Hormonal responses differences when playing against an ingroup or outgroup (Oxford et al., 2009)

Via gamepolitics, this study examined FPS players’ hormonal responses towards an in-group versus an out-group in a competitive setting. I can understand the evolution part, but I’m a complete beginner in neurobiology, I can’t understand the entire results section.


For 14 teams of three young men, salivary testosterone and cortisol were assessed twice before and twice after competing in within-group and between-group video games that simulated violent male–male competition. Men who contributed the most to their teams’ between-group victory showed testosterone increases immediately after the competition, but only if this competition was played before the within-group tournament. High-scoring men on losing teams did not show this immediate effect, but they did show a delayed increase in testosterone. In contrast, high-ranking men tended to have lower testosterone and higher cortisol during within-group tournaments. The results are consistent with the hypothesis that men’s competitive testosterone response varies across ingroup and outgroup competitions and is muted during the former. The testosterone response during the between-group competition also suggests that violent multiplayer video games may be appealing to young men because they simulate male–male coalitional competition.

I’ve emailed one of the authors for clarification, if the study has an implied link with aggression.

Here’s his response: 

The testosterone response – challenge response – is commonly associated with male on male fighting for dominance and thus the implied link to behavioral aggression.  We found this challenge response for the between-group but not the within-group matches, suggesting the different social contexts elicit different hormonal responses, one of which is commonly associated with increased propensity for behavioral aggression.  Our point though is that these games are appealing for a reason; they may tap an evolutionarily old system for male-male coalitional competition.

From their literature review, the video game paradigm is just another instrument to test evolutionary predictions (IMO). They referred several studies that tested this hypothesis from animal studies to village dominoes competition in Dominica to a Spanish professional basketball game.

When men are facing challenges, their testosterone levels don’t rise uniformly because it depends on internal and external differences. Individuals’ responses to challenges are related to individual differences in competitive striving. The challenge must have some significant influence to his niche-relevant status. For example, a professional RTS gamer who wins an RTS match would have this testosterone response than would a professional bowler playing the same game. This means, any competition where a person has a status-related stake would have this challenge response.

There is a competitive response in terms of social context, when the challenges come from an in-group (i.e. friends or someone you have a connection with) versus an out-group (i.e. strangers or a rival guild or clan, someone you hate or even the opposite gender). The evolutionary explanation why a challenge response to in-groups is lower than toward out-groups is a matter of cost-benefit trade-off. Males must simultaneously maintain in-group status and social support from the in-group in order to compete against out-groups. Basically, you can be a bully, but if you alienate your social group, you’d be left in a disadvantaged position and must fend by yourself.

The authors briefly mentioned cortisol’s role in competitive situations, but the article is focused on testosterone. So I’ll leave it at that.


Participants: 14 three-member teams of undergraduate men. Average age is 19, they were offered class credit and to make the video game competition relevant, cash rewards are offered based on their performance: winners are given 45$ and losers get 15$. That’s pretty good way for generalization. One might ask if the researchers would have tested actual friends, this would entail scheduling hell. Nevertheless, the authors argued through the minimal group paradigm that it was sufficient to instil ingroup identification. This sounds good to me since they’ve managed to get the predicted results from a relatively weak ingroup.

Video game used: Unreal Tournament 2004. They used this game because it’s easily replicable, high experimental control, and requires some skill on the part of players, namely marksmanship and teamwork.


Teams played one practice session per week for three successive weeks, the practice session last for 2 hours and they played in coop mode. The teams played separately according to weekday and time of day. Therefore, the teams did not have any significant contact with each other. Aside from familiarizing with the game, team members also familiarize with each other, such as sharing demographic information, video game experience, having a team name, and calling each other through their screen name.

The experimental room is arranged like a LAN party room, the opposing teams are separated in that they could hear each other, but not see each other. The team members can see and interact with each other.

On the fourth and fifth weeks were the actual tournaments. The experimental orders were randomized, so a team might either begin playing death match which is the in-group condition or onslaught which is the out-group condition. Then the next week, the team would play the other condition. Play time during those two weeks is 30 minutes.

The in-group condition is plain death match where performance is based on frag scores. So the winner gets 45$ while the other two players get 15$.

The out-group condition used the same map as in death match, except two teams are competing in achieving a goal. The game mode Onslaught is where teams must capture successive power nodes in order to be able to destroy the opposing team’s power core, while preventing the opposing team from doing the same.

During each experimental tournament, four saliva samples are taken. One is taken 30 minutes before the start of the tournament, one immediately before, one immediately after and a final one 30 minutes after.


You can refer to the abstract.

They found their predicted results with a few interesting variations. One variation is the timing of the testosterone levels being sooner than typical because players get immediate feedback and information on their performance. Higher ranking players in relation to other individual players, regardless of being in the winning or losing team, showed an increase in testosterone level, again the immediate feedback and information on their performance. The authors contend that the status within the context of the video game is more relevant to them. My interpretation is akin to compensation in regards to high ranking players of the losing team, they still feel good of their own achievement despite the performance of his team members.

This study is pretty good, although some of its flaws should be resolved in a future study, such as controlling information feedback and perhaps try a playing a competive non-violent video game.

Oxford, J., Ponzi, D., & Geary, D. C. (2009) Hormonal responses differ when playing violent video games against an ingroup or outgroup. Evolution and Human Behavior,


One thought on “Hormonal responses differences when playing against an ingroup or outgroup (Oxford et al., 2009)

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