Jasper Carrott & conditional co-operation

In the unlikely event  of you needing a general theory of human behaviour , one of the most promising candidates is the idea that people are strong reciprocators (pdf): they are inclined to co-operate, but go out of their way to punish defectors.  However, a new paper provides only mixed support for this view.
In a contribution to the tellynomics literature, Donja Darai and Silvia Gratz studied Jasper Carrott’s show Golden Balls. The final round of this offers contestants a classic one-shot prisoners’ dilemma game. If both cooperate, the jackpot is split evenly between them. If one defects and the other cooperates, the defector gets all the cash.  But if both defect, no-one gets it.
They found that - consistent with the conditional co-operators theory - most people (54.5%) choose to co-operate. They also found that people who had contributed most to the jackpot in earlier rounds co-operated less. This too is consistent with the theory - though the quirk here is that contributions to the jackpots are due to luck nor skill; they depend upon the draw of balls in earlier rounds.
They also found that people who thought that they should not have gotten to the final - because they had drawn low-value balls in earlier rounds - co-operated more. This too is consistent with conditional co-operation theory; people repay favours.
However, they also found that people who lied about their cash draws in earlier rounds were not punished by defection in the final. This suggests that people are loath to punish anti-social behaviour, even if it is in their interests to do so.
Darai and Gratz found some other interesting things. One is that the bigger the jackpot, the less likely people are to co-operate:

The cooperation rate decreases with a continuous increase in the jackpot size. But, surprisingly, the rate declines sharply from 73.6% for jackpots below £500 to roughly 50% for jackpots above £500

This is more evidence for the abundance effect: proximity to wealth makes people more anti-social.
They also found that:

The more similar both final players are with respect to their age, gender, race, and place of residence, the more likely player i is to cooperate with her opponent. This finding suggests, that in-group biases may be present.

An exception to this, however, is that two women are less likely to co-operate with each other than are mixed pairs.
One other quirk is that English players are significantly less likely to co-operate than are Welsh, Irish or Scots. I’ll leave others to ponder the meaning of this.

Comments

[Tim Worstall]

"One other quirk is that English players are significantly less likely to co-operate than are Welsh, Irish or Scots."

Aha! Proof perfect of my contention that while the Celts might be naturally socially democratic societies, England isn't.

[Phil]

I would really like to see the 'English' figure disaggregated by region - I'd be willing to bet that most regions of England would fall closer to the "Celtic" norm. My experience of living in the north-west and the south-east of the country suggests that there are much higher concentrations of grasping, resentful, self-righteous sod-you-Jack merchants down there. My personal theory is that London isn't so much a city as a pathology.

[Laban]

One other quirk is that English players are significantly less likely to co-operate than are Welsh, Irish or Scots

Robert Putnam : "the more people of diverse ethnic backgrounds live in a community, the lower the level of trust among the community’s citizens"

[Tom Addison]

Another quirk is that the Scots are shite at football.

[Julian Ware-Lane]

Not all English conform to stereotypes, not even those born and residing in Essex.

I would be interested to see a breakdown for voter preference - surely socialists are more likely to cooperate?

[Yannis]

This study is not very new. Golden Balls was already studied by Richard Thaler, one of the godfathers of behavioral economics: http://ssrn.com/abstract=1592456. Seems that Donja Darai and Silvia Gratz basically redo what was already done in that paper.

[Jack Potts]

"the more people of diverse ethnic backgrounds live in a community, the lower the level of trust among the community’s citizens" I guess thta's just the way of the world now - shame really...