A lot of attention was given to the fact that Trump did better with Democratic leaning groups especially Hispanics, while Biden did better with Republican leaning groups like married men and evangelicals. I think that we should look at this beyond the demographic level and look at it this way: the blue got redder and vice versa.

You could see this same effect at the county and even state level. Five out of seven electoral units that saw two party vote swings towards Trump (California, Hawaii, Illinois, New York and Washington DC) had higher than average vote shares for Biden and thus Hillary Clinton (and all significantly higher than average at that), with the other two being considered swing states (Nevada and Florida). West Virginia and Wyoming, the two most Republican states in both 2016 and 2020, moved 5% and 7% towards Biden, more than twice the national shift.

Here is an image that shows the correlation between swing and 2020 vote share (it was -43%):

enter image description here

What caused the apparent depolarization in the 2020 presidential election? Was it higher turnout bringing out voters that vote in unexpected ways? People switching sides? Some combination or something else?

  • google.com/amp/s/www.nytimes.com/2021/06/30/us/politics/… -- here is an article which talks about the demographics. Jul 2, 2021 at 23:12
  • 2
    Keep in mind that US election campaigning is not so much about convincing undecided voters (like in most democratic countries) but mostly about mobilizing the own supporters. A swing does not necessarily mean that one party is now less popular than the other. It just means that one party was more successful at getting "their" people to vote than the other. When Hispanics voted more Republican, then that does not mean that their leaning changed. It means the Republicans had more luck mobilizing Republican Hispanics than the Democrats had at mobilizing Democratic Hispanics.
    – Philipp
    Jul 12, 2021 at 13:40
  • 3
    Was it really depolarizing? I guess if you look at vote-percent-by-race metrics, but what about vote-percentage-by-party? How many Republicans voted for Biden, how many Democrats for Trump? I don't know the answer here but it seems like it'd be a more useful dichotomy to examine than race.
    – Ryan_L
    Jul 13, 2021 at 21:36
  • When I look at the chart it looks like something very close to random noise and like the value of the r^2 or p value or Chi-square would all indicate a statistically close to insignificant relationship. Did you evaluate the correlation by any of those measures?
    – ohwilleke
    Jul 15, 2021 at 0:47

2 Answers 2


The 2020 Election Was Not Depolarizing

I disagree with your premise. The 2020 election was not a depolarizing election.

Both 2016 and 2020 were incredibly polarized elections relative to prior years and any distinction between 2016 and 2020 is more or less trivial and misses the forest for the trees.

While there were slight tweaks to the outcomes in some swing states that changed the result, the overall shifts were modest in magnitude everyplace that it mattered. The evidence supporting this conclusion includes the following points.

Coalition Changes Can Inaccurately Look Like Depolarization

Also, redrawing coalitions does not imply depolarization. Democrats have in recent elections shed working class voters while picking up college educated voters from Republicans. This has certainly happened.

enter image description here

More Demographically Polarized Voters

But that doesn't make partisan divides any less stark. The electorate is more divided by ideology, by race, by gender, by age, by geography (especially at a county or smaller level), and by religion than ever (see 2020 exit polling from multiple sources summarized here).

Consider, for example, the extent to which the nation has become polarized politically along economic lines:

enter image description here

More Polarized Voter Ideologies

In general, we have gradually shifted over the last decade or two from a political system in which voters are ideologically situated on a bell curve of ideology, with elected officials being more bimodal with moderate left and moderate right peaks, to one in which the distribution of voters by ideology increasingly mirrors that ideological distribution of federal elected officials in the U.S.

enter image description here

More Polarized Members Of Congress

For example, both the House and Senate have no ideological overlap between the parties (something that was not true until recently).

enter image description here

Record Low District Splitting

Similarly, the total number of split districts in 2020 was 16 (3.7% of Congressional Districts). This is the lowest level of district-splitting, both in percentage and absolute terms, since 1920, a full century earlier. That year, 11 out of 344 districts produced a split result (3.2% of Congressional Districts). In 2016 and 2012, 35 and 26 districts (respectively) split their tickets.


You have missed a very important point here, which is that both the 2016 and 2020 Presidential elections were less about politics than about the personalities of the candidates.

In 2016, you had many voters who were unwilling to vote for a woman, and one without much in the way of credentials or personal appeal. Against her, you have someone with a familiar name, who had spent his life creating a false impression of business competence.

Four years later, that person had demonstrated his actual lack of competence, most prominently in mishandling the COVID pandemic, and had other aspects of his character exposed. Running against him you had someone with a long record of moderate competence, and a reasonably likeable personality.

Those differences, not politics, were what tipped the elections. They were personal rejections of the candidates, more than of their politics. This is demonstrated by the results of the House and Senate elections, where political results were much closer.

  • 6
    There are a lot of statements here that could do with some backing. Clintons lack of "personal appeal", Trump's "false impression of business competence", that Trump's lack of competence and mishandling of the pandemic were a driving force and that Biden was "reasonably likeable". I'm not saying I disagree with these things, only that just because we see it that way doesn't mean it can be stated as fact without any backing. Jul 4, 2021 at 9:09
  • 11
    Wasn't Hillary Clinton the US Secretary of State? How many more offices does the US have that provide better presidential credentials?
    – Jontia
    Jul 4, 2021 at 17:49
  • @Lio Elbammal: So how, other than having experienced the last few years, would you suggest those statements be backed?
    – jamesqf
    Jul 5, 2021 at 3:09
  • 7
    @james Other people experienced the past few years differently to you and, given you weren't the only one voting, their opinion matters too. Support for the statements would probably come from polls or public reactions. Jul 5, 2021 at 6:24
  • "In 2016, you had many voters who were unwilling to vote for a woman". Of course but is it the only reason and is it the main reason we can find in polls ? The wording is awkward because it leaves the feeling that it is the main reason.
    – Genorme
    Jul 13, 2021 at 6:51

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .