(Updated February 26, 2023)
More Americans have become increasingly and deeply divided by political party over roughly the past quarter-century. Last November’s election showed just how evenly split and polarized we are. Polarization is no longer news (see these 2012 and 2017 posts). But, what is new in the last few years is the tidal wave of research helping us understand the nature, history, breadth, and sources of this development. More has come into focus, especially on the connection between politics and personal identity. Journalist Ezra Klein’s 2020 book, Why We’re Polarized, is an excellent, informed analysis. I use this post to bring in the latest work and insights in order to describe, especially for the general reader, what happened. In Part 2 I will address polarization’s explanations, prognoses, and policies.
The bumper-sticker version of what follows is that division along party lines has spread from politics to many seemingly apolitical and private realms. Polarization has become deeper and more bitter as party identity–being a Democrat versus a Republican–has absorbed more and more of Americans’ other identities.
Two Kinds of Polarization
One kind of polarization (among several) happens when the divide between opponents on a specific issue widens, each side taking increasingly extreme positions. Not much of this has happened. Americans generally still cluster in the middle on most issues, even contentious ones.[1] And the middle is where they’d like to see the politicians operate.
The second kind of polarization, however, has riven the country: Americans have increasingly lined up their views on political issues–and on much, much more–with their party affiliations. Chasms have opened up not between policy positions (gun control yes or no; foreign aid yes or no), but between the political parties; rank-and-file members have followed suit. Abortion provides a telling example.
In the 1970s and ‘80s, the two party leaderships were not much apart on abortion policy. Rank-and-file Democrats and Republicans (whom I will abbreviate as D’s and R’s) held similar views. Today, however, there is a huge difference between the parties and the members on abortion. R’s moved toward more restrictive views and D’s toward less restrictive views, spurring ferocious battles. Yet, the average American position remain little changed, in the middle.[2]
This graph shows the percentage of Americans in the General Social Survey (GSS) from 1972 to 2018 who said “Yes” to the question of whether “it should be possible for a pregnant woman to obtain a legal abortion if she is married and does not want any more children.” For Americans as a whole there was little change in respondents’ answers (44% said yes in the 1970s, 48% said yes in the 2010s), but the divide between parties widened greatly.[3]
The widening chasm between parties appears in other domains of American public life (in just a year the party gap on funding Ukraine widened from 4 to 25 points) and in private life, too (on, for example, taking some future cancer vaccine).
What happened?
A few cautions: One, this is not about Trump. Party polarization started years before he rode down the golden escalator. Second, while more Americans have been registering as independents, about 30 percent now, most analysts see independents as more or less D or R “leaners.” Third and critically, most Americans have not caught the polarization fever. During the height of the January, 2023, battle over the House Speakership, a Politico reporter was hard-put to find even Republicans in Speaker-to-Be McCarthy’s home town who recognized his name or noticed the whole to-do. Amazing as it is to you and me, most Americans are neither active nor that much interested in politics. (More on that later). Still, those who are much interested have changed the nation.
What Happened?
First, the politicians polarized. Starting over 40 years ago, D and R office-holders began differing on issues more than they had for many years. R’s moved farthest from the shared middle ground. For example, in the 1970s D and R legislators held relatively similar positions on environmental policy. Richard Nixon had established the EPA. But by the 2000s, R and D office-holders sharply disagreed on environmental issues. Sticking to your party’s positions as the parties moved apart became the rule on Capitol Hill. And beyond Capitol Hill. Judges and justices, for example, increasingly decided cases in line with the politics of the president who appointed them. States also increasingly diverged in their policies, dividing the country into red-state versus blue-state blocs. It used to be that states tended to copy policies from neighboring states; nowadays they copy policies from states with the same red or blue tint.
Americans have noticed escalation of party conflicts and view both parties equally as too tolerant of extremism. Many nonetheless have themselves polarized, dividing more and more along party lines and more emotionally on political issues, social values, and lifestyles.
