When a related study was posted on ScienceDaily, it reminded of a long forgotten post and was misplaced in my files. The two studies are related in that it dealt with stereotype priming within computer-mediated environments. Matthew Eastin (University of Texas at Austin) had published a study on the effects of playing as black avatars within a first-person shooter.
The current study examines the impact of racial representation on character identification and postgame play hostility. Examining data from Black and White participants, results suggest that cueing racial attributes influences identification and elicits stereotyping and hostile outcomes. Specifically, White players displayed more hostile thoughts when playing as a Black character than they did playing as a White character. Black participants had more hostile thoughts when playing against a White opponent than they did playing against a Black opponent. This research supports previous literature suggesting that self, character, and opponent interact to influence game-play outcomes.
Be mindful, this study is done by those experienced in video game research (Matthew Eastin and Vincent Cicchirillo) and ethnic minorities in media (Osei Appiah).
In any game with a story and human protagonist(s), as players, we connect these characters to our self during our gametime. The question is we either identity “with them” (i.e. in most video game genres) or “as them” (i.e. in most first-person shooters) and it depends how much we’d identify with our character. The authors explained the identification theory that people would unconsciously assess their character’s similarity to themselves and identify with them according to the assessed level of similarity.
They further elaborated identification with the distinctiveness principle that people are motivated to distinguish themselves (or their own social group) from others because it helps them maintain one’s self-esteem through social comparison. Three factors are specified in their paper: difference, separateness and position.
Difference refers to distinctiveness based on perceived intrinsic qualities, abilities, opinions, or physical characteristics of a person. Separateness represents feelings of boundedness, independence, or physical/perceptual distance from others. These feelings of separateness may be seen by physical boundaries (e.g., gated housing community) or the use of symbolic boundaries (e.g., military officer’s uniform stripes used to differentiate rank). Lastly, position is associated with a relational orientation such as the role or socioeconomic status of the individual or group.
In the case with African-Americans, one might distinguish themselves based on skin colour, living in a black neighbourhood and based on their socioeconomic status. It is surmised that video game characters and players who share similar distinctive qualities would have players show more favouritism and identification with the characters. This is especially the case when an individual is part of a minority group or a group with a salient and valued social category, such as age, gender, race, number of people within the social group or (I might add) interests.
If you go back or remember the abstract where it writes that white participants displayed more hostile thoughts playing as a black character than a white character. This is based on cultural stereotyping of blacks being attributed to traits like aggression, hostility, criminality and such. In effect, white participants are acting out as a black player or character (from their own knowledge base, of course). This doesn’t mean that the white players are racist, well not openly or consciously. The authors explained that simply knowing these cultural stereotypes, especially if they’re salient and controlling for racial prejudice, and a mere exposure of a black person or anything (symbols) related to blacks can automatically trigger these stereotypes. You can thank generations of racial prejudice, lack of direct interracial contact, repeated media exposure of stereotypes, and confirmation bias among other things.
As for African-Americans about their views of these stereotypes, they obviously don’t endorse them, they’re less likely to think or behave in ways consistent to these stereotypes. Some even try not to act out these stereotypes in fear of reinforcing them.
Participants: 178 undergraduates from what I guessing is a university in Ohio. Average age is 22 (SD = 3.90), equal gender ratio. Participants who identified themselves as Black (n = 95) and white (n = 83) were included in the analysis. Participants were recruited from a 10$ reward or extra course credit, the usual.
Aggressive thoughts: the word completion task, a fill in the blanks of a word. There are 98 words and participants are given 3 minutes to complete as many as possible. These words can completed and coded either as neutral, aggressive, ambiguous or nonword. Now remember this measures only aggressive thoughts, not behaviours.
Identification: four item answered on a scale of 7. The four item asked the character’s similarity to the participant based on cultural background, appearance, basic values and how strongly one identifies with the character.
Video game experience: Participants are asked to list their five favourite games and rated them on various dimensions, like violent content and such.
Video game used: Unreal Tournament: Game of the Year Edition. Dr. Eastin’s game of choice, given his previous studies it’s not surprising. Participants play on a map where there’s a central mirrored column, this column is designed to allow participants to see their avatars’ physical appearance, in this case, skin colour (black or white). This study does not address any black characters (i.e. speaking role or active in narratives), since this could change the player’s behaviours or perception of the character. For example, Louis from Left 4 Dead. So you could say that participants are playing black action figures and anything that would prime stereotypes would be based on the physical traits of the avatar.
