Thursday, February 09, 2006

Blame dame games

A recent study has been reported by the National Bureau of Economic Research, summarized under the title Women Shy Away from Competition:
The payment for the first task was awarded on a non-competitive basis by paying a piece rate of 50 cents for each correct answer. Payment for the second task was a competitive winner-take-all "tournament." Losers received nothing and the person in each group with the largest number of correct answers was awarded $2 per correct answer. For the third task, participants chose either piecework payment or the tournament compensation.

Men and women answered the same number of problems correctly under both compensation systems. But when allowed to choose compensation rates for the third task, 75 percent of the men chose tournament compensation while only 35 percent of the women did so...

In this experiment, large gender differences in tournament entry can be observed, even in a case where women are as good as men, where discrimination is absent, and where the time spent on each task is limited...

This summary puts attention on the women. But women performed rationally! If all participated in the tournament, a 25% preference for competing would be rational, but since not everyone may participate in the tournament, the rational likelihood is higher. 35% sounds pretty close to optimal (since fewer than 4 will participate on average, but probably at least 2 will, the optimal likelihood is somewhere between 25-50%, probably around 30-45%).

Men's 75% likelihood of choosing the tournament, on the other hand, shows a preference for competition that clearly does not optimize cash rewards. The study's summary buries the lede:

About 75 percent of the men believe that they won the second-task tournament of four participants.
That's remarkable: it's not just that men prefer to compete, but they are deluded as to the quality of their performance and their odds of winning, three times so. Isn't that the real story?