Political Issues. People’s party affiliations presumably should line up with their views on issues. But for much of the 20th century, that was only weakly true. Abortion, the environment, and immigration are examples of divisive issues today about which the two parties and their members differed little decades ago. Pew researchers looked at survey respondents’ views on ten policy issues. In 1994, 64% of R’s were more conservative on those topics than was the average D; in 2017, almost all of the R’s were. The same shift from the center is evident for D’s. Let’s be specific.
The key polarizing issue is race. Strikingly, sixty years ago R senators were more likely than D senators to vote for the historic 1965 voting rights law, but in 2022 no R senator and all D senators voted for a bill to reinforce that law. D and R Americans have come to disagree greatly about the reality of discrimination, against whom it happens,[4] ethnic slurs, affirmative action (although both R’s and D’s are against it)[5], what, if anything, ought to be done for greater racial equality,[6] and about pretty much everything racial.[7] This party divide happened not because R’s became more racist–they have generally become more tolerant–but because White D’s became more liberal on race.[8]
Polarization occurred on many other topics, too: abortion, as we saw, immigrants, immigration policy, health care, climate change, the death penalty,[9] gun permits, tax fairness, and so on. Yet newer issues repeat the pattern. Initial confusion about Covid policy sorted into partisan stances. Partisans on both sides of the 2020 election controversies agree that democracy is under threat but for vastly different reasons. For 71% of R’s the threat is voter fraud; for 81% of D’s it is voter suppression. And, as I noted above, the Russian-Ukrainian War quickly went from cross-party consensus to party divide.
Increasingly, people’s stances on a variety of issues line up in a partisan way. One can tell what people feel about one issue, say taxes, by knowing what they feel about another, say guns. Clearly distinct D and R packages of opinions have emerged.[10] And more than before, D’s and R’s disagree about what issues actually matter.
World Views. Values, morality, and even facts increasingly have a blue or red cast. D’s and R’s disagree about what defines being American: R’s point to birth, language, and religion, D’s to endorsing principles like democracy and freedom. Similarly, D’s and R’s disagree greatly about whether American society has improved or worsened since the 1950s, for example, whether it has become too soft and feminine (68 percent of R’s say it has versus 19 percent of D’s).
Americans generally trust institutions less than they did a few decades ago, but especially since the 1990s D’s and R’s have trusted different ones. D’s express greater confidence in the press, science, and elections (even before 2020); while R’s trust business, police, religion, and the military. Partisans even object to their relatives or friends associating with people in “opposing” institutions–for example, the police for D’s and universities for R’s.
R’s and D’s have diverged on personal issues, too: on the proper roles of men and women, sex, sex education, homosexuality, equal opportunity in America, and much more.[11] And D’s and R’s increasingly disagree about facts: whether extreme weather events have become more frequent (a 30-point gap), crime rates have risen, the motivations of terrorists, the causes of poverty,[12] and the plausibility of evolution and of astrology (D’s more often say that both are plausible than do R’s).[13]
Behavior. Is all this just verbal scrimmaging, posturing for pollsters? No. Differences have widened in behavior, too. In voting certainly. Voting the straight D or R ticket has become much more common (although the 2022 midterms showed, under serious provocation, a bit more ticket-splitting). Party also correlates with how Americans spend their time and money.[14] D’s and R’s tend to shop at different places–Trader Joe’s versus Sam’s Club, for example; drive different cars–Subarus versus Chevies; drink different coffees–Nespresso versus Mr. Coffee.[15] D’s and R’s watch different kinds of television shows and, of course, follow different news media. Experimental studies suggest that Americans are now likelier to discriminate against someone, say in awarding a job, based on the candidate’s party than based on their race.[16]
D’s and R’s increasingly live in different places. There are red and blue regions, red and blue states, red and blue counties, and red and blue zip codes within counties; one researchers described it as a form of political “apartheid.” Partisans live in different kinds of places, D counties typically thriving economically and R counties more often declining. About 15 years ago, journalist Bill Bishop wrote The Big Sort, describing early evidence of this process; it has accelerated since. Indeed, identifying with places–particularly as a city or a country person–has become a defining feature of which side people are on.