Participants are randomly assigned to play either a white or black avatar against a white or black opponent. They’re given a tutorial of the video game. After that, they’re given biographies and pictures of their avatar and their opponent cueing them of their racial features. I assume the race of the player behind the opponent is not revealed? Play time is 15 minutes. After the game, participants (9%) who did not correctly identify their avatar’s race and their opponent’s race were excluded from analysis. It makes think who are these participants who couldn’t correctly identify in-game race.
Analysis using ANCOVA. Controlling for gender and video game experience, they found that Black participants were statistically significantly more likely to identify with a Black avatar than a white avatar. As for white participants, they were significantly identify with a White avatar than a black one. Effect size is large, but that’s a partial eta-squared. Well there’s no surprise, except for the White participants which the authors didn’t expected.
Next is their analyses on avatar race and opponent race, so participant race is not factored in. They’ve found that players’ who played a Black avatar against a White opponent showed greater identification than playing against a Black opponent. Participants who played a White avatar against a White opponent showed lower identification than playing against a Black opponent. Effect is size is medium-large, again partial eta-squared. The authors didn’t discuss much of it, but I find it interesting in how physical cues can elicit such identification and it reminds me of a virtual reality study of participants’ avatars’ level of attractiveness can change their behaviours in a money-splitting game. This is quite interesting and relevant in how participants playing an avatar as a minority (ethnic or sexual orientation, etc.) might learn a thing or two about stereotyping behaviours online.
Third analysis is the interaction between race and aggressive thoughts. Controlling for gender and video game experience, they found significant interactions effects in that White participants playing a Black avatar scored higher on the aggressive thoughts measure than playing a White avatar. Effect size is small-medium. On the other hand, Black participants did not scored differently when playing a White or Black avatar. Referred earlier, the authors argued that White participants playing a Black avatar in a violent video game elicit Black cultural stereotypes, namely aggression and hostility. It’s because the in-game violence and the black stereotypes go hand in hand quite nicely. A reminder, participants did not act out these stereotypes deliberately or consciously, it would have been the case 60 years ago.
The authors offered another explanation which made me shudder, not that it’s an invalid explanation, but the thought of imagining the following. White participants feel liberated to embrace black culture and showing this embrace through clothing, language style and nonverbal expressions. They also indiscriminately embrace the good and bad stereotypes. In short, white people imitating black people. The second author calls this as a form of cultural voyeurism.
Where do they get their info on Black culture? Is it from their Black friends or from the media, such as music, television, movies and maybe video games? A quick glance in google scholar reveals little research on racial stereotypes in video games. However, it is getting a lot of attention and there are several games (e.g. Grand Theft Auto or 25 to Life) that some people would say is racially stereotyping.
Fourth analysis is the interaction between aggressive thoughts and the participant’s race and opponent’s race. There were no significances found, but they reported a statistical trend (p = 0.06, close but not significant). That statistical trend showed that Black participants scored higher on aggressive thoughts when they played against a White opponent. White participants reported no significant difference against a White or Black opponent. Please take this with grains of salt, they argued that it could be that violent video games provide a safe environment for Blacks to “indirectly address social grievances and cathartically release frustrations against Whites” due to the legacy of slavery and racial segregation. Or IMO, that Black participants who are part of a minority group (and may have stronger in-group favouritism) are more hostile to outgroup members in particular those that are in sharp contrast to them (say skin colour).
Well, it’s quite a nice study, but a lot of work and grant money needs to be done. The authors offered a suggestion for future studies. The examination of in-game violent behaviours, say the differences in excessive in-game violent behaviours or random violence vs. required violent acts in a given game. IMO, other suggestions in this line of research should investigate characteristics of minority video game characters (are they stereotypical or fleshed out nicely?), video game characters’ behaviours and attitudes towards characters of minority group, names associated to a minority group (say Tyrone, what would others behave to a player with that name), does it generalize to other parts of the world or does stereotypes adversely affect prosocial behaviours online?
The authors take-home message is something echoed in old media research. Media producers, from radio to video game designers, should be aware that whatever they produce is consumed by people with perceptions based on real life and quite often interpret at that way. If whatever they watch is in agreement with their beliefs (bad or otherwise), then it would reinforce such beliefs. On a loosely related subject, a media producer whose aware of such phenomenon (and I’m sure nearly every video game designer are) could use this to their advantage to engage players in highly emotional or upsetting situations in order to bring out serious discussion.
Eastin, M.S., Appiah, O., & Cicchirllo, V. (2009). Identification and the influence of cultural stereotyping on postvideogame play hostility. Human Communication Research, 35, 337-356.