7

On the Census.gov website, there is a feature to see how much of a population has a bachelor's degree or higher in education. I want to see the county-level correlation between how much of the population in a county is college-educated and how much "bluer" the county got. A county getting redder (voting for Trump more in two-party vote share for his re-election) would have a negative score. I am focusing on the swing because I want to see how the US electorate is changing not what it is now.

  • You need to include many other factors. For example, "education" is not a uniform thing. STEM fields, for example, are very likely to be more heavily republican while "soft" sciences (grievance studies) will very much tend towards democrat. Also, you need to correlate many other factors, such as the number of welfare cases in the county, which will clearly correlate democrat. – puppetsock Dec 3 '20 at 1:21
  • Related reading: skeptics.stackexchange.com/questions/49969/… The Question is very different, but both Answers address how voters changed their support without education being considered as a reason for that shift. – computercarguy Dec 10 '20 at 0:03
11

Using data from the American Community Survey, as well as the county-level results from Fox's feed for 2020, and from the MEDSL dataset for 2016, I've investigated this - below are the results. I found that there was a fairly decent correlation between the two variables, with a correlation coefficient of around 0.51. A linear trend-line (red) drawn through the data shows a gradient of around 0.16. On the graph below, a positive change on the y-axis indicates a two-party swing towards the Democrat candidate.

If we instead plot a LOWESS curve (magenta) over the data, we see that the increase only seems to occur up until about 35%, at which point the curve flattens, with a few counties in Virginia (around the Pentagon) dragging it down towards the end of the scale.

enter image description here

The code used in this answer can be found on GitHub.

  • 1
    Frankly with a loess fit that looks more likely to be an increase up to 30% or so, after which it is flat. – Fizz Dec 1 '20 at 17:15
  • @Fizz yeah, you could be right there. The data is provided in a nice format with my code if you want to have a crack, or I'll have another look at some point. – CDJB Dec 1 '20 at 17:27
  • 3
    Please note that washing together all the colleges might lead to Simpson's paradox. There is a difference between a math/engineering diploma, and a liberal arts degree (majoring in gender studies and underwater basketweaving) the latter now being so strongly pro-Democrat that even the suspicion of leaning Republican will lead to students being ostracized. – vsz Dec 2 '20 at 9:29
  • 3
    @vsz: There's also a confounding issue, where looking at distinctions between STEM degrees and "other degrees" means you're also looking at gender and race-based distinctions too; STEM degrees in general, and engineering degrees in particular, are heavily biased towards white males, who are themselves the most conservative-leaning block in the U.S. Including type of degree in the analysis would also pull in all that other baggage; STEM may be more conservative than non-STEM, but a huge part of that is the unbalanced race and gender mix; what they happen to study is secondary. – ShadowRanger Dec 2 '20 at 15:22
  • 6
    @vsz On average, every group becomes more Democrat when they become more educated. The number of liberal arts degrees (your disparagement is not necessary or called for) is not sufficient to drive this effect. Besides, you conspicuously drop the sciences, which are very liberal as well (probably because the GOP constantly spews science-denying nonsense and tries to decrease science funding ...). – Azor Ahai -him- Dec 2 '20 at 18:59
8

The State Level Correlation

enter image description here

(Source)

The question is asked in terms of county level data, but it is worth adding that the correlation at the state level, using the less quantitative comparison of which candidate won, and using rank order rather than percentages for education, which is good for showing trend, is very striking

There is a very strong correlation between the rank order of states from highly educated to less highly educated, and whether they supported Trump or Biden in the 2020 Presidential election.

Out of 50 states and the District of Columbia that voted in the 2020 Presidential election, 47 out of 51 voted for President in accordance with their rank order, and there were two exceptions in each direction.

The Exceptions At The State Level

The only exceptions are that Nevada and Wisconsin supported Biden (by very thin margins) despite being less educated than some states that voted for Trump, and that Kansas and Utah supported Trump despite being more highly educated than many states that supported Biden (and in the case of Utah, despite also being highly urban and having a population with the lowest average age in the United State).

In the case of Nevada, Kansas and Utah, the extent to which those states are more or less religious than states with comparable levels of education does a great deal to explain the difference (with more religious states tending to favor the GOP). Nevada is more secular. Utah (obviously) and Kansas as well, are more religious.

