Monday, April 21, 2014

The Confidence Gap and Possible Effects on Persistence and Pay

An article in The Atlantic by Kathy Kay and Claire Shipman is well worth reading and pondering.  On average, women are less confident than men, with harmful consequences for equitable advancement based on ability.  "Confidence, " says psychologist Richard Pette, "is the stuff that turns thoughts into action."  A person with low confidence tends to try less hard, to give up more easily, to negotiate less successfully, and to face fewer challenges that lead to growth.  Could it be that men's overconfidence is putting women at a disadvantage?

 Numerous surveys show the same thing: women as a whole are less confident than men.  In my own university's survey of its students, the gender difference on self-assessed confidence is one of the largest and most robust signals  in the data.  The effect has been noted in studies of math performance and as a reason why fewer women than men run for political office.  To be sure, there are underconfident men and overconfident women.  But the balance is tipped, and female self-confidence goes against social norms, as the mixed response to Sheryl Sandberg's messages makes clear.

This is not merely about social norms and politics: it might affect your salary.  At nearly all elite universities, women full professors earn less than men do, on average.  The data are available from the AAUP Annual Report on the Economic Status of the Profession.  There are notable exceptions.  What causes the gap, and explains the exceptions?

In one sense, the cause of the gaps is clear: individual women are valued less than individual men. 
The real question is why that is so.  Is it because the women are less accomplished?

At MIT, we have studied this question since the gender inequities between salary, lab space, and leadership roles of faculty were first highlighted in the late 1990s.  I am confident (and so are the women) that our female faculty are not less accomplished than our men (as measured, for example, by the percentages receiving significant prizes).  In the 1990s their salaries were less - at least in part because they did not negotiate as hard as the men.  These differences were brought to the attention of department heads and deans who set salaries.  As a result of improved institutional review, our women full professors now earn about the same as our men (a little more, actually).

I do not know whether our female faculty are more accomplished than those of our peers, but I do know they earn more relative to men at their own institution.  And while I cannot prove that confidence plays a role, I do know that examining the facts carefully made a difference in salaries.

Salary equity, and perhaps equity in other matters, is not achieved by "fixing the women".  It is achieved by understanding the gender differences in confidence and ensuring that promotions, salaries, professional opportunities, and rewards are given not on the basis of confidence, but on ability.

So the next time you see a hand shoot up in your classroom, check the gender of the questioner.  If women are being called on less than men, consider how you might encourage the men to give a little more space to others.

1 comment :

Anonymous said...

These articles give important counter-arguments to the confidence article published by The Atlantic:

http://www.policymic.com/articles/87879/it-s-not-the-confidence-gap-here-s-what-s-really-holding-women-back

and

http://www.theguardian.com/commentisfree/2014/apr/23/female-confidence-gap-katty-kay-claire-shipman

Personally, I think it's a combination of all of the above. Certainly lack of confidence plays a role, but improving women's confidence will only take us so far. As the author of this post mentions, women tend to be less valued. I agree that looking at the facts can help to remedy inequalities (such as pay), but convincing people to see the inequalities in the first place (or, better yet, removing the cause!) can be very difficult. I'm glad MIT has seen the light, but many places have not.