Violent content enhances video game performance (Bösche, 2009)

Tewi Inaba (Touhou Project) is seeking revenge over the deaths of virtual bunnies killed in this experiment

It’s the beginning of January and I am already facing hurdles. A rejection letter from a graduate school because my degree is not an US equivalent degree (at least I got a refund), more TOEFL scores to send, one still haven’t received my transcripts (they assured me they are still processing the materials, but I’ll send another copy), one that I haven’t submitted yet because my dad (who is my major source of funding) doesn’t want to submit a bank statement until I get admission.

A study published in the Journal of Media Psychology has found that violent content in a video game increases video game performance. The abstract has drawn my attention as the author, Wolfgang Bösche, had interesting semantics with inhibition.


This study assesses the impact of violent video game content on players’ game performance. According to the desensitization hypothesis (Carnagey, Anderson, & Bushman, 2007), violent content may elicit negative affective responses and inhibitions, which in turn should interfere with performance. On the other hand, the players might understand virtual violent acts as a digital form of roughand-tumble play, associated with positive emotions and mobilization, which in turn should raise performance. To test these competing hypotheses on game performance, N = 50 males with no prior violent gaming experience were exposed to three different versions of a custom-made video game in which the actions to be performed were identical, though they were audio-visually presented to appear either nonviolent, moderately, or extremely violent. The results show no indication of an initial inhibition of aggressive behavior, that is, performance is elevated and remains so if the action is presented audio-visually as being violent. This supports the notion that being involved in violent video game activity is perceived as an essentially harmless acting-out of playful fighting behavior.

What brought this study into being (IMO) was his creative stretching of the “central assumption” of violent video games’ inhibition effects, a “normal negative reaction to violence” leading to a decrease or hesitation in helping behaviours, to another activity: video game play performance.

The author’s writing style would likely upset any hardcore gamer in less than a paragraph. The article’s tone makes it seem the author has a strong opinion about video games. However strong it may be, I am unsure of his position whether he sees video games as positive or negative social influence. He described two schools of thought on video game effects, one which is clearly seen as anti- violent video games and the other as pro-violent video games. For example, he referred that the anti-video game would see violent video games acting as a “desensitization-training procedure” that phrase alone evoke images of Jack Thompson and Dave Grossman. He even used quotes to push this perspective.

What he has to say about the pro-video game group is straight to the point. People like to play violent video games because of the positive feelings associated with mock or play aggression, which he supported with the sales records between the bloody vs. non-bloody versions of Mortal Kombat. He then brought it the reality-fantasy distinction argument and tore it down, and took it towards another perspective when he referred to play-fighting among animals which is more real and physical than video games. Such play-fighting among animal, and possibly humans, is crucial for developing social and other competences. What an unusual, yet convincing, argument. So, this study will be testing two opposing theories, the desensitization theory vs. the mock aggression theory, using video game performance as a direct measure.


Participants: 64 male students in a German university. Average age is 24.34 (SD = 4.06), 14 were excluded because they had played at least a video game in the past week and/or had had a long video game experience. The final count comes to 50 participants with no violent video game exposure. Why he only tested male participants was not elaborated and this is a critical confound considering previous research on gender differences in gaming performance.

Video game used: It’s an electronic whack-a-mole game. There are three versions: nonviolent, moderately violent and extremely violent. Below are the images for each version. The sound effects are different corresponding to its versions’ level of violent, so a chewing sound and a “heehee” for the nonviolent version to sounds of pain to the extremely violent version.

Graphics used between the three versions of the game (should be in colour)

The music (a computerized instrumental version of Zillertaler Hochzeitsmarsch), background picture and mouse pointer are the same across all versions. You might ask why the bunny in the extremely violent is different from the rest, he elaborated later on in the discussion section that he was following the criteria for “a positive emotional context for desensitization” to occur: exciting background music, humorous, cartoonish characters, sound effects and rewards for acting violently.

Why he didn’t used any commercially-available video games hearkens to balancing between experimental control and external validity. As an experimenter, you’d have to be sure that the noise level (i.e. anything extraneous from the experiment, but relevant can affect the results of your study like music, lighting, even time of day) in your experiment is minimized or else you can’t distinguish noise from your intended independent variable. In his experiment, we know that he changed two formal features between the three versions: graphics and sound. I prefer strict experimental control for replication purposes before going to more complicated studies (i.e. actually using a commercial video game). In addition, (IMO) he used a neutral genre that everyone has the same expectations from a whack-a-mole game. If he were to use an First-Person Shooter games, everyone will expect that it will involve violence. Hence, there’s a possibility of anticipation effects which could prime participants’ aggressive cognitive schemas (I recently learned about this from a class). Another reason for concern is if the stimulus can be considered a video game, I consider it a video game since there are clear and explicit rules, quantifiable feedbacks, challenges and goals.

