The Geography of Partisan Prejudice

Editor’s note: The maps in this article have been corrected to address problems with two entries in the underlying data. People searching for some counties were shown different counties, and some saw information that didn’t match the county they’d searched for.

We know that Americans have become more biased against one another based on partisan affiliation over the past several decades. Most of us now discriminate against members of the other political side explicitly and implicitly—in hiring, dating, and marriage, as well as judgments of patriotism, compassion, and even physical attractiveness, according to recent research.

But we don’t know how this kind of stereotyping varies from place to place. Are there communities in America that are more or less politically forgiving than average? And if so, what can we learn from the outliers?

To find out, The Atlantic asked PredictWise, a polling and analytics firm, to create a ranking of counties in the U.S. based on partisan prejudice (or what researchers call “affective polarization”). The result was surprising in several ways. First, while virtually all Americans have been exposed to hyper-partisan politicians, social-media echo chambers, and clickbait headlines, we found significant variations in Americans’ political ill will from place to place, regardless of party.

We might expect some groups to be particularly angry at their political opponents right now. Immigrants have been explicitly targeted by the current administration, for example; they might have the most cause for partisan bias right now. But that is not what we found.

In general, the most politically intolerant Americans, according to the analysis, tend to be whiter, more highly educated, older, more urban, and more partisan themselves. This finding aligns in some ways with previous research by the University of Pennsylvania professor Diana Mutz, who has found that white, highly educated people are relatively isolated from political diversity. They don’t routinely talk with people who disagree with them; this isolation makes it easier for them to caricature their ideological opponents. (In fact, people who went to graduate school have the least amount of political disagreement in their lives, as Mutz describes in her book Hearing the Other Side.) By contrast, many nonwhite Americans routinely encounter political disagreement. They have more diverse social networks, politically speaking, and therefore tend to have more complicated views of the other side, whatever side that may be.

We see this dynamic in the heat map. In some parts of the country, including swaths of North Carolina and upstate New York, people still seem to give their fellow Americans the benefit of the doubt, even when they disagree. In other places, including much of Massachusetts and Florida, people appear to have far less tolerance for political difference. They may be quicker to assume the worst about their political counterparts, on average. (For an in-depth portrait of one of the more politically tolerant counties in America, see our accompanying story on Watertown, New York.)

To do this assessment, PredictWise first partnered with Pollfish to run a nationwide poll of 2,000 adults to capture people’s feelings about the other party. The survey asked how people would feel if a close family member married a Republican or a Democrat; how well they think the terms selfish, compassionate, or patriotic describe Democrats versus Republicans; and other questions designed to capture sentiments about political differences.

Based on the survey results, Tobias Konitzer, the co-founder of PredictWise, investigated which demographic characteristics seemed to correlate with partisan prejudice. He found, for example, that age, race, urbanicity, partisan loyalty, and education did coincide with more prejudice (but gender did not). In this way, he created a kind of profile of contemporary partisan prejudice.

Next, Konitzer projected this profile onto the broader American population, under the assumption that people with similar demographics and levels of partisan loyalty, living in neighborhoods with comparable amounts of political diversity, tend to hold similar attitudes about political difference. He did this using voter files acquired by PredictWise from TargetSmart, a commercial vendor. Voter files are essentially data snapshots about all American adults, based on publicly available records of voter registration and turnout from past elections, along with data about neighborhood variables and demographic traits. In this way, PredictWise was able to rank all 3,000 counties in the country based on the estimated level of partisan prejudice in each place. (For more technical detail about the methodology, click here.) “What I find most striking is that we find a good degree of variation,” Konitzer says. Some states, like Texas, show a real mix of prejudiced and nonprejudiced counties; whereas Florida is very consistent—and fairly prejudiced—from place to place.

Nationwide, if we disregard the smallest counties (which may be hard to pin down statistically, since they have fewer than 100,000 people), the most politically intolerant county in America appears to be Suffolk County, Massachusetts, which includes the city of Boston. In this part of the country, nine out of every 10 couples appear to share the same partisan leaning, according to the voter-file data. Eight out of every 10 neighborhoods are politically homogeneous. This means that people in Boston may have fewer “cross-cutting relationships,” as researchers put it. It is a very urban county with a relatively high education level. All these things tend to correlate with partisan prejudice.

