Cannon fodder NPCs, dehumanization and aggressive behavior (Greitemeyer & McLatchie, 2011)


Via Psychology Today, Dr. Art Markman (University of Texas) posted about an article published in Psychological Science about videogames and the dehumanization effect, authored by Tobias Greitemeyer (University of Innsbruck) and Neil McLatchie (University of Sussex).


Past research has provided abundant evidence that playing violent video games increases aggressive behavior. So far, these effects have been explained mainly as the result of priming existing knowledge structures. The research reported here examined the role of denying humanness to other people in accounting for the effect that playing a violent video game has on aggressive behavior. In two experiments, we found that playing violent video games increased dehumanization, which in turn evoked aggressive behavior. Thus, it appears that video-game-induced aggressive behavior is triggered when victimizers perceive the victim to be less human.

Dr. Markman wrote a good gist of the article, I’ll fill in some of the details.

Dehumanization is “a process in which people disengage their moral self-sanctions, thereby relieving themselves of feelings of guilt over their aggression actions” (Bandura, 2002). So, perpetrators of violence equating others, either a group or an individual, to a non-human status, either as an animal or robot, can relieve themselves from cognitive dissonance or feelings of guilt.

Moral disengagement (see Hartmann & Vorderer, 2010) can play a similar role in dehumanization as we justify our aggressive behaviours because we don’t feel guilty in destroying a human-looking object, rather than killing. On a similar note, we might say we objectify other groups, such as sex or ethnic minorities. In videogames, this is true that we perceive the faceless, same skins, or clones as mere bots or “mobs” in MMO vocabulary:  “Non-player entities whose primary purpose is to be killed for experience, quest objective, or loot”. Red shirts and cannon fodders are recognized tropes within television and movies as well. We see the countless stormtroopers, orcs, and zerg being cannon fodder for the protagonist and perhaps to serve the seriousness of an epic battle.

So, does being primed to dehumanize others in videogames, also translate or persist to other human beings in a social psychology experiment?

Experiment 1

This experiment looked into inter-group dehumanization.


Participants: 60 students from the University of Sussex. Average age is not mentioned, 36 women and 24 men.


Ten-Item Personality Inventory: Greitemeyer and McLatchie decided to use subtle methods to measure dehumanization. The inventory asked participants to rate how each of the Big Five personality factors apply to Britons (an in-group) and to immigrants (an out-group). This was used in previous studies and it was shown that neuroticism and agreeableness was rated as the least human, whereas openness and conscientiousness were as the most human traits.

Videogames used: Lemmings as the prosocial videogame, Lamers as the violent videogame, basically an evil version of Lemmings, and Tetris as the neutral videogame. These games were picked because they are similar in eliciting similar levels of arousal and mood.


Participants are randomly assigned to play one of the three videogames for 15 minutes. Then they complete the Ten-Item Personality Inventory. The participants were told that they were participating in two completely unrelated studies, so as not to arouse suspicion, just so you know.


Their analysis found that those in the violent condition rated the out-group (i.e. the immigrants) as possessing fewer human traits than the in-group in comparison to the prosocial and neutral condition, whom the two latter groups did not differ from each other. On the flip side, those in the violent condition attribute the out-group with more non-human traits than the in-group and in comparison to the other conditions. Sex differences were not a significant covariate.

Experiment 2

Let’s try inter-personal dehumanization, but without a prosocial videogame, and with different dehumanization measures.

Participants: 40 students from the University of Sussex. 29 women and 11 men.


Human uniqueness and human nature traits: a 20-item of positive and negative human traits, such as broadmindness, disorganized, active, patient, etc. Participants rate themselves and another participants (who I shall describe later) on these 20 items.

Attributions of emotions: a 16-item of emotions, 8 of them are primary emotions (e.g. pleasure) akin to basic emotions that people believe are not uniquely human and are shared with animals. 8 are secondary emotions (e.g. hope) of which looked like complex emotions or mixes of the primary emotions that makes people human. Participants were asked to rate the extent the other participant is likely to feel for each emotion.

Aggression measure: Participants were asked to write an essay on a subject, in this case, the British National Party. The Participants were informed that “another” participant (a confederate) is also writing the same essay and both are to judge and rate the essays. Of course, the participant never saw that confederate (i.e. doesn’t exist, but don’t tell the participant). Anyways, the “other” rated the participant’s essay as a poorly written essay in a means to piss off the participant. To measure aggressive behaviour, the participant was asked to judge the “other” by answering four questions on a 7-point scale. These four questions were framed in that the “other” is applying for a research assistant job at the university. I gather that being pissed off might make you not wanting to help that wanker from getting that job.

Videogames used: They picked based on previous studies: Wolfenstein (not sure which versions) as the violent videogame and 3D pinball as the neutral videogame.