Almost everyone in Nevada also lives in highly urbanized area whose economy has social liberalism (legalized gambling and in some places legalized prostitution) at its core, and Nevada's population is younger than average (and young people tend to favor Democrats). Kansas, in contrast, is more rural than most blue states and also older than average.

Wisconsin isn't quite that easily explained, but was very close electorally (Republicans control the state's government), is still close to the middle in terms of education, and has Democratic leaning neighboring states with whom it shares a great deal culturally in an election that was heavily driven by cultural issues.

States That Flipped

Most of the states that flipped from Trump in 2016 to Biden in 2020 were more educated: AZ, MI, PA and GA.

WI is again an exception, but as noted previously was very close and is still above average in education among states that voted for Trump in 2016.

Exit Poll Analysis

We also know, below the level of the county at the level of the individual voter, what was going on in 2016 from exit polls that confirms that there was a major, education driven shift:

enter image description here

enter image description here

(Source for second chart).

The 2020 exit polls showed more of an across the board shift away from Trump from 2016 to 2020. Men shifted away from Trump and this was particularly true of college educated men.

enter image description here

The exit poll data from some key swing states was more of a mixed bag:

enter image description here

enter image description here

2

Voting for Trump or Biden in this election isn't a measure of Democrat or Republican. This is a completely different election than it was even 4 years ago and likely completely different than 4 years from now. So many people are fed up with Trump that Republicans are voting for Biden, so that makes the premise of this Question, well, questionable.

I understand that there's a lot of concentration on just education being a deciding factor of party affiliation, but there's so much more to it than just that to the point that trying to justify party affiliation on any one thing doesn't make any sense.

I didn't count the number of people on this list, but I'd guess there's at least a couple hundred Republican officials that voted for Biden, or at least worked against Trump during this election cycle. This doesn't mean they are Democrat, it means they broke from their party. That's it.

https://en.wikipedia.org/wiki/List_of_Republicans_who_opposed_the_Donald_Trump_2020_presidential_campaign

There's even a website dedicated to Republicans who are against Trump. Scrolling through the list, there's likely thousands, maybe tens of thousands of people listed as Republicans going against Trump.

“Donald Trump is not representative of the Republican Party that I fell in love with.”
These are Republicans, former Republicans, conservatives, and former Trump voters who can’t support Trump for president this fall.

https://rvat.org/

Cindy McCain, Senator John McCain's widow, states she is endorsing Biden and why. She doesn't state specifically why she's voting against Trump, but you can assume that the reasons she states for voting for Biden are where Trump is failing badly or doing the exact opposite.

As a proud, lifelong Republican I have done my share of campaigning for our party’s candidates. My decision to endorse Democratic presidential nominee Joe Biden was not taken lightly, and I owe it to my fellow Republicans to explain why I’m convinced he is the best choice on this year’s ballot to lead the nation as president of the United States.

https://www.usatoday.com/story/opinion/2020/11/02/mccain-why-republican-votes-biden-column/6113484002/

This article has been revised at least once, since the title was originally "Nearly 350" to now "Nearly 600 Prominent Republicans Voting for Joe Biden". Again, this doesn't mean they changed parties.

With two weeks until Election Day, former Republican National Committee Chairman Michael Steele has endorsed Democratic nominee Joe Biden for president, joining 593 other prominent Republicans who are voting blue come November 3.

These Republicans have broken party lines to speak out against President Donald Trump and publicly endorse Biden. One of the most notable GOP groups is the Lincoln Project, whose ads have severely attacked the president. The anti-Trump group is led by several high-profile Republicans, including John Weaver, Rick Wilson, George Conway, Reed Galen and Jennifer Horn.

https://www.newsweek.com/nearly-350-prominent-republicans-voting-joe-biden-1540611

Even after the election, Republicans are standing up to Trump to say that Biden won the election, even in the face of death threats and other retaliation.

