Hostility differences between cooperative versus competitive play in video games (Eastin, 2007)


Most research on violent video game play suggests a positive relationship with aggression-related outcomes. Expanding this research, the current study examines the impact group size, game motivation, in-game behavior, and verbal aggression have on postgame play hostility. Consistent with previous research, group size and verbal aggression both displayed a significant positive relationship with hostility. From these results, avenues for future research on anti- and prosocial outcomes from group gaming are offered.


The differences between competitive and cooperative play is how players achieve a desired goal. That is whether a goal is interdependent with other people’s goals or not. For example in cooperative play, “I need to kill a dragon, but I can’t do it alone. So having others who share the same goal would help”. Whereas in competitive play, “I want to kill that dragon myself, but there are others who wants to. I’d better be the one who did the deed.”Therefore, between cooperative and competitive play different strategies will be adopted.

Researchers have found that competition is related to aggression. If you think of hockey, fights break out often for the slightest offences. Or the case of the headbutt of Zinedine Zidane. So, the literature says that competition is associated with frustration from being unable to achieve a goal, anger, and arguments. On the other hand, cooperative is associated with feelings of cohesion, camaraderie and agreement. Therefore, there are differences in experiences between cooperative (coop) and competitive play and this can translate into other situations as well.

Group size within MMOs can range from 2 players to thousands in a single guild. As group size increase, so does the feelings of competitiveness and disagreements. In addition, subgroups (i.e. cliques) would form that would compete with other subgroups. Decreased perceived individual achievement and competitiveness and increased perceive anonymity within a group are factors of this in-group competitiveness.

Personally, I think individuals would compete depending on the balance between their individual identity and group identity. So, soloists vs. groupists or selfishes vs. altruistics. There might be cultural differences between individualistic, like the U.S.A. and collectivistic, like China individuals. In theory, collectivisitic guilds would be able to hold a larger number of players before antagonism between members sets in, since they emphasize group identity before individuals. But I’ll need to look into it.

Although, a player’s interactions with other players is generally smaller, the type of interactions can range from strangers to intimate friendships. Generally, strangers group together for a single and one-time purpose, such as completing a quest. While more stable groupings, like guilds, have more long-term goals and purposes, such as social networking or fun between friends. In any case, group sizes within virtual worlds are pretty much like those in real life.

In-game behaviours are more likely to prime the relevant thoughts. For example, doing more in-game aggressive acts like shooting are more likely to elicit more aggressive thoughts. Or merely the decision to agress is partly responsible for this cognitive priming. Eastin will investigate state hostility, wish I knew what the definition is.

Verbal aggression, such as insulting another, is pretty much associated to aggressive behaviour. However, verbal aggression within computer-mediated communication isn’t examined much. Although, I could say that flame wars are very much aggressive. Swearing as part of a normal conversation is something I’d like to investigate, but I digress.

So his hypotheses from the text above gives us four hypotheses:

1- Competitive play will elevate state hostility than coop play
2- Larger groups will elevate state hostility than smaller groups
3- Number of kills is positively related to state hostility
4- Verbal aggression is postively related to state hostility


Participants: 162 undergradute students (48% male, 52% female) from Ohio State Uni. Age (M=21.50, SD=0.8) nothing interesting.

Procedure: The game used is Unreal Tournament (unknown edition or is it really the first?) Participants are randomly assigned into a group size (2, 4, or 6 players) and play mode (coop or competitive mode, i.e. deathmatch). In coop, player groups are to fight another player group of the same size, while competitive play is one player fight other players, either against 1, 3, or 5 others. The goal is to score the highest, your score in competitive and group score in coop, obviously.

All participants play at different physical locations in campus, not seeing other players and are using Teamspeak to measure verbal aggression. In coop, they can communicate and listen with only teammembers. In competitive, they can communicate with everyone. All participants are given 10 minutes training, then the testing session last 20 minutes, consistent with previous research.

Questionnaires are given before (demographics, game experience, trait aggression) training and after testing session.


State Hostility: the Anderson, Deuser, Deneve state hostility scale is a 35 questions on a 5-point likert-scale. Basically, asking questions about their mood, like “I am angry”

In-game behaviour: kill scores (M=21.2, SD=15.67, range 0-75)

Verbal aggression: I quote: “any utterance having hostile connotation and, more specifically, expressions of negative affect, verbally abusive utterances including character and competence attacks, negative comparisons, profanity, or references to destruction or physical harm” Then it’s scored in a ratio, so the number of verbal aggression divided by the total number of verbal acts.

