This article came off the press today and I spotted it during the evening on Ohio State University’s news site while looking for ways to make Qualtrics work the way I want for a summer project. Graduate colleague Jodi Whitaker (Ohio State University) and my advisor Brad Bushman (Ohio State University) has a paper as of this writing, in press at Communication Research.
Video games are excellent training tools. Some writers have called violent video games “murder simulators.” Can violent games “train” a person to shoot a gun? There are theoretical reasons to believe they can. Participants (N = 151) played a violent shooting game with humanoid targets that rewarded headshots, a nonviolent shooting game with bull’s-eye targets, or a nonviolent nonshooting game. Those who played a shooting game used either a pistol-shaped or a standard controller. Next, participants shot a realistic gun at a mannequin. Participants who played a violent shooting game using a pistol-shaped controller had 99% more headshots and 33% more other shots than did other participants. These results remained significant even after controlling for firearm experience, gun attitudes, habitual exposure to violent shooting games, and trait aggressiveness. Habitual exposure to violent shooting games also predicted shooting accuracy. Thus, playing violent shooting video games can improve firing accuracy and can influence players to aim for the head.
If I can finish my paperwork for the summer project, I can move on to someone’s work.
The question they sought to answer, which concerned parents among others, is whether violent shooters can teach how to accurately shoot a gun. If a response is a ‘no’, then I would like to know about how and why a shooter video game cannot impart hand eye coordination skills. Also, please tell me how and why some researchers using other videogames have found that surgeons who played videogames had better laparascopic skills or the whole cognitive science literature with action videogames. Please email Daphne Bavelier (University of Rochester), Walter Boot (Florida State University), C.S. Green (University of Wisconsin-Madison) among others whose work are related to this study.
The authors posited several theoretical reasons that explain why shooter videogames would teach a person how to shoot a gun accurately. First, shooters use operant conditioning to effectively shape players’ skills according to set criteria. Enemies are aversive threats that when dispatch out, players are rewarded for avoiding to themselves, a negative reinforcement. On the other hand, players who learned a particular set of skills over the years have made a living nightmare for other players because of their skilled ways of getting headshots. They are likely to have been rewarded frequently by the computer’s positive praise (e.g. those headshot sound bites) over the years, a positive reinforcement and reward mechanism, according to social learning theory. Of course, a real gun instructor would say to aim for the chest…
A last theoretical reasoning posited by the authors is mechanistic transfer where training tools that closely approximate would greatly facilitate the skills and behaviours of one training context to the next, but actual context. This mechanistic transfer actually makes sense when you consider how swordsmen learned their skills through wooden swords and not solely by watching a sword fight. The same logic would apply for shooting games where a gamepad controller is not an accurate approximation in comparison to light gun controllers.
Participants: 151 undergraduate students, average age is 21. If I remember, they were mostly recruited from the communication department.
A series of questionnaires were used to account for possible differences among the participants.
Trait Aggression: the 29-item of the Aggression Questionnaire by Buss & Perry. Response scale is a 5-point agreement scale.
Attitudes Towards Guns Scale: a 17-item scale through an 8-point agreement response scale.
Attitudes Towards Guns and Violence Scale: a 23-item scale through a 3-point agreement response scale.
Firearms training: yes or no. I once asked the first author, Jodi Whitaker, if she took into account as to how recent they fired a firearm.
Videogames played: Participants listed their three favourite videogames. These were coded whether they involved shooting.
Videogames used: Resident Evil 4 as the violent and shooting videogame. Wii Play’s target practice game as the non-violent and shooting videogames and finally Super Mario Galaxy as the non-violent and non-shooting videogame. I believe the Wii-version of Resident Evil 4 was used. Two versions of the shooting videogames were used, one version was using the gamepad controller and the other version is the pistol-shaped controller.
