All of this writing and data analysis is a lot of work! So after more than five years of posting, I've finally launched a Patreon to help pay the bills.

12/13/19

No significant evidence of sexism in candidate favorability scores

Research has repeatedly affirmed that systematic and institutional sexism, rather than widespread personal bias among voters, are the primary obstacles to women in politics. Nevertheless, Amanda Arnold, writing for the Cut, asks: Why Do People Really Dislike the Women Running for President?

To find out, the Cut had YouGov ask people whether "they like or dislike four female candidates...and why" - but there are some serious problems with this approach. For one thing, asking people "why" is only going to get you so far in trying to understand personal motives, since this relies entirely on self-reporting. But more to the point: like-or-dislike numbers about women just aren't that useful if we can't compare them to the same numbers for men. And evidently, the Cut didn't bother to ask.

Fortunately, The Economist did. "Favorability" ratings are often used as a proxy for likability, and every few weeks The Economist and YouGov collects favorability ratings for all of the candidates running for the Democratic nomination. Here's how they look:

The main thing to notice here is that there is no significant relationship between a candidate's gender and their net favorability rating. Elizabeth Warren is viewed the most favorably, and Michael Bloomberg the least. Women do have a slightly lower net favorability average then men, but it's well within the margin of error: -14 for women, -12.75 for men. And if we exclude the two clear outliers in this set (Bloomberg and Williamson), that pattern reverses: in that case, men average -11.27 favorability, versus -11 for women.

What happens if we break these numbers down by gender?

If we simply look at the women in this race, it's tempting to conclude that sexism is indeed at work, since men view these candidates way more unfavorably than women do. But look at the scores for male candidates, and it's clear why that is: men view all Democratic candidates less favorably than women do. And in fact, on average, men give male candidates a rating of -19.5, and female candidates a slightly better rating of -19.25.

That, trend, too, is well within the margin of error. The trend among female respondents is bigger, but not by much: they give male candidates an average rating of -6.5, and female candidates an average rating of -9.5.

Looking at these numbers, I don't really see anything statistically significant going on. No strong evidence of misogyny at work in candidate favorability ratings, even when you break them down by respondent gender. If one really wants to understand sexism in electoral politics, one might ask questions like "how come so few women ran for president in the first place" - but that's the sort of question that shifts scrutiny away from voters and onto party elites, so I doubt we'll see many articles about it.