Nonprofit, nonpartisan journalism. Supported by readers.

UCare generously supports MinnPost’s Second Opinion coverage; learn why.

No clear link has been found between video games and violence

Rockstar Games
"Grand Theft Auto V" topped $800 million in world-wide sales on its first day of release last Tuesday.

After it was reported that Aaron Alexis, the gunman in last week’s tragic Washington Navy Yard shooting, may have played military-style video games for up to 18 hours a day, some media pundits began to suggest a direct link between the games and his violent actions.

It’s a connection that seems to gain favor in some corners of the media whenever there is a mass shooting — or other type of violent crime.

The belief that violent video games can lead to real-life violence is not a new one, of course. As far back as 1982, the then-Surgeon General C. Everett Koop was warning parents about “the aberrations of childhood behavior” that could result from playing video games.

But, as I’ve noted here before, several decades of research on the topic has found no clear evidence linking video games, even notoriously brutal ones like Call of Duty or Grand Theft Auto, to outbursts of violence.

What the research shows

Writing in the Guardian last week, British psychologist Pete Etchells explains (with British spellings) why the issue is not as black and white as the media pundits suggest:

There is some evidence to suggest that there is a link between playing violent video games and showing more aggressive tendencies, at least in the short term. For example, in a study published in the Journal of Experimental Social Psychology this year, participants played either a violent or non-violent video game for 20 minutes per day over 3 days. After playing the game, they then played a competitive task in which, if they won, they could blast their opponent with an unpleasant noise. The researchers found that participants who played violent games blasted their opponents in the secondary task for longer, which was interpreted as an increase in aggressive behaviour.

However, it also depends on the context in which these sorts of games are played. A study by Seth Gitter and colleagues, published in Aggressive Behaviour this year, showed that if participants were asked to play a violent video game with a positive goal in mind (for example, protecting a friend in a zombie game), they showed reduced levels of aggressive behaviour compared to participants who were asked to simply kill as many zombies as possible. In other words, it’s not the simple act of playing violent video games that dictates whether they have a negative effect.

Context is also important, says Etchells, when interpreting the results of studies that investigate the long-term effects of video games on behavior:

A 2012 longitudinal study looking at the behavioural development of 165 teenagers over the course of 3 years found that when pre-existing emotional, family and social problems were accounted for, any aggression-increasing effects of playing violent video games disappeared. However that same year, Teena Willoughby and colleagues at Brock University in Canada looked at 1,492 children from the ages of 8/9 to 17/18, and found that those who played violent video games over a long period of time also showed increased development of aggressive behaviours.

A big reason for the conflicting results of these and other studies, says Etchells, is that they measure aggressive behavior in different ways and do not control for the same confounding factors (variables like age, geography, or socioeconomic status that might also explain a study’s outcome).

And, of course, the long-term studies are observational. They can show only a correlation, not a causation, between two things.

No definitive evidence answer

“The question as to whether playing violent video games negatively affects behaviour hasn’t been completely answered yet,” Etchells concludes. “Moreover, because ‘violent video game’ (much like ‘screen time’) is quite a broad concept, we’re probably not capturing the subtlety of any effects in an adequate way at the moment. To really get an understanding of what’s going on, we need to be looking more at the way in which these sorts of games are being played – for example, no one has yet really looked at if and how the multiplayer aspect of video games (playing in the same room together, playing online together) has any sort of effect.”

“So until there is more definitive evidence,” he adds, “it doesn’t seem right to imply that there is a clear and known effect. And it certainly isn’t right to tenuously highlight links between video game use and violent behaviour whenever it is vaguely possible to do so. It detracts from figuring out if there is another underlying cause instead.”

You can read Etchells’ commentary on the Guardian website.

Comments (8)

  1. Submitted by Rosalind Kohls on 09/24/2013 - 01:46 pm.

    Not that I’m in favor of either

    I’ve always thought it strange that tobacco advertising influences behavior, but violent video games don’t.

    • Submitted by Todd Hintz on 09/24/2013 - 03:56 pm.


      Influencing is one thing; mowing people down is another. Sure advertising can sway people–just look at the number of dolts who buy things like teeth whiteners. But then to say they go on to shoot people left and right is just a bit of a stretch.

      I think the bigger issue here is people who are mentally deficient having access to guns. Sane people aren’t likely to take a semi auto weapon or a shotgun and start popping off kids or their coworkers left and right.

    • Submitted by Paul Brandon on 09/25/2013 - 10:35 am.