I don’t recall that well since she presented her findings about a year ago in the research group I was in. What I remembered is that we were having a long discussion about the videogames. My fellow graduate colleagues and I were nit-picking every structural differences between Wii Play and Resident Evil 4 such as the former featuring bull’s-eyes whereas the RE4 featured zombies, we were wondering if the bull’s eyes were human-shaped would have similar effects. I was being nit-picky about whether shooter games increased interest on firearms.
Pistol accuracy: An airsoft training pistol was used to measure firearms accuracy. This pistol simulates a 9mm pistol’s weight and recoil using pressurized air. Participants fired the gun with an ammo count of 16 hard Velcro balls to a mannequin from a distance of 6.1 meters or 20-feet. That mannequin scared the crap out of me when I first saw it in a dark room. I forgot its name, but right now it’s stored somewhere in a box.
Participants completed all questionnaires, then they were randomly assigned to play one of the videogames for 20 minutes. Afterwards, they proceeded to shoot the mannequin and were finally debriefed.
A 2 X 5 MANOVA was conducted with the type of shoots (headshot and the rest) to the mannequin as the within-subject factor and the videogames conditions along with the controller types as the between-subject factor.
Being so late at night writing this post, I present you this graph from the article itself.
The results remained significant after taking gender, trait aggression, attitudes, firearm experience and videogame experience into account.
They found an interesting correlational result with the questionnaire data to the shooting data. It seemed there is a positive correlation between higher violent shooting game exposure (which was the number of shooting games listed as their favourite) to the number of total hits (r = .20, p < .02) and marginal significance for headshots achieved (r = .14, p = .09). Understandably, getting headshots with a real gun is much harder to achieve, thus probably explaining the marginal significance. In retrospect, maybe we should’ve videotaped the participants, but then again this is all correlational with no causational explanation.
I’m sure the U.S. military is happy to hear that shooting games helped train firearms skills, although I think they already knew that but kept it themselves. Of course, they still need to teach the rules of engagement and other elements not present in commercial first person shooters, but I really can’t think of one and soldiers should probably have a good idea to what is missing.
As the results demonstrated, more experienced players in shooting games and those who played Resident Evil 4 in the pistol condition tend to be more successful at getting hits and headshots to the mannequin. The authors argued that this is a function of operant condition and rewards from social learning, in particular to the target shape and controller type. The more approximate between the training task and actual tasks, the greater mechanic transfer there is. Since headshots are highly rewarded (e.g. voice praise and statistics), the participants were incentivized to try getting headshots in the real life experiment. For a firearms instructor, I imagine it’s a terrible habit that needs to be put down from the cocky rookie.
Speaking of headshots and rewards, I remembered speaking out whether other parts of the body would have also been targeted (e.g. the groin) and rewarded as well as the headshot. At the time, Sniper Elite V2 was not announced… I wonder if we can replicate that by manipulate reward values to different parts of the body, say emphasizing the groin or the chest instead of the head.
The authors listed their limitations. The airsoft gun does not simulate every aspect of a pistol, such as fire and gun smoke. Firearms experience did not even affect the gun data, although I pointed that how recent they fired a gun might’ve been a greater influence. One videogame per condition were used, so who knows if Sniper Elite V2 might’ve been a better learning tool than RE4. The participant did not report where they were aiming at for each shot (e.g. chest, groin, arm, head, etc.). The shooting distance is a 6.1 meters or 20 feet, I don’t know if that’s a realistic distance at a gun range. Long-term research on shooting games and firearms skills is needed, although I think the military should do that kind of research…
The authors left this final note: playing a violent shooting videogames does not influence a person to shoot a real gun towards an actual person, not in the context of the experiment. In the case of a zombie apocalypse, at least they’ve got the useful skills, but will need the will to fire.
Related readings: http://www.citeulike.org/user/waiyen/tag/game_interfaces
Whitaker, J. L., & Bushman, B. J. (2012). “boom, headshot!”: Effect of video game play and controller type on firing aim and accuracy. Communication Research . DOI: 10.1177/0093650212446622