Back in 2011 Matthew Yglesias observed that DW-Nominate scores showed Democratic women were more liberal than Democratic men, but then updated the post with a link to this 2009 paper from Brian Frederick showing that the effect went away when you controlled for a district’s partisan profile.
We were curious if that had changed at all five years later, so I mapped male and female House Democrats’ progressive scores compared to their districts’ partisan profiles.
It’s not a major difference, but just in case you needed any more reasons to elect more Democratic women to Congress — however small — here it is:
At least if you rate the importance of a progressive score the way we do — where current congress and crucial votes are weighted more, etc. — the Democratic congresswomen above are on average consistently more progressive than men. The distinction is fairly small, as it’s a difference of approximately two points, but it’s still noticeable.
Democratic congresswomen also technically vote more progressive than men in R+ districts as well — when they get elected there. But since there are only three Democratic women in the House currently representing districts that lean Republican [1, 2, 3], we don’t have enough data to make an accurate determination.
If you extrapolate this graph, the takeaway is that simply electing more Democratic women to Congress will automatically shift it to the left. Only 31% of Democratic elected officials in the House are women, so the return on a political investment in a campaign to equalize representation within the Democratic Party would be high. Embarrassingly, women hold only 19% of all seats, but make up 51% of the country.
This is an opportunity to make Congress more progressive and more representative of the American people. Let’s get to it.
UPDATE: For you statistics people, the p-value of Democrats’ progressive scores based on gender (also adjusting for PVI) is 0.08. In non-statistical terms, this means that after accounting for PVI, there is a presumption that women are more progressive than men, but it’s not very significant — much like we said in our above analysis.
However, if you look solely at gender and ignore the partisan lean of their district, the p-value is a low 0.007, meaning there is a very strong presumption against the fact that Democratic women and men vote the same — where the men vote, on average, approximately 4.8% less progressive than women do. Though this is likely a symptom of more liberal districts voting for female candidates, rather than proof that women in fact vote more progressive.
Including gender along with PVI also better predicts a House Democrat’s progressive score, where PVI alone has an adjusted R-squared of 0.3972 — but gender and PVI together have an adjusted R-squared of 0.4033. (And for you non-stats people, higher in this case is better — but as you can see, it’s not much higher.)