This includes Republican election clerks, secretaries of state such as Georgia’s Brad Raffensperger who has faced death threats, Aaron Van Langevelde of the Michigan state canvassing board who recognized he had no legal discretion not to certify Biden’s 154,000 win in the state, the Republican state legislators in states in Michigan, Pennsylvania, and elsewhere who resisted Trump’s attempted gambit to have states appoint competing slates of electors to create a January battle over counting Electoral College votes in Congress, the Republican Govs. Brian Kemp of Georgia and Doug Ducey, who rejected Trump’s calls to declare their states’ elections fraudulent despite Trump’s personal entreaties and attacks, and the federal and state judges of both parties who applied sound legal principles in rebuffing Trump’s flimsy lawsuits.* And on Tuesday, even Trump’s Attorney General Bill Barr, who had scandalously backed Trump’s claims of a potential for voter fraud in the past, told the AP that the Department of Justice has uncovered no kinds of fraud or irregularities that could change the outcome of the election.

https://slate.com/news-and-politics/2020/12/republican-officials-who-have-gone-against-trump-barr-ducey-kemp.html

A Gallup poll does say that there are fewer Republicans this year than in 2016, but polls are notoriously wrong. The article continues on to say that Trumps political decisions were the reason for the switch, not education. With Trump dismissing or actively angering his supporters, he changed how people voted regardless their educational standing. This article even includes quotes from people who are Republican and even voted for Trump in 2016, but didn't this year.

Republican ranks have shrunk this year. In January, more Americans said they were Republicans or Republican-leaning than Democrats or Democratic-leaning — 47% to 45%, according to a Gallup poll.
...
The polls reflect Trump’s odd political strategy. Instead of trying to expand the coalition that supports him, he has focused on bolstering his relatively narrow base of conservatives, white evangelicals and men without a college degree.

Along the way, he has alienated suburban women, college-educated voters in both parties, young people and — thanks to his chaotic response to the pandemic — older voters.
...
“I’ve been a Republican my whole life,” Stallings told me. “I still believe in small government, the 2nd Amendment and individual liberty. I voted for Trump in 2016. But I can’t vote for him a second time.”
...
“I voted for Trump with misgivings,” she said. “I believed once he felt the weight of the office, he’d knock off the bluster and the silliness. But he didn’t, and I think it’s cost the country a lot.”

https://www.latimes.com/politics/story/2020-11-01/bidens-secret-weapon-anti-trump-republicans

There's plenty of other reasons why Republican numbers are dropping, such as the older white population dying who make up a considerable number of Republicans, minorities leaving the party, and Republicans taking a hard stance on "family values" that women and LBGTQ+ disagree with.

The GOP is aging out. There’s an advantage to having older voters on your side — they vote more heavily than younger ones. The disadvantage is that they don’t stick around forever.

The GOP also depends heavily on white America. White voters furnish nearly 9 out of 10 Republican votes. But it’s the same problem here. White people vote at a higher rate than do minorities but are a declining proportion of the electorate. They were 87% of the voters in 1992. The number was 67% in 2020 and will continue to drop.

The GOP is paying a stiff price for defaming recent immigrants. Republicans are congratulating themselves on their inroads this year with Latino voters. Inroads? A gain of a couple percentage points isn’t a whole lot to brag about in an election where 2 out of 3 Latinos voted for President-elect Joe Biden. A 2019 poll found that 51% of Latinos believe that the GOP is “hostile” toward them, with an additional 29% believing that the GOP “doesn’t care” about them.

Asian Americans have also turned away from the GOP. They are America’s fastest-growing ethnic group and have the profile of a Republican bloc. They have the nation’s highest average family income and are twice as likely as other Americans to own a small business. As late as the 1992 presidential campaign, they voted 2-to-1 Republican. In 2020, they split nearly 2-to-1 Democratic.
...
There was no gender gap until the GOP adopted evangelicals’ version of family values, including opposition to abortion. Women are now the Democrats’ largest voting bloc and their loyalty has increased, reaching record highs in the past few elections. And Republicans’ embrace of evangelicals’ position on gay rights has alienated members of the LGBTQ community. They are now second only to Black Americans in their Democratic loyalty. The GOP’s rigid stance on social issues has also eroded its standing with college-educated voters. Once heavily Republican, they sided with Biden by an estimated 12 percentage points in 2020.

https://journalistsresource.org/studies/politics/elections/election-republican-party-future/

Race is also considered a major factor, along with education, when deciding which party to support. Gender is also a considerable factor, too.

Republicans hold wide advantages in party identification among several groups of voters, including white men without a college degree, people living in rural communities in the South and those who frequently attend religious services.

Democrats hold formidable advantages among a contrasting set of voters, such as black women, residents of urban communities in the Northeast and people with no religious affiliation.

https://www.pewresearch.org/politics/2020/06/02/in-changing-u-s-electorate-race-and-education-remain-stark-dividing-lines/

This doesn't even get into single-issue voters, which is a major can of worms, usually centered on abortion and business/economy.

You must log in to answer this question.

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