cwbeijer@gmail.com / About / Archive / Other media
Sexism and the 2020 primaries: some early polling - 3/30/19
Democratic primary analysis in 2016 was often dominated by the notion that sexism was inclining men to vote for Bernie Sanders rather than Hillary Clinton. This, I noted at the time, significantly misunderstood the challenges that face women who want to run: while all kinds of institutional and systematic sexism stands in their way, Democratic voters, on the contrary, tend to prefer women.

This narrative, Kevin Drum notes, is emerging once again. But this time around, we have an advantage: there are so many men and women running that we can look for patterns, rather than trying to tease out some underlying trend from just Sanders and Clinton. With that in mind, I decided to look through the polling to see if I could detect any differences in who men and women are supporting. All of this is drawn from YouGov / Economist's latest favorability ratings:


Note that I calculated two measures based on the favorability scores. The first, Net Preference, combines "Very favorable" and "Somewhat favorable" poll responses into a general score, and then subtracts from that "Very unfavorable" and "Somewhat unfavorable" responses. The second, measure, the Likert Score, multiplies "Very favorable" responses by 4, "Somewhat favorable" by 3, "Somewhat unfavorable" by 2, and "Very unfavorable" by 1 - and then adds these numbers together. This method, as I've noted elsewhere, has its problems, but it is often used by social scientists to measure not just general approval but enthusiasm.

So what do these numbers tell us? Two different stories! If you look at Net Preference, it appears that women candidates are being penalized by men - their score, on average, is about 5 percentage points lower. But if you look at Likert Scores, the trend reverses: men are more enthusiastic about women running for office than women are, or than anyone is about the men.

My read of these numbers is that they are giving us the sort of mixed message that you only really see when the trends are ambivalent to nonexistent. This is in part simply because most of the candidates still have very low name recognition and are returning polling numbers that aren't very reliable. But this also reflects the fairly intuitive fact that Democratic primary voters generally support putting more women in office. It remains to be seen, of course, whether the early enthusiasm we've seen from men for the women in the race translates into greater net support as everyone's name recognition improves.