Did Men or Women Shift More in This Election?

A common thread in a lot of stories about the 2020 election is that Trump solidified his appeal to men but bled votes among women, both overall and within racial groups. Is this right?

It's strange that this has taken hold since the exit polls mostly contradict this story and that is the data source most stories tend to use. Or rather tend to use when it suits their the pre-existing narrative.

The national exit poll shows:

1. Men shifted Democratic (by margin) slightly more than women, thereby narrowing the gender gap.

2. White men shifted sharply toward Biden (8 points!), while women shifted slightly toward Trump, thereby significantly compressing this gender gap.

3. Black men and women both shifted significantly toward Trump but by equal amounts so this gender gap did not change.

4. Hispanic men and women both shifted significantly toward Trump but men (8 points) more than women (5 points) so this gender gap did, in fact, increase.

5. White noncollege men shifted 6 points toward Biden while white noncollege women did not shift at all so this gender gap was compressed.

6. White college men shifted dramatically toward Biden (11 points!) but white college women by only 2 points, strongly compressing this gender gap.

So how much do you hear about these shifts? Not much, since it doesn't seem to fit with much of what media analysts want to say. That's bad but on the other hand is the story told by the exit polls really true. I've got my doubts.

Consider these data from AP/NORC VoteCast, the best available source for data on 2020 and States of Change, the best available data source for 2016. This comparison yields the following.

1. Men did not shift and women shifted slightly toward Biden therefore slightly widening the gender gap.

2. White men and women both shifted toward Trump by roughly equal amounts thereby keeping the gender gap stable.

3. Black men and women both shifted slightly toward Biden and by roughly equal amounts thereby keeping this gender gap stable.

4. Hispanic men and women both shifted sharply toward Trump but women (12 points) more than men (7 points) so this gender gap decreased significantly.

5. White noncollege men and women both shifted 7 points toward Biden so this gender gap stayed the same.

6. White college men shifted 2 points toward Trump, while white college women shifted 7 points toward Biden, thereby substantiall increasing this gender gap.

So, what to make of all this? Neither data comparison shows much of a change in the overall gender gap--a slight decrease in the exits, a slight increase in the VoteCast/States of Change (SOC) comparison. The VoteCast/SOC results cast doubt on exit poll findings suggesting big gender gap compressions among whites overall and among college and noncollege whites. Similarly, the exit poll finding of an increase in the gender gap among Hispanics is contradicted by the VoteCast/SOC results.

Interestingly, neither comparison suggests an increase in the gender gap among black voters, which directly contradicts a common story in postelection coverage.

In several months, we will have better data on all this and will be able to adjudicate, for example, whether the exit polls are right that the gender gap decreased among white college voters or whether the gap increased among these voters (which I suspect is probably the case, as suggested by the VoteCast/SOC comparison). But for now I would be suspicious of any stories that pronounce with any degree of certainty about gender gap trends in this election.