Wishful identification in videogames and the effect on aggression among adolescents (Konijn, Bijvank & Bushman, 2007)

In this study, they investigated on game characters, mainly the protagonists, effect on adolescent boys as role models. Indeed, we often look up to other people who serve as a role model in order to interact with society, how we ought to act in different situations, how to talk with other people, etc. So who do we look up? That depends on what you’d like: say your parents, a celebrity, a fictional character (say Sherlock Holmes or Gil Grissom from CSI), you get the picture. So anyone, real or fictional, can be a role model for children and prior research has been done on older media. Anyways, they want to see if game heroes have any effect on players in terms of aggressive behaviours. Their findings suggested that participants are more aggressive when they played violent video games and when they identify with the main character within said game. The effect is pronounced with realism and immersiveness.

There are two types of identification explained in the study: similarity identification and wishful identification.

  • Similarity identification is identification where a person’s role model has similar characteristics to one’s own, which leads to liking that character more than anyone else. For example, a nerd is likely to identify a nerdy character in a show. Or macho guy like a character who acted like a tough guy.
  • Wishful identification is where characteristics of a character are attractive to that person who does not have them. For example, a weakling likes someone who’s strong and has guts to act out while he does not possess them.

So, like I mentioned before, what people identify with will likely learn their behaviours and attitudes. So in most violent video games, heroes behave in an aggressive manner. The researchers did make a good argument that boys are generally rewarded for what we see as gender-typical behaviours. For example, being tough, competitive, bravery, the saying “boys don’t cry”, etc. So, in a sense, heroes in violent video games engender both similarity and wishful identification since they display manly characteristics. This is relevant for adolescents who are searching an identity.

So what makes video games different from television and other media, well first the player plays the protagonist him/herself. In my opinion depending on game’s narrative, a player may either go through the story linearly and the character will develop accordingly without much player input. Or the game is open-ended, like Grand Theft Auto, where the player while may identify the character, but player has input into the character’s development. Say a criminal becomes reformed thanks to the player’s actions and vice versa. But this is my opinion.

According to the researchers, there are characteristics as the level of identification one may develop: violence and realism. Violence because being aggressive is what men are like, right? As they pointed out in their paper, “real men are not sissies”. Realism, unless you have lost your sense of reality, is where you know that being Master Chief of Halo is impossible because he’s a fantasy superhuman, while being a member of a SWAT team or the tragic Max Payne is something possible and easier to relate to. So, in a sense, what determines realism isn’t solely graphics, but that these heroes are believable and possible. Although the researchers believed graphics to be a key to realism and immersion, therefore wishful identification, but I don’t buy it with my earlier reasoning.

Other factors included are player traits like aggressiveness and sensation seeking. I guess activities that satisfy these traits attract people to them and video games are one of them and would likely generate identification.


Games are selected based on four categories: violent-realistic (America’s Army, Killzone and Max Payne), violent-fantasy (Doom 3, Quake and Metroid Prime), non-violent-realistic (Pro Evolution Soccer, Sims 2, Tony Hawk), non-violent-fantasy (Mario Kart, Mario Sunshine and Final Fantasy?). These games were chosen according to a group similar to the experimental subjects. So the researchers didn’t pick the games and they did it so it can be generalized.

Participants: This is interesting: 99 boys from middle school in the Netherlands from “VMBO-class” which they say that these boys have the lowest educational ability and are mostly from low-income households. (see wikipedia here, says here that 60% of students are enrolled in that level) These are important factors to aggression according to common knowledge and should be noted. Applicability to U.S. is uncertain due to cultural homogeneity, but may apply to other homogenous societies. Another thing, it would’ve been a larger sample, but 9 had to drop out because they don’t understand how to play video games, and 4 who don’t follow instruction. Well it can’t be helped since not 100% of youth population play video games.
Measurements: questionnaires measuring their trait aggression, sensation seeking and their playing history. Given two weeks prior to testing. Although there are few questions, but Cronbach’s alpha is okay, but I’m still skeptical of its validity…so few questions…

During testing, participants are randomly assigned to play in one of the 12 games. Play time: 20 minutes. After that, aggression measure testing: This is done in previous experiments: two players compete by blasting noise against each other on 25 trials. What they’re looking for: the first trial measures unprovoked aggression by looking at how loud the player sets, from level 0 (no noise) to 10 (loud as a smoke alarm). Then reciprocal aggression on subsequent trials.

Please don’t say that it’s unrealistic, well we can’t do anything worse than that or someone will sue. Participants are convinced that it’s realistic and it’s been done and studied before. Even if there’s other ways, we must have consistent measures or someone will burnout from stress, SO from now on I’ll call the experiments using the blastnoise as the Bushman’s noise blast experiment series. Update: actually it’s called the Taylor Competitive Reaction Time test, thanks to Ferguson’s (2007) article critizing the meta-analysis of video game research.

After that, they are given questionnaires measuring for wishful identification, immersion, realism, ratings on video games they played in the experiment. Again, they asked few questions, but Cronbach’s alpha is okay, but still…validity…


They used hierarchical regression analysis (don’t know how it’s done) to analyze data.

Wishful identification to main characters in violent video games, violence in video game and sensation seeking are positively related to aggression. What about trait aggressiveness? It came close, but not close enough to be significant. But they did rule out the possibility that more aggressive players identify more with aggression characters. But I don’t buy it, needs more research and other studies.