Americans moving just in order to have neighbors of the same party accounts for little of this sorting. Sorting mostly follows from movers looking for places that they like and can afford. Because place tastes differ–say, for communities with large yards and big churches versus communities with “walkability” and ethnic food diversity–by party movers tend to get neighbors of the same party indirectly.[17] Geographical clustering has raised the number of “landslide counties,” those where a D or R candidate wins overwhelmingly. While the nation is evenly divided politically, there are fewer and fewer evenly divided communities.
Party has become tied to yet more intimate aspects of Americans’ identities, to, for instance, their faith (previously discussed in these 2013 and 2018 posts). D’s in particular have increasingly reported having no religion, not attending services, and not praying, while R’s have stayed more engaged in churches.[18] A study of Twitter users–a highly political group–found that between 2015 and 2018 the percentage who used political labels to describe themselves overtook the percentage who used religious labels. (This case may be a Trump effect.)
Party also shapes Americans’ personal relationships. About half of Americans say that they have no or very few important relationships with people on the other political side. Many report breaking online or even in-person ties over politics (D’s somewhat more than R’s).[19] Americans increasingly marry someone of the same political affiliation and object to a child marrying across party lines (again, D’s more than R’s). Wary parents may not need to scupper a political intermarriage, since children and parents agree politically more than before.
Finally, political commitments are driving life-and-death behavior: Covid. After an initial period of comity, political actors sorted out what Americans should think and what they then should do about Covid, washing hands, wearing masks, social distancing, and, of course getting vaccinated. Even though it was an R administration that deployed Big Government at “Warp Speed” to develop the vaccines, vaccination became something that D’s did and so R’s did not.[20] Death by party allegiance followed. In the early part of the pandemic, whether Americans died from Covid depended mainly on factors like their ages, incomes, jobs, and neighborhoods. But since vaccinations became easily available, party identification became a critical factor. Many R’s died for their refusal to vaccinate. (This capped 20 years in which red-state death rates increasingly outpaced blue-state death rates.[21])
Emotion. Heightened emotions suffuse these party divisions, what researchers call “affective polarization.” Partisans have come to distrust, fear, and hate the other side more. A simple example is that Americans rate presidents of the other party much more negatively now than they did before.[22] But emotion–especially hostility–shows up in many ways, such as objecting to having in-laws of the other party. Also, increasingly what people find offensive is less that their opponents are wrong and more that their opponents seem to them dishonest and immoral.
Review
Political party now divides more Americans in more ways than it used to. From matters only political to matters deeply personal, these face-offs are bathed in intensifying emotions.
Among western nations, such polarization is particularly American (although Canadians may be catching the bug). Americans report bigger political divisions than other nationalities do, have divided faster, differ much more by party on fundamentals such as what makes for a good citizen, and are the most hostile toward political opponents.[23] Fierce polarization is not new in American history; we have had earlier periods of intense, bloody confrontations. But, today’s polarization is unprecedented for at least a century.
One recurrent question is whether the change has been symmetrical, D’s and R’s equally moving from central ground and entrenching against the others. The answer varies by issue, but over the range covered here, both sides have consolidated and defined their positions. Klein argues that the polarization is asymmetrical because the D’s are much more internally diverse culturally and ideologically. I am not sure. On many topics discussed above, such as race, guns, and religion, the D’s have moved the farthest from the average position.
