Trump's strongest soldiers

White non-college married couples

I sometimes look askance at charts like this one.

Part of this is my longstanding distaste for the gender gap as a frame of analysis. Gender politics can be pretty zero sum, and you can’t really isolate whether or not they’re a function of men shifting right or women shifting left. If you’re a candidate, dialing up female-coded rhetoric is likely to turn men off, and vice versa.

The second reason is that by combining two or more demographic subgroups together, you can almost always show big differences between groups.

But lately, I’ve been studying gender differences more, because of what I suspect are errors of attribution in the discourse. Discussions of the youth gender gap get quite a lot of buzz, for example, but actually, most of it is due to the wide gender gap among single voters — and young voters are overwhelmingly likely to be single. To call it a youth gender gap isn’t strictly accurate. It’s really a gender gap among singles that persists with voters as they age into their 30s and 40s, and for a few, further beyond. Married voters show no such gap. The “youth gender gap” isn’t real among the minority of young voters who are married.

The NBC News chart technically takes into account three different demographic characteristics: race (subsetting to whites), education, and gender. But marital status and age also play a big role in shaping partisanship. What if we could combine all of these relevant variables into one visualization, using a much larger dataset, the 2024 AP VoteCast exit poll, with interviews of 121,000 voters.

What I’ve put together is a treemap of the American electorate, showing presidential vote choice among all specific combinations of voters.

And what see here is that “white non-college men” and “white college women” are far from the only cohesive voting blocs you can look at. “White college women” range all the way from White Married College Women 50+ at Harris +6 to White Single College Women Under 50 at Harris +43. They both lean Democratic, but there are substantial differences within this group, mostly based on whether they are married or not.

What stands here as a much more cohesive voting bloc on the right is a set of groups ironically at the top left: White Non-College Married Voters, representing 20% of the electorate, with each component group backing Trump by margins of around 40 points. Here, both genders are basically unified: older married men were Trump +49 and older married women were Trump +38. There is no age divide among these voters, with younger married voters in this demo as Trumpy as older ones. The internal unity of this group makes it a voting bloc worth studying.

In fact, there’s no single bloc of voters on the left that can form a counterweight to the White Non-College Married voter. The linear opposite, nonwhite college singles, would be quite Democratic, but a very small group. And realignment has made it so that you can’t form a logically cohesive counterweight from all nonwhite voters. Black voters remain the only solidly Democratic group in here, but after that, you’d have to pick out very specific combinations of nonwhite voters (mostly single women).

The only thing that even comes close to being an equivalent bastion for Democrats are unmarried women. According to AP VoteCast, they were 15% of the electorate and were Harris +33. That’s not quite the size of pro-Trump White Non-College Married voting bloc, nor were they as strong in their support for Harris, but it’s just as logically cohesive as a voting bloc as White College Women, if not more so.

At some levels, these are arbitrary groupings. Why do they matter?

Because think about all the ink spilled about about single women or about young women versus young men. These groups aren’t very big compared to this old-fashioned group that alone supplied Trump with an 8 point popular vote cushion. The marital divide in our politics is understudied, this group especially so.

Now, let’s look further at Harris’s strongest soldiers — single women. No other gender/marital combination was stronger in their support for either candidate.

But it’s also a pretty cohesive group along racial and education lines, with the exception of White Non-College Single Women. White College Single Women are more Democratic than their Asian and Latina counterparts, and with the exception of “Other,” are second only their Black counterparts in their Democratic loyalty.

The larger story here is how much the gender divide varies by whether one is married or not, and how that should shape how we think about the political coalitions. It makes sense to think of married couples as voting blocs situated within demographic groups — they were only 5 points apart in their 2024 voting patterns overall. But single men and single women are on different planets politically. Dividing the electorate by gender and marital status produces much more illuminating picture, with gaps between groups larger than almost all others.