Psychology

Snot fair!

Exactly when is something perceived as "not fair"?

AS THE bankster phenomenon has so eloquently illustrated, Homo sapiens is exquisitely sensitive to injustice. Many people grudgingly tolerated the astronomical incomes of financial traders, and even the cosmological ones of banks' chief executives, when they thought those salaries were earned by honest labour. Now, so many examples to the contrary have emerged that toleration has vanished.

Surprisingly, however, the psychological underpinnings of a sense of injustice-in particular, what triggers willingness to punish an offender, even at a cost to the punisher-have not been well established. But a recent experiment by Nichola Raihani of University College, London, and Katherine McAuliffe of Harvard, just published in Biology Letters, attempts to disentangle the matter.

Dr Raihani and Ms McAuliffe tested two competing hypotheses. One is that the desire to punish is simple revenge for an offence. The other is that it is related to the offence's consequences-specifically, whether or not the offender is left better off than the victim.

Until recently, the temptation would have been to advertise for undergraduate volunteers for such a project. Instead, Dr Raihani and Ms McAuliffe decided to follow a new fashion in psychology and recruit their human guinea pigs through a system called Mechanical Turk. This arrangement, run by Amazon, a large internet firm, pays people registered with it (known as Turkers) small sums of money to do jobs for others. That allowed the two researchers not only to gather many more volunteers (560) than would have been possible from the average student body, but also to spread the profile of those volunteers beyond the halls of academe and beyond the age of 21.

Dr Raihani and Ms McAuliffe asked their Turkers to play a game. In it, the volunteers were paired and given small sums of money. One member of a pair could then take a predefined sum from the other, or not, as he chose. After that the other could, at a certain cost to himself, impoverish his opponent to a greater degree.

The first player might receive ten cents, 30 cents or 70 cents. The second player always received 70 cents. The first player was then allowed to take 20 cents of the second player's money. Finally, the second player could reduce the first player's total sum by 30 cents, but at a cost of ten cents to himself-in other words, he lost money too by doing so.

The crucial point of the game is that in all three cases the second player suffers the same absolute loss if the first chooses to take money from him. The offence, in other words, is identical. But in the first version of the game he remains ahead if he does not retaliate (50 cents v 30 cents), in the second he comes out equal (50 cents v 50 cents), and in the third he ends up behind (50 cents v 90 cents).

The upshot was that in the first two cases about 15% of second players chose to retaliate if they had money taken. This was more or less the same as the number in all three versions of the game who "retaliated" even though they did not have money taken (a course of action allowed by the rules). In the third version, though, more than 40% of second players retaliated when money was taken from them-even though the outcome was still that the first player ended up ahead, with 60 cents to the second player's 40 cents.

On the face of things, this result suggests that what really gets people's goat is not so much having money taken, but having it taken in a way that makes the taker better off than the victim. That will clearly bear further investigation, for example by looking at the case where the first player begins the game better off than the second. It is intriguing, though, that even such trivial sums of money can provoke thoughts of revenge. In light of this, the fate awaiting those astronomically paid bankers could be a particularly nasty one.