Paradoxically, on some issues, particularly moral and social ones, D’s and R’s have actually moved in the same direction. But because they moved at different rates, the gaps between them nonetheless still widened. Homosexuality is an example. Since 1973, the GSS has asked people what they thought about “sexual relations between two adults of the same sex.” In the 1970s and 1980s, D’s’ and R’s’ answers differed by about three percentage points; only 13% of all Americans said that homosexuality was “not wrong at all.” By the 2010s, the party difference in that reply had grown to 20 points. Both more D’s and R’s came to say “not wrong at all” over those thirty years, but the D’s (57%) much faster than the R’s (37%).[24] A similar example is attitudes toward ethnic minorities. Members of both parties have “liberalized” on race-ethnicity, but D’s earlier and faster. Americans as a whole have moved leftward on cultural issues, but that has prompted a fierce backlash among those who virulently reject those moves, inflaming polarization.
Still, remember that most Americans are not involved in these fights. About 40 percent of American adults do not vote even in presidential elections, 60 percent not in the midterms. Most people discuss politics only occasionally,[25] a minority follow cable news. To be sure, interest in politics has risen,[26] but most Americans are still not into politics. Even Twitter is overwhelmingly about nonpolitical topics. Similarly, not all topics are subject to party division[27] and there is no party cleavage on some core values, such as the primacy of family and friends. Those who partake in this polarization tend to be, as one might expect, the more educated, those more interested in politics,[28] those who follow politics as others follow sports. Nonetheless, their widening divisions and heated rhetoric are splitting all of America.
It’s About Me and You
What has happened, researchers largely agree, is that more Americans have defined who they are primarily by their politics and less by other identities–their ideology, religion, region, marital status, hobbies, etc.–identities which have now become increasingly subordinated to party. Politicized Americans have adopted one of two “mega-identities,” as Lilian Mason puts it, R or conservative versus D or liberal, with all the associations those tags now evoke. Confrontations between the two mega-identities intensify each one and heighten “team” loyalty. The way Klein frames it, “the question for the unengaged citizen is: what will this policy do for me,” while the engaged citizen asks, “what will support for this policy position say about me.”
Party increasingly says man or woman, religious or skeptical,[29] country person or a city one,[30]. The parties come to be homes to different kinds people and people want to be at home with people like themselves. Indeed, as I will discuss more in the next post, some people even adjust their self-descriptions–their religion and their ethnicity–to better fit with their party. And they increasingly care enough to despise the other.
Politics is increasingly identity politics for both sides (masculine, Christian, patriot versus questioning, agnostic, global-minded) and about what side you are on. Polarization, however, is not only about individual identity; it is also about institutional identity. Some churches, schools, workplaces, clubs, and even families have been split by the political wars and others have mobilized to fight those political wars, enlisting individuals who may otherwise have stayed on the sidelines.
How and why did we get here? I’ll turn to that in a second post.
ENDNOTES ——————————————————————
[1]. See studies one, two, three, four.
[3]. The graph and reported numbers exclude respondents who claimed to be true independents. I combined those independents who “leaned” one way or the other with the party partisans.
[4]. For example, R’s reject the idea that Blacks suffer serious discrimination, but overwhelmingly D’s say they do. R’s are likelier to say that discrimination against Christians and Whites was a problem than to say that discrimination against Blacks, gays, Latinos, or Jews was a problem; D’s go strongly the other way. See also studies one, two, three, four.
[5]. In a 2022 Pew survey, 87% of R’s and 62% of D’s said that ethnicity or race should be not a factor at all in college admission decisions.
[7]. Interestingly, R’s will vote for Black candidates if those candidates agree that Whites are being put upon by Black demands.
[8]. Studies one, two, three, four, five, six, seven.
[9]. From the 1970s through 1990s, R’s were about 15 points likelier to approve of capital punishment than were D’s. Then, more D’s started disapproving and by the 2010s, the gap was 33 points (my analysis of the GSS item CAPPUN).
[10]. Studies one, two, three, four.
