Affirmative Action for Women Works

The strongest criticism against affirmative action is that in trying to achieve diversity in the office or at school, fairness is sacrificed.

The fear is that a white male with excellent credentials will be passed over and a job, promotion, or college acceptance will get handed to someone with inferior skills.

And on the surface, this critique makes sense: if these women or non-white men were just as good -- or better -- than their Mitt Romney-esque counterparts, wouldn't they already be well-represented?

Not only may that thinking be mostly wrongheaded, but it could be the case that when affirmative action isn't used, a candidate pool suffers -- at least when looking at women.

The reason seems to be that women are less likely than men to be competitive and could keep themselves out of the applicant pool, according to very interesting new research by Muriel Niederle of Stanford, Carmit Segal of the Pompeu Fabra University in Barcelona, and Lise Vesterlund of the University of Pittsburgh.

The trio set up an experiment where equal numbers of men and women were paid for accomplishing different tasks. All of the tasks required the participants to solve simple math problems, but they differed in how the subjects were rewarded.

The first task paid off $0.50 for each correct answer.

The second task was a tournament in which the top two scorers got $1.50 for each correct answer, the rest nothing. This task was supposed to represent what happens when someone applies for a job or school.

The third task then asked the subjects to choose which one of the first two payment schemes they wanted to be paid in. Men proved to be more competitive and chose the tournament style 74 percent of the time while women chose it only 31 percent of the time.

But this result wasn't being driven by the fact that men were stronger performers (although on average they were), but by overconfidence. Men who, based on their previous scores, didn't have a good chance of winning the tournament still chose to enter. High-scoring women, on the other hand, were 36 percent less likely to choose tournament-style than men who got similar scores.

Now enter affirmative action: for the fourth task, the participants were told that they could choose to be paid either by the number of problems they got right or by entering into an affirmative action tournament where one of the winners was mandated to be a woman. This type of tournament had the effect of increasing the chance that a woman would win and lowering the chance that a man would win.

Under this scenario, 83 percent of the women played the tournament while only 45 percent of the men entered. Interestingly, now neither men nor women were being overconfident when choosing to play in the tournament.

Why the change in behavior? For women, it could be that they were either more eager, or likely more comfortable, competing against other women. For men, it could be that affirmative action reduced the excitement of competing against other men or reduced the pressure to.

But the main question still remains unanswered: Was there any differences in the quality of winners between the regular and affirmative action tournaments?

The results, while mixed, show that the differences were small.

While the number of high-performing entrants was the same in the two different tournaments, the proportion of women among high-performers was much greater under affirmative action.

Affirmative action would still be costly, however, if the minimum requirements were very high, the researchers found. The red line in this graph shows the number of higher quality men in the affirmative action tournament that would have to be passed in order for at least one woman to win. For example, if the minimum requirement was 13 questions answered correctly, then no men would have to be passed over. But if the minimum requirement was 19 questions correct, then three men would have to be passed over.

The black line shows what would happen if affirmative action is enforced after a tournament takes place. Because of the different affinities towards competition when affirmative action wasn't involved, it would be much costlier to apply it after-the-fact.

(Incidentally, this scenario is similar to the University of MIchigan affirmative action program which was eventually taken to the Supreme Court in 2003. Applicants were not aware that the university used affirmative action.)

The upshot is that affirmative action is probably much less expensive than previously believed, and in some circumstances, not costly at all.

So, why are women less likely to compete in the first place? Could it be genetics?

That's the question being tackled by University of Chicago economist Uri Gneezy in research I wrote about here. The short answer is that it's likely cultural differences that are driving the different the levels of competition -- and not genetics. Read the post for a more in depth explanation.

Add Your Comment

Report Abuse

Archive

Categories

Links