Can experimental economics shed light upon the Trayvon Martin case? I ask because of a recent paper by Jeffrey Butler and colleagues.

They got people to play a simple dictator game, in which one subject was given some money to divide between himself and a partner. Third parties were invited to watch and to choose to punish the dictator if he made what they thought was an unfair division.

They found that, on average, observers chose to punish unfair offers even when they had to pay for doing so. This is a routine finding; we have a taste for fairness and are willing to pay for that taste.

But here's the thing. Observers were then split into two groups, depending on whether they drew a blue or red chip from a bag. They then found that observers were keener to punish unfair decisions by dicators from the "other" group. Whereas observers were willing to pay 33 cents to fine a dictator €1 if he was in their group and made an unfair offer to a group member, they paid 47 cents if the dictator was in the other group and treated one of the observers' group unfairly. That's a 40% difference.

We are, then, much more willing to punish members of the "out-group" for an offence than we are to punish members of our own group. You can easily imagine an evolutionary reason for this; moral norms evolved to protect the group, especially from predation by outsiders.

This is where the Martin case enters. Blacks and liberals regard Zimmerman (perhaps wrongly, but no matter) as one of the out-group and are therefore keener to punish him than those who regard him as one of "them". There is of course nothing new in this. It's consistent with a vast range of behaviour, from the horrible to the trivial:

- Demagogues who want to commit genocide first take steps to establish their victims as an out-group, and a threatening one at that.

- In the deep south, blacks were lynched for things that wouldn't have been crimes at all had a white person done them.

- The attempt to distinguish between strivers and skivers is a means of legitimating welfare cuts not just because it tries to paint skivers as undeserving, but because it paints them as an out-group, who can be punished harshly.

- One reason why MPs' expenses created such a furore - when such behaviour had been going on for ages - was that they had grown distant from the public, who thus regarded them as outsiders and so were keener to punish them.

All this is consistent with the idea that social capital has a dark side (pdf) - a desire to hurt outsiders. As Charles says: "Social solidarity, tradition and community (all good) can end up being defensive and fearful of 'otherness' (bad)."

But there's second message. Professor Butler created social divisions in the weakest and most trivial way possible - by simply drawing chips. If divisions formed in this way can be sufficiently strong to generate significant differences in willingness to punish, isn't it likely that real-life social and ethnic divisions can be easily deepened by ideology, rhetoric and other cues? And mightn't such divisions have some terrible effects?