Procedure: So, participants started out writing the essay and evaluate the “other” participant’s essay. Then they played one of the two videogames for 15 minutes. Then, they were asked their thoughts about the videogame. Then they get their essay evaluation from the “other”, it was bad.

So, participants went on to complete the Human uniqueness and human nature traits inventory and the attributions of emotions. And then finally, they were asked to judge the “other’s” job eligibility in getting that research assistant job.


Dehumanization: those in the violent condition associated less positive human-uniqueness and human-nature traits to the “other” participant. They associated them with more negative human-uniqueness and human nature traits to the “other” participant. As for emotions, participants attributed the “other” with less secondary emotions. As for their self-assessment on all three measures, there were no differences between the violent and neutral condition. In essence, participants in the violent condition treated the “other” as less human and despite the fact that even the neutral condition group was pissed off by the wanker “other” for giving a bad evaluation. So, I reason to say that those in violent condition were easily pissed off or treated others dehumanly. Sex was not a significant covariate.

Aggressive behaviours: those in the violent condition (M = 3.65, SD = 0.65) judged the “other” less positively than those in the neutral condition (M = 5.24, SD = 0.82).


The take-home message is that a dehumanization process was found in violent videogames and also adds up to the known ways, such as priming, that can increase aggressive behaviours. The outcome of this dehumanization effect has translated or persisted towards other people, either an out-group or an individual.

The authors noted that violent videogames did not increase the in-group or the participants’ own humanization, so they didn’t ascribed themselves as more human, but merely dehumanized out-groups or individuals who pissed them off. The authors cautioned that the causal process is a bit murky in terms of whether dehumanization is a mediator for aggression or that aggression is a mediator for dehumanization or in simple terms, which comes first as a causative factor. Both pathways are supported in their analysis. However the authors reasoned, methodologically, that dehumanization came first since they measured aggression after measuring dehumanization.

The authors mentioned something interesting that perhaps people might dehumanize others through mechanistic type, by attributing others to machines or objects. This is a very possible and quite relevant to videogames as the earlier mention of “mob” where the definition of a non-player entity whose sole (and mechanistic) purpose is to be killed for rewards. Going further, the mobs are simple and predictable entities that made easier to dehumanize and perhaps feel a bit of pleasure in killing them or not feeling any remorse. Since the mobs are all alike, it gets easier to become desensitized and to dehumanize since the mobs, despite some having a face, are all becoming “faceless” killing machines. They’ve become the antagonists’ red shirts, cannon fodder for the protagonist and more importantly an expendable and invisible minority group (from a certain point-of-view) for everyone else.

The authors asked how violent videogames affect dehumanization. They did not speculate what, but I’ll offer my insights. First, the dehumanization process is a structural problem inherent in video game design. Video games cannot render each individual enemy as a unique individual, in that the required resources are just astounding at this time. Sure, there are attempts for differentiation through skin palettes, multiple voice actors for one character type, the presence of conversations, social presence, multiple A.I. personalities, and such. But the varieties may be just so few, by the time we’re half-way through a game, we might have become familiar with them and might already start dehumanizing them. Because cannon fodders are part of the game design, it just becomes part of the culture. Breaking the cannon fodder mould would mean some innovation in videogame design.

Of course, the NPC’s characteristics might affect dehumanization, perhaps robots or animals might not affect the dehumanization process towards real human beings. Perhaps, having some persistent antagonists, dubbed as mauve shirts by tv tropes, whose role is the cannon fodder whose purpose is to lose and come back again on the next level. I reasoned that since players would be interacting with the same characters that they might start relating to them as rivals rather than rats. The situation may be different in MMO, specifically player-versus-player environments, where players don’t dehumanize others because they know people are controlling avatars and do show human behaviours and feelings, unless they are assholes themselves.

I am not sure if the issue of dehumanization is being addressed in videogames. The most relevant case would be America’s Army, Robertson Allen’s (University of Washington) piece in Games and Culture explored the characterization of enemies in the game.


This paper explores the characterizations of enemies in military-themed video games, with special attention given to the games Conflict: Desert Storm and America’s Army. I demonstrate how the public enemy of America’s Army is one not confined to any nationality, ethnicity, or political agenda. This marks a significant departure from games such as Conflict: Desert Storm. I argue that the production of this abstract enemy—what I call the ‘‘unreal enemy’’—is significantly shaped by a biopolitical system that intertwines the military and electronic entertainment industries. This arrangement delocalizes power, distributing it through a network of institutions and subjects. Throughout, I use ethnographic examples that explore how this abstract enemy has been constructed and juxtaposed against more concrete and personal figures, such as the America’s Army Real Heroes, individuals upheld as the embodiment of personal achievement in the U.S. Army. I conclude by asserting that the unreal enemy of America’s Army is, ultimately, an enemy that is not exclusive to a video game, but one that exists as an anonymous specter, ever present in the militarized American cultural imaginary.