We now assume that the other political side is much more extreme than it actually is, as Matthew Levendusky and Malhotra have found. In a 2012 survey, they found that Republicans rate fellow Republicans as more hard-line on taxes, immigration, and trade than they actually are; and Democrats rate Republicans as even further to the right.

These distortions lead us to make worse decisions. Most obviously, politicians refuse to compromise on things like border walls and budgets, even when it hurts the country. But regular people’s judgments get warped too. For example, parents are less likely to vaccinate their children when the other party’s president is in the White House, according to a 2019 working paper by the Stanford Ph.D. candidate Masha Krupenkin. Regardless of who is in power, mutual-fund managers are more likely to invest in funds handled by fellow partisans, a bias that does not lead to better returns.

The irony is that Americans remain in agreement on many actual issues. Eight out of 10 Americans think that political correctness is a problem; the same number say that hate speech is a concern too. Most Americans are worried about the federal budget deficit, believe abortion should be legal in some or all cases, and want stricter gun regulation. Nevertheless, we are more and more convinced that the other side poses a threat to the country. Our stereotypes have outpaced reality, as stereotypes tend to do.

By contrast, the North Country, in far upstate New York, just east of Lake Ontario, seems to be more accepting of political differences. The same seems to be true in parts of North Carolina, including Randolph, Onslow, and Davidson Counties. In these places, you are more likely to have neighbors who think differently than you do. You are also more likely to be married to someone from the other side of the aisle. It’s harder to caricature someone whom you know to be a complicated person.

Other research has also found that more educated and politically engaged people tend to be more politically prejudiced. But the PredictWise analysis also detected a correlation with urbanicity and life stage. Older Americans and people living in or near sizable cities, from Dallas, Texas, to Seattle, Washington State, seem to be more likely to stereotype and disdain people who disagree with them politically.

We don’t know what is causing what, unfortunately, as is often the case in sociological research. We just know that being older and living in or near a city seem to go along with partisan prejudice in general. This may be because, according to decades of research into how prejudice operates, humans are more likely to discriminate against groups of people with whom they do not have regular, positive interactions. (In Europe, some research suggests that anti-immigrant sentiments tend to be higher in people who live in homogeneous neighborhoods near—but not among—immigrants.)

And in America, people who live in cities (particularly affluent, older white people) can more easily construct work and home lives with people who agree with them politically. They may be cosmopolitan in some ways and provincial in others.

Americans now routinely guess one another’s partisan leanings based on what they eat, drive, and drink (Dunkin’ Donuts? Republican; Starbucks? Democrat), according to a working paper by the University of Pennsylvania Ph.D. candidate Hye-Yon Lee. And based on these unreliable cues, they say they’d be more or less likely to want to live, work, or hang out with one another.

We are now judging one another’s fundamental decency based on whether we eat at Chipotle or Chick-fil-A. This may seem silly—harmless, even. But it is uncomfortably reminiscent of stories from conflict zones abroad. In Northern Ireland, for example, an outsider visiting during the Troubles had no way to tell unionists and nationalists apart. They were pretty much all white Christians, after all. But the locals themselves routinely guessed one another’s identity based on their names, the spacing of their eyes, their sports jerseys, the color of their hair, their neighborhood, or even how much jewelry they wore­. This process came to be known as “telling.” If a reliable cue didn’t exist, people would make one up. It was a way to move about in the world in a time of profound tribalism, during which 3,600 people were killed.

In parts of America, it is markedly more uncomfortable to be perceived as a Democrat right now. In other places, it is very isolating to be outed as a Republican. To get a sense of these differences, we asked PredictWise to do two other rankings—this time reflecting the directional flow of partisan prejudice. The resulting maps reveal places where Democrats are the most dismissive of Republicans and vice versa.

In general, Republicans seem to dislike Democrats more than Democrats dislike Republicans, PredictWise found. We don’t know why this is, but this is not the only study to have detected an imbalance. For example, in a 2014 survey by the Pew Research Center, half of consistently conservative respondents said it was important for them to live in a place where most people share their political views—compared with just 35 percent of consistent liberals. But a more recent survey, conducted in December by The Atlantic and the Public Religion Research Institute, found that Democrats were the ones showing more ill will—with 45 percent saying they’d be unhappy if their child married a Republican (versus 35 percent of Republicans saying they’d be unhappy if their child married a Democrat). So it’s hard to know exactly what’s going on, but what’s clear is that both sides are becoming more hostile toward one another.