Trait aggression: The Buss-Perry aggression questionnaire, 29 questions on a 7-point scale on what characterisitcs does a participants agree with.

Perceived arousal: 24 words to rate on a 5-point scale, from very unlikely to very likely. Basically, whether the mood you felt is related to the word.

Video game experience: The new media experience questionnaire which ask participants of their lifetime history (played since…), monthly history (many times per month) and daily history (hours per day).

Gender is also included in the analysis.


They analyzed the data using a mixed multilevel analysis model. There’s some complicated details, but I think an expert should say something here about controlling factors and fixed variables.

It seems that trait aggression is significantly positively associated with post-game state hostility. Game experience, perceived arousal and gender are not associated.

There were close statistical significance differences between cooperative and competitive play, in that competitive players scored higher on state hostility. But not significant enough. More studies!

There are significant differences between group sizes, in that 6-players groups scored higher on state hostility than 2 and 4-players groups. The 2 and 4 players groups don’t differ each other. So this study pretty much consistent with earlier research on group size and aggression, nothing special. Although, Eastin wrote that anonymity is something that needs to be looked closely at since it can potentially affect competitiveness, especially on the internet IMO. And let’s not forget that when you hear a girl’s voice, you’d react differently.

The quality of the group can be a potential factor in aggression, since the study does not assess whether the participants know each other. For example, there are different interaction behaviours between clan members and strangers.

In-game killing was not associated to state hostility scores, while verbal aggression do have a positive relationship. What is a bivariate relationship? Eastin wrote there is a significant bivariate relationship between in-game killing and hostility, same goes for verbal aggression.

This is in contradiction to the previous research on the relationship between Violent video games and aggression. Eastin argued that as games are more complex with Teamspeak, team play, killing method and coop vs. deathmatch, the effect might become attenuated. It’s logical, but from a cultural-historical point of view, I think it has more with cultural changes like how I really see the 80’s as stereotypically violent and unsophisticated.

Update (05/03/08): I was showering when I thought about in-game kill scores and state hostility. Kill scores represent successful aggressive acts, what about unsuccessful aggressive acts? They’re still aggressive acts and If the GAM (citation needed) says that the more aggressive acts you do, the more you become hostile. So I think Eastin’s result may have a possible confound. And I e-mailed him this question to find out.

There are several caveats in how verbal aggression is measured. The target and intent of the aggression. Team member performances might be influential on players’ state hostility and verbal aggression. Like some players saying “you suck” on newbie team members. Or players would say “pwned!” or “suck it” to their recently deceased opponent and that opponent heard what they said , thereby having an effect on both the insulter and the insultee. IMO, this might explain the coop and competitive group non-significant results. Another consideration is to assess one’s multiplayer experience, instead of general game experience, some like me might only be playing solo, while others play with others.

Well that’s it, and I pretty much said what Eastin wrote in his paper with what little opinion I have.

Eastin, M. S. (2007). The influence of competitive and cooperative group game play on state hostility. Human Communication Research, 33 (4), 450-466.


7 thoughts on “Hostility differences between cooperative versus competitive play in video games (Eastin, 2007)

  1. @ JR

    It’s mostly likely that I got that through my university library servers. And that was by luck too, since I was looking for Dmitri Williams’ latest article and Eastin’s article was there beside it.

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

  3. Pingback: Addictive Group Play Might Make Johnny an Angry Boy « Educational Games Research

  4. Pingback: Video Game Violence and the Female Game Player: Self- and Opponent Gender Effects on Presence and Aggressive Thoughts (Eastin, 2006) « VG Researcher - Psychology

  5. Pingback: Aggressive players have aggressive personality (Peng et al., 2008) « VG Researcher - Psychology

  6. 1. In reference to your point (see below):

    “In-game killing was not associated to state hostility scores, while verbal aggression do have a positive relationship. What is a bivariate relationship? Eastin wrote there is a significant bivariate relationship between in-game killing and hostility, same goes for verbal aggression.”

    A bivariate correlation is basically a correlation between two variables. In this case, it is showing the relationship between how many in-game kills and hostility (and also verbal aggression and in-game killing).

    2. References for the GAM:

    (Anderson & Bushman, 2002b; Anderson & Dill, 2000; Anderson & Huesmann, 2003).

    The GAM has more recently been updated by Buckley and Anderson (2006) to the General Learning Model (GLM).

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 )

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