Procedure: Participants are individually tested and randomly assigned to one of the three conditions (nonviolent n = 20, moderately violent n = 20, extremely violent n = 10) They are given instructions on how to play and the instructions were carefully worded as to prevent any priming. For example, they used abstract words and phrases, like “targets”, “click” and “pointing with the mouse cursor”. In the extremely violent version, the participant is tasked to give a “taste” of the hammer so they learn to behave. A single point is award for every successful click, a single point is subtracted for every missed click or missed target.

Play time for a single round is 8 minutes. This round is separate into 2-minute blocks, each block is separated by short pauses. A 5-minute distraction is given between the first round and the second round. Total play time is 16 minutes.


There’s the usual practice effect, scores increase over time, irrespective of game version. What is interesting is that the average (and across time) nonviolent video game score is statistically significantly lower than the scores of the moderately and extremely violent scores. The scores between the moderate and extreme do not differ. So, from his stretching logic of the two theories, desensitization theory stipulating that violence should reduce performance because of the “potentially fearful and disgusting stimuli” versus mock aggression theory stipulating that such violence (being in a magic circle) is exciting and enhances enjoyment (not so sure about this) facilitates performance. The results supports the latter.

Alternative explanations must be addressed in order to ascertain the theory’s validity. The graphical difference between the carrot and the hammer is not a concern, the carrot and hammer operates under the same operational rules (i.e. clicking the carrot or the hammer over the rabbit’s eye generate the same positive response). Therefore, there are no speed-accuracy tradeoff found in the data. The graphics in the game might not be “disgusting and violent” enough to generate fearful reactions, even though they comply with the criteria I mentioned earlier. Or participants are already desensitized which would not explain the superior scores in the high violence scores, argued the author. He asked for more studies looking for more specific properties that can generate disgust and hence impair performance. Perhaps, some shock images are in order. Third, arousal could increase performance, but it can also lead to more mistakes, so it has been dismissed. Fourth, demand characteristics of the game versions would not be a factor since there are no additional constraints or incentives. The game only involve subtracting points and the violent game don’t punish players by death, so no big pressure there.

The author noted a peculiar observation between participants’ academic engagement. When participants were asked if they want to participate in future studies by leaving their email address, 90% of the extremely violent game group left their email in comparison to 25% from the other two groups.  But 25% of 20 participants is 5 participants, which in absolute numbers may render this observation moot given that there are only 10 participants in the extremely violent group.

The author concluded stating the study does not mean it goes against the media violence and aggression link, it might even support it. He argued that such mock aggression, while it can be recognized that it is not meant to seriously hurt anyone and is just play aggression (and it might help promote social competencies which is left out in the discussion section), the evolutionary-biological roots stemming from mock aggression experiences can prepare someone for aggression in real life situations. Again, he ends with a tone that can upset gamers: ” The effectiveness of such virtual preparation and training with widely available VVGs (violent video game) must still be assessed more thoroughly, however, because most of today’s VVGs typically contain seriously inappropriate mental models of fighting and combat.”

Bösche, W. (2009). Violent content enhances video game performance. Journal of Media Psychology, 21 (4), 145-150. DOI 10.1027/1864-1105.21.4.145.


5 thoughts on “Violent content enhances video game performance (Bösche, 2009)

  1. Pingback: The Psych Files podcast’s on violent video games « VG Researcher – Psychology

  2. Correct me if I’ve missed something, but couldn’t the results also be attributed to the violent feedback being more intuitively appropriate to the game scenario?

    It seems you could test this by also including versions of the game where you’re whacking bells (the dome-shaped ones you see on reception desks) instead of animals. If one version would provide feedback via a ringing sound, while the other produced blood and a cry of pain, I’d theorize that the nonviolent version would see better player performance because the feedback created a more intuitive game scenario.

  3. You are quite right that situational appropriateness might be one of many factors in player performance. It might be possible that creating equivalent feedback would equalize player performance between the nonviolent and violent versions.

    However, substituting rabbits for bells would still pose a between-version equivalency problem, despite the lack of intuititve appropriateness. Furthermore, finding a situationally appropriate feedback (or even a prosocial feedback) for a nonviolent game is a significant problem in game design. Perhaps, platic rabbit balloons for players to pop?

  4. Popping balloons. . .that actually raises another matter. Could desensitization to violence correlate with a general acclimation to disturbing/schocking stimuli? This may just be quibbling over the definition of “violent”, but it seems possible that a nonviolent scenario involving shocking, context-appropriate feedback (such as a loud popping noise in the balloon example) could see similar performance increases.

    • In the end, someone should replicate the findings with adjustments that we have discussed. The paper itself contains enough information to replicate the study, albeit a few programs would be difficult to get. Better contact the study’s author.

      As far as I understand desensitization, I don’t find popping balloons to be disturbing/shocking. It could be your wording, but I get your meaning. I don’t think there’s any meaningful connection between the balloons and violence. And when I get into grad school, feedback characteristics will be one of my top priorities.

      If you’re interested, there was a study by Staude-Muller (posted on July 2008) that used a modded FPS game to compare violent and nonviolent version and verified desensitization.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s