Gender and incumbency, or: girl-on-girl political action

An interesting pattern I noticed in the 2006 midterm elections is that women seem to run disproportionately against other women. More precisely, it seems that female incumbents face female challengers more often than male incumbents face female challengers.

Being the intrepid social scientist that I am, I dug up this paper on the subject, which confirms my suspicions: they report that women are both more likely to seek the opposition nomination and more likely to be the opposition candidate when the incumbent is female as opposed to when the incumbent is male.

What the paper does not do a very good job at is figuring out why this happens, although they do offer several hypotheses. One possibility is that this is a deliberate strategic decision on the part of parties, to try to neutralize a perceived gender advantage on the part of the incumbent. Another is that it might be a strategic decision on the part of the individual candidate: if a woman has won office, women should be more inclined to run for office. And, most simply, it may just be that some congressional districts will randomly have a high concentration of female politicians, and that those districts will clearly both be won more often by women and contested more often by women.

Shooting from the hip, I would say that the first hypothesis (which was my initial thought when I observed this anecdotally) is the least convincing: if we think that men are at a real or perceived disadvantage facing women in a general election, one would imagine we would see a lot more women in Congress. Occam's Razor inclines me towards the last hypothesis, but I believe getting the true story from the statistics will require a little "persuasion" with an instrument or two. I'll keep you posted on my progress.

Edit: Link now goes to paper.