The question of America’s Army dehumanizing its political enemies is up to the media scholars (see Mantello, 2009) to scrutinize, it would be interesting to examine, but this is not my cup of tea. On a broader outlook, I believe the first-person shooter genre would be the focus of dehumanization/videogame line of research.

Some lingering questions remain that I cannot speculate about: what the short-term effects? Do they persist more than 10 minutes? What if the videogame actively demonize some character groups? What are the long-term effects? Greitemeyer and McLatchie suggested a spiralling effect where dehumanization increases aggressive behaviours which in turn increase dehumanization. In an isolated spiral system, that may hold, but other factors come in that may weaken this spiral effect, say narrative closure where the enemy and hero reconcile, huh just blabbering on this point right now.

Greitemeyer, T. and McLatchie, N. (2011). Denying humanness to others: A newly discovered mechanism by which violent video games increase aggressive behavior. Psychological Science, 22(5):659-665. doi: 10.1177/0956797611403320


3 thoughts on “Cannon fodder NPCs, dehumanization and aggressive behavior (Greitemeyer & McLatchie, 2011)

  1. peered over @ Markman’s comments section and something interesting from commenter SDGundamX:

    Dehumanization is indeed a real problem in our society. Probably the best example of it is how the media portray women in advertisements and popular entertainment (i.e. mostly as objects to be desired or used).

    But I have a problem with the notion that violent video games cause dehumanization. The author wrote “Violent video games often treats the victims of the violence abstractly” and I think this is the key point that ironically undercuts the author’s entire argument.

    All players are aware, whether the are blasting each other or computer-controlled aliens, that they are playing a game–an abstraction of reality. The rules work differently within the game from reality. Unless they’ve had some sort of psychotic break with reality, they don’t believe they are killing “real people” when they are playing. While violence may be rewarded in the game, again–unless someone has had a psychotic break–they don’t suddenly assume that violence will be rewarded in reality once they stop playing the game. Hence, I find it really difficult to believe that playing violent video games has any significant effect on the player’s actions outside of the game.

    One problem I’d like to point out about the study that claims to show violent video games cause dehumanization is that the construct seems questionable. The researchers in the study claim that when you dehumanize someone you will associate them with personality traits most typical of animals. What empirical evidence is there that this is the case in actuality? For example, have those who participated in the Rwandan massacres been clinically shown to exhibit such behavior? If so, how do these participants in real-life massacres compare with the video game players in the study in terms of their performance on this experiment? I would love a link to the original article so I could see what evidence the researchers have given that shows their construct is valid and not purely hypothetical.

    Markman reply:

    Thanks for the comments.

    Check out some of the research by Nick Haslam and his colleagues. There really does seem to be a general link between dehumanization and aggression.

    If you look back in the history of genocides, you will see a definite stream of dehumanization as part of the process. Nazi propaganda, for example, tended to portray Jews, Gypsies, and other groups as animals. I think that certainly played a psychological role in the Nazi genocide during the Holocaust.

    SDGundamX reply:

    I’m aware of the strong link between dehumanization and violence (I know you wrote “aggression,” in your reply but let’s be honest–people are much more concerned about game playing leading to violent behavior than aggressive behavior). What I’m not aware of is a study that empirically demonstrates that Haslam’s theoretical model of dehumanization can be operationalized to measure what it claims to be able to measure (i.e. dehumanization). In his own article on the topic he admits that “it remains to be seen whether the model provides a useful framework for research” (Haslam, 2006, p. 262).

    In other words, a study that demonstrates people playing violent video games show signs of dehumanization using Haslam’s model really doesn’t tell us much if we don’t know that Haslam’s model accurately can be used to measure dehumanization in the first place. Also, it should be pointed out that Haslam designed the model not to only include the extreme cases of dehumanization (of which violence against another would be included) but to “extend the scope of dehumanization as a concept” (p. 262) to areas that are much more subtle. So even assuming that both the construct and the measurement are valid and we are detecting dehumanizing attitudes arising in players of violent video games, how sure are we that this effect is significant (not in the statistical sense, but in the sense that it has an effect on their actual behavior)? Without a longitudinal study how do we know if these effects last beyond a few minutes after game playing? There seem to be so many unknowns here that it hardly makes sense to get excited over these findings.

    Haslam, N. (2006). Dehumanization: An Integrative Review. Personality and Social Psychology Review, 10(3), 252-264.

  2. Pingback: Being a victim of videogame aggression and your social cooperation (Rothmund et al., 2011) « VG Researcher

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