games and bad driving. Young men are already more aggressive and worse drivers, I believe. They have much much more experience playing video games and simulations, as well. Do video games cause bad driving? Or do the same things that lead to bad driving also lead to a higher interest in video games and simulations?

New comments have been disabled.