Conflict and protest are vital to democracy. But whenever people begin to caricature one another, anywhere in the world, predictable tragedies occur. Fixable problems do not get fixed. Neighbors become estranged, embittered, and sometimes violent. Everyone ends up worse off, sooner or later. “This is the great danger America faces,” Representative Barbara Jordan of Texas said in 1976. “That we will cease to be one nation and become instead a collection of interest groups: city against suburb, region against region, individual against individual. Each seeking to satisfy private wants.”

Partisan prejudice is different from other forms of prejudice. It is not yet embedded in all of our institutions, the way racism has been. But the evidence shows that it distorts our thinking, just like other kinds of prejudice. “Just like with race, the problem is that when people stereotype, they miss the variation within a group,” says Stanford University’s Neil Malhotra, who has researched political behavior for more than a decade.

Fundamentally, partisan prejudice is another way for one group of humans to feel superior to another. New research suggests that it is now more acceptable in some areas of life than racial prejudice. In a 2012 experiment, the political scientists Shanto Iyengar and Sean Westwood gave nearly 2,000 Americans implicit-bias tests and found that partisan bias was more widespread than racial bias. About 70 percent of Democrats and Republicans showed a reflexive bias for their own party. (Take a version of this test here.)

Of course, it can be harder to tell someone’s political leanings than someone’s skin color. And it’s hard to develop an implicit-bias test that mimics realistic, everyday encounters. But when people think they can guess someone’s political leanings, they discriminate accordingly.

In a 2014 study, Karen and Thomas Gift at Duke University sent out 1,200 resumes, tweaking some to suggest a candidate with previous experience in a Democratic or Republican organization. And employers seemed to notice. In a conservative county in Texas, a Republican applicant had to submit about five resumes for each positive callback. By contrast, a Democratic applicant needed to submit seven resumes to get a callback. (And the Republican candidates had a similar disadvantage in a liberal California county.)

What makes this kind of prejudice unusual is that it is currently very easy to defend. What is wrong with discriminating against someone based on political values? After all, unlike race or sexuality, politics is something you choose. If you choose unwisely, maybe you deserve to be judged accordingly.

Yes and no. We have more choice over our politics than over our sexuality, without a doubt. But the vast majority of people follow their parents’ lead when it comes to party affiliation, just as they do with religion. In fact, some researchers have even found that political tendencies are significantly influenced by genetics, with identical twins sharing even more political opinions than fraternal twins.

Most people adopt a political team at a young age and very rarely change—regardless of whether they make more money or need more government help at different life stages. Political preferences are not rational or linear decisions, even though they feel that way. “People bind themselves into political teams that share moral narratives,” Jonathan Haidt writes in his book The Righteous Mind. “Once they accept a particular narrative, they become blind to alternative moral worlds.”

About four in 10 Americans identify as independent today, but even they pick sides. Most independents consistently lean either right or left in their voting behavior over time and tend to exhibit similar prejudices as people who claim a specific party.

As politics have become more about identity than policy, partisan leanings have become more about how we grew up and where we feel like we belong. Politics are acting more like religion, in other words.

This is partly because partisan identities have begun to line up with other identities, as Lilliana Mason describes in her book, Uncivil Agreement. Making assumptions about people’s politics based on their race or religiosity is easier than it was in the past. Black people get typed as Democrats; people who go to church on Sunday are assumed to be Republicans. (But as always, stereotypes still mask complexity: About half of black Americans go to church at least once a week, for example, a far higher rate than that of white Americans.)

In other words, partisan prejudice now includes a bunch of other prejudices, all wrapped up into one tangled mess. “Americans are really divided, but not in terms of policy; they’re divided in terms of identity,” Mason says. “And the more identities come into play, the more salient they are, the harder it will be to agree, even if policy positions shift.” Politics are becoming a proxy battle for other deep divisions that have almost nothing to do with environmental regulation or tax policies.

Hope is embedded in all these maps: This kind of prejudice is malleable. That is why it varies so much from place to place. By cultivating meaningful relationships across divides, by rewarding humility and curiosity over indignation and righteousness, people can live wiser, fuller lives. They can also learn to speak one another’s language, which means they might one day even change one another’s minds. This happens organically in some places, we now know. Maybe it’s time to think of these outliers as rare and interesting, worthy of our attention, before they become extinct.