[11]. On gender roles, see, e.g., study one and two. Both D’s and R’s have become much more willing to say that premarital sex is “not always wrong,” but the gap widened between from 11 to 16 points (my analysis of the GSS). On sex education, huge differences exist. Many have studied the changes in attitudes toward homosexuality; a recent one is here. Equal opportunity, here.
[12]. Explaining inequality, see, for example, here.
[13]. Evolution: GSS item EVOLVED, 2006-18: D’s were much likelier to say yes, humans evolved from other animals, even after controls for race, age, gender, religion, and education (party mattering as much as having a college degree). Astrology: ASTROSCI, 2006-18: How scientific is astrology?: D’s likelier to say more so, even after controls.
[14]. Studies one, two, three, four, five.
[15]. And then there is buying guns or not: In the 1970s, gun ownership rates differed by 8 points (reported in 48% of D households, 56% of R households), but in the 2010s, they differed by 26 points (24% versus 50%)–my analysis of the GSS variable OWNGUN.
[16]. See studies one, two, three.
[17]. Studies one, two, three, four, five.
[18]. As late as the 1980s GSS, D’s were only 2 points more likely to say they had “no religion” than were R’s; in the 2010s, the gap was 13 points (GSS item RELIG). As late as the 1980s GSS, D’s were only 4 points less likely than R’s to report going to church services at least once a month; in the 2010s, the gap was 13 points (GSS item ATTEND). (See graph below.) D’s became a bit likelier since the 1990s to say they prayed rarely or never and Republicans a bit more likely to say they prayed daily (GSS item PRAY). See also studies one, two, three, four.
[19]. On the political composition of people’s personal networks, see studies one, two. On cutting ties, studies one (the Luntz quote), two, three.
[20]. Hand-washing, here. Distancing: Studies one, two, three. Vaccination: Studies one, two, three, four document growing party differences and studies five and six show party effects after controls for covariates. Covid vaccination resistance even spilled over into R flu-vaccination resistance.
[21]. “From 2001 to 2019, the death rate in Democratic counties decreased by 22 percent, according to a recent study; in Republican counties, it declined by only 11 percent. In the same time period, the political gap in death rates increased sixfold.”
[22]. Studies one, two, three. More generally, on the increase in negativity toward the other party: studies four, five,and six. For a modest dissent on whether negative partisanship is ascendant, see study seven.
[23]. Studies one, two, three, four, five. On explaining this cross national difference, see, e.g., here.
[24]. My analysis of GSS item HOMOSEX.
[25]. My analysis of the American National Election Survey (ANES), item vcf0732, 1984-1998. A 2021 survey asked: “How often do you discuss politics and government with your friends?”Twenty-one percent said a few times a week or more, 24% a few times a month, 55% less often.
[26]. I analyzed an ANES question (vcf0310) that asked respondents how interested they were in the political campaign of the year. From 1952 into 2000, the percentage very interested lingered at or under 30% and then, starting in 2004, that percentage jumped to 40 and up to almost 50 percent by 2016.
[27]. For example, although GSS respondents’ views on whether it is better “if the man is the achiever outside the home and the woman takes care of the home and family” show the expanding gap pattern (in the 1980s, half of both D’s and R’s agreed; in the 2010s, 77% of D’s and 63% of R’s disagreed–item FEFAM), their views on whether women should “leave running the country up to men” remained similar (with both about 65% of D’s and R’s disagreeing in the 1970s and 85% disagreeing in the late ‘90s, when the GSS stopped asking the question, an example of near-completion of a change process–GSS item FEHOME).
[28]. Using the ANES, 1978-2016, I calculated the absolute difference between D and R thermometer ratings (vcf0218 and vcf0224). Through 2000, the mean absolute difference had risen from 30 or less ; after 2000 it rose sharply to nearly 40 by 20012 and 2016. Gaps were greater for college graduates and those very interested in campaigns, especially after the turn of the century. See also here.
[29]. Studies one, two, three.
[30]. See an earlier blog post. Also studies one, two, three.
MADE IN AMERICA 