Spirit of the Stairway — Measure of Violence of “Violent” Computer Games

The books that the world calls immoral are books that show the world its own shame.
Oscar Wilde

In the posting about so-called “violent” computer games, I mentioned that I had an issue with one of the measurements. I think it produced measurement artifacts and was not an actual measure for the detrimental effects of violent computer games. But I couldn’t phrase it at the time. A nice “spirit of the stairway” moment, but here we go.

If I recall the issue correctly, the argument was that “violent” computer games desensitize players to violence. A method used to test this effect was to use a scale with descriptions of different forms of violence (e.g., ostracizing someone, hitting someone, etc.) and letting the person rate them on how violent they are (e.g., on a scale of 1 to 10, one being not violent at all and ten being absolutely violent). Half of the participants did play a “violent” computer game beforehand, the other half did not (e.g., they played another game like Tetris). If the players who did play the violent computer game rate the violent behaviors as less violent than the ones who did not, you can show that “violent” computer games desensitize these players. You could further argue that this would mean that they don’t see violence as that bad, and perhaps even that they might be more prone to violent behavior themselves.

So much for the argument for this measure.

In principle, it’s a nice measure. But my issue with it is that it does only reflect the current situation the players are in, not anything within the players. It over-represents the situation in the rating values — they do not represent the disposition of the person. But let’s make the issue clearer with an analogy.

Suppose you have two groups of people, like in the example above, and you give one group a book about whales, and the other a book about cats. Now you let them rate a list of animals on a scale denoting size (again from 1-10, one being very small and ten being very large). I’m willing to bet that you would find a statistically significant difference between the whale-readers and the cat-readers when it comes to the size ratings. But would you be making the assumptions that the whale-readers are larger? Or more prone to “large” behavior? Or did they simply had a different frame of reference represented?

Sure, it’s dangerous to use analogies. You could easily use the different perception as the crucial issue. They might not be larger but they might feel this way, which when it comes to violence is the only thing that counts. But one criticism still stands: It’s an effect of the scale, specifically that the scale was without fixed and clearly denoted end points. By giving the players of “violent” computer games another frame of reference, you changed the scope of the scale. Similarly, for the whale-vs-cat group, the whale group has the largest living animals in the back of their mind. No wonder they (likely, it’s untested as far as I know) underrated the size of the animals.

Whether it’s animal size or violence, these end points should be specified with a case example. And that might be hard to come up with. For animals it’s easy — a blue whale for the living ones. But what would you use for violence? Torture? Murder? Genocide? The sadistic killing of a deranged individual?

I mean, seriously, reading a book about serial killers (by John Douglas if I remember correctly) and what they did with their victims redefined what I consider as “very violent”. Does that make “lesser” forms less relevant. No. But it puts them into perspective.

And frankly, I think that’s missing a little in today’s Western world. Luckily, extreme cases of violence are extremely rare. But I wonder whether this also means that we react overly sensitive to things that are not really that much of an issue. And perhaps this is the main criticism to “violent” computer games. They give us a glimpse at what people are capable of. Take the torture scene in GTA V. Sure, it’s over the top, it’s cliche, but are we really to believe things like this do not happen? By the “good” guys (and girls)? And what’s the problem here, that a person is tortured, or that this information is used to kill someone?

So, if games with violent scenes are criticized, perhaps the arguments should be better than a few points on a scale that is more a measurement artifact than an actual effect. After all, that might be an issue — and that these games show us a world that does or can exist, whether we want it or not.


  1. Frame of reference is a vague term describing contextual learning that has implicit effects on memory and habit-formation. The concept of time is not weighed into your argument as the amount of contextual learning greatly influences the likelihood of habit formation. To draw another analogy: doing/seeing something repeatedly leaves a contextual after-image that is more intense with more learning — also related to neural plasticity and neural network modification.

    Desensitization is a well-established principle, and the only thing it really implies is that your reaction to something is less intense than would be expected from the population. Such a reaction is definitely there, otherwise desensitization wouldn’t be a construct.

    I pose another analogy that is also flawed (because it is about desirable behavior), but that illustrates a different perspective: If I rate regular sexual intercourse as less arousing, will I look for things to increase arousal? Will I be more likely to condone and/or perform sexual activities that are more “extreme”?

    Now say for some reason violence is arousing to a person (i.e. with violent tendencies), do you think the analogy upholds?

  2. A lot of good points.

    First of all, I think that the arousal or desensitization tapers off. Sure, you can learn from observation, and you might even establish habits, but based on the data I am not convinced that so-called violent computer games actually make the person more prone to condoning or actually showing violent behavior. Even here when it’s rating behaviors in terms of violence — why should a person also be more likely to do these behaviors? Would this person even be able to do? It might not be as obvious as the size-analogy, but there’s a difference between play-shooting a gun, and shooting an actual gun. Or to look at sociopaths, I guess that some find out that actually doing what they imagine they would like is not what they expected it to be.

    Regarding increased sexual arousal — nice analogy, but here’s the same issue. If we are talking about sexual arousal in general, I think it tapers off and thus can increase over time. I am hesitant to do the comparison, because it does not fall under the heading of violence, but look at people interested in SM. They don’t escalate until they off themselves or others. They may have tastes that are not mainstream and might include things that look like violence, but here the same: Arousal tapers off over time and can increase again. I would not call SM violence however, as it is consensual. If actual violence is arousing for a person, I would not call the person kinky, but criminally insane. But even here I would expect it tapering off over time until it rises again. At least arousal in general sense, that sex with the same person can become “routine” is another issue.

    BTW, I left out the definition of violence, so here’s that addendum: I would consider something as violent when it is an action intended to harm/cause pain to another person without this person’s consent. That excludes e.g., a surgeon performing an operation (if the patient’s consent is given/implied), SM, “violent” sports like boxing or MMA, and similar activities.

    But all these are side notes as I am not convinced that the measure actually detects an enduring change in the person.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.