Wishful identification to characters in non-violent video games is not significantly related to aggression. Is it because boys don’t identify with non-masculine models or is there a methodological problem (video games used in study)? or non-violent characters are not as realistic or convincing to these adolescents boys.

Immersion and realism has no effect on aggression, so you can’t disprove or prove realistic games as murder simulators. But they affect wishful identification; so having a realistic and stereotypically aggressive hero is problematic. To parents, don’t let your kids have role models who are realistic, believable and possible. They may act out the heroes they realistically believe in. But fantasy characters like the Master Chief are okay, since kids are taught to distinguish fantasy and reality with the help of cartoons in TV. Actually follow the lego rule, an advice from a gamer parent. However, the results of this study is not authoritative and like all scientific study should be viewed as tentative and not rightly or certainly conclusive. Like always more study and more funds is required to acertain the accuracy of results.


Well, I’m glad that they started nit-picking game characteristics and see what influences our behaviours, instead of looking directly into content. Take That! The researchers identified limitations: what if players identify more with the victims of violence? I have a great idea, what if we make players play the ending level of Metal Gear Solid 3 and have them watch the ending scene. It made me cry. Make the players realize what they’ve done throughout the game is wrong, in the sense that it brought pain and suffering to Naked Snake. In addition, what if the hero suffers more violence and pain than he creates like he’s the victim in the story that’s something to consider and Japanese media has done that. Empathy in video games is the keyword for this argument.

The amount of time (20 minutes) is another limitation that I see throughout the literature, because of individual differences players might interpret or react to games differently. Given the number of participants in the study, this should not pose a problem because having a sufficient number of participants would minimize the effect of individual differences. However, it is still unreasonable to assume that gamers would have sufficient time to identify characters as a whole. In 20 minutes, players would only have about a first impression perspective on the character and not their whole personality. For example, it’s like seeing Jack from Lost as a competent leader at first, but we later see him as an insecure guy.

Third, the narrative design of each game is different, some games develop their characters at different pace, different qualities and depth and perhaps focus, this may affect the results. However, that question was beyond their scope since they’re interested into immersion and realism of which I presume they’re basing the assumption that players are assuming the role of the protagonist.

Another thing, gender-typed behaviour and modelling is dependent upon the person’s knowledge about gender-typed behaviour, so a parent might teach a kid everything about being a guy excluding any aggressive behaviours. However, this is improbable in some cultures IMHO due to ascribed values for some behaviours. For U.S., machismo in some parts of the country is considered good attitude.

More criticism, er… I mean observations. Confounds can be considered in the methodology, such as the categorization of video games and prior experience with the selected video games. My observation is that they grouped the games regardless of genre, they did so in order to generalize the results. This is quite valid since a study showed that males are engaged into a game regardless of point of view (Farrar, Krcmar & Nowak, 2006). Prior experiences with a selected game may affect the results because the experienced player may have a different perspective about the game in general and the protagonist. Furthermore, I would to conduct a comparison study between video game identification and television identification to see if there are any significant differences. Theoretically, there should be some differences, but we must ascertain that theory anyways.

Finally, their research is correlational in nature I should repeat that.

Some recommendations for developers: keep creating fantasy characters, impose in-game limits to behaviours, like don’t shoot civilians, don’t create unnecessary destruction or make the hero cringe or indicate his negative reaction to the player’s bad behaviour (some of those suggestions is already, but must repeat myself anyway); make well-developed and complex heroes, make heroes who has bad reaction to violence he makes, like regretting to kill people or saying his attitudes about violence. Make good stories about the follies of war. Brothers-in-Arms is a good example.

Well I’m done here. I’ll have start studying for my GRE, if I want to get into grad school.

Konijn, E. A., Nije Bijvank, M., & Bushman, B. J. (2007). I wish I were a warrior: The role of wishful identification in effects of violent video games on aggression in adolescent boys. Developmental Psychology, 43, 1038-1044.

Farrar, K. M., Krcmar, M. & Nowak, K. L. (2006). Contextual of violent video games, mental models, and aggression. Journal of Communication, 56, 387-405.

Ferguson, C. J. (2007). Evidence for publication bias in video game violence effects litterature: A meta-analytic review. Aggression and Violent Behavior, 12, 470-482.


6 thoughts on “Wishful identification in videogames and the effect on aggression among adolescents (Konijn, Bijvank & Bushman, 2007)

  1. And I’m not even sure, if the participant selection has any significant effect in the study. Although, the authors did mention that limitation and asked for more studies are needed.

  2. Pingback: Technological advancement and violence exposure (Ivory & Kalyanaraman, 2007) « VG Researcher - Psychology

  3. Pingback: Appeal of violent video games to lower educated aggressive adolescent boys (Lemmens & Bushman, 2006) « VG Researcher - Psychology

  4. Pingback: Media Violence Researcher’s op-ed on Grand Theft Childhood and reactions « VG Researcher - Psychology

  5. Pingback: Technological advancement and violence exposure Level 2 (Barlett et al., 2008) « VG Researcher - Psychology

Leave a Reply

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

WordPress.com Logo

You are commenting using your WordPress.com 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 )

Google+ photo

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

Connecting to %s