      You might start with the fact that

      tobacco is an addictive drug.
      The act of smoking is physically rewarding.
      That said, tobacco advertising does not cause people to smoke (peer pressure does that). It is designed to shift an established habit from one brand to another.
      Huge amounts of money are spent on how to do that.
      And of course, unlike video games, tobacco ads show people in real world situations.

  2. Submitted by Paul Brandon on 09/24/2013 - 02:58 pm.

    A proper study

    Would randomly assign children at birth to at least three groups:
    One would view a high (specified) level of violence.
    The second would view a low level.
    The third would view NO violence.
    All three would be followed for at least twenty years, and their actually violent behavior (not statements about violence, or behavior on some simulated task) would be measured.

    Obviously such a study would be difficult and expensive, and has not been done. So, we are left with flawed approximations which cannot support an inference of causality.

  3. Submitted by Dennis Tester on 09/24/2013 - 07:06 pm.

    I tend to believe the conclusions

    from the study done by Seth Gitter, et al.

    In that “Participants who played a violent game in which the violence had an explicitly pro-social motive (i.e., protecting a friend and furthering his nonviolent goals) were found to show lower short-term aggression (Study 1) and show higher levels of pro-social cognition (Study 2) than individuals who played a violent game in which the violence was motivated by more morally ambiguous motives. Thus, violent video games that are framed in an explicitly pro-social context may evoke more pro-social sentiments and thereby mitigate some of the short-term effects on aggression observed in previous research.”

    Games in which the objective is to eliminate bad guys (terrorists, for example) instead of guys in general could actually result in a positive, vicarious feeling of accomplishment. Certainly for those who possess the warrior gene.

    • Submitted by Paul Brandon on 09/25/2013 - 10:30 am.

      First of all

      there is no “warrior gene” — genetics are not that simple.
      Second, I checked the Abstract of the Gitter, et. al. study.
      It’s not clear that their measure of violence goes beyond performance in a video game and verbal statements about that performance. No indication of any real world measure of violent actions.
      And (contrary to the preponderance of the literature) they start out with the assumption that video games cause violent behavior.

  4. Submitted by Kim Millman on 09/25/2013 - 10:04 am.

    Problem with most studies

    I have a career in analyzing information and the problem with many studies is that they were designed poorly from the beginning. In some instances, studies appear to be intentionally flawed. Many colleges and other prestigious institutions that perform studies have hidden conflicts of interest by way of in direct large corporate contributions or affiliations.

    For example, some cell phone studies were designed to never find what they were allegedly looking for. For example, one study was set up in a small town with elderly patients trying to find an association with cell phones and brain tumors. It wasn’t too hard to determine after reading the study’s design that obviously there would be no correlation to cell phones. The area did not have good cell phone reception and the study participants were too old to have ever used cell phones much in their lifetimes. The study also classified a large amount of cell phone use as being two or three times of use per week. The result of the study was widespread press releases saying there was no correlation to cell phones and cancer. The media failed to read the study or analyze it to determine how ridiculous it was from the beginning.

    Another extremely flawed study that is still broadly cited today as conclusive evidence that there is no correlation between native job loss or wage erosion from immigration. From an isolated macro perspective, the study was correct. The problem was that it had no applicability in broader sense nor was it accurate from a micro perspective. The study is still being cited in many government reports for public policy makers.

    The question for mass killings is not whether violent videos cause aggressive behavior in the average person, it is whether they cause or facilitate the behavior in a certain type of person. The specific correlations with mass killings appear to be, (1) mental illness, (2) most of the violators appear to have taken anti-depressant drugs or prescription sleep medications in close proximity to the time of the crime; (3) most have the means to gain access to an unlimited amount of fire power; and (4) many are now emulating either each other or the video game personas.

    Setup the study on the types of people committing the crimes.

  5. Submitted by Paul Brandon on 09/25/2013 - 02:22 pm.

    Mental Illness

    has a low correlation with violence; it is not viewed as a major cause of violent behavior. The other relationships that you cite are still correlations.
    I’d appreciate a citation on the (possible) link between antidepressants and sedatives and violent crime. My career was in behavioral psychology with an interest in behavioral pharmacology — I’m not aware of any support for this.
    These are very widely used (and abused) substances. As someone pointed out above, it’s akin to pointing out that most killers have previously consumed milk.

    I agree on guns — it’s very difficult to kill large numbers of people without guns or bombs, and -everyone- has easy access to guns.

Leave a Reply