(Updated April 26, 2023)
A political identity–being a liberal or Democrat (a “D”) versus being a conservative or Republican (an “R”)–has become viscerally more important to more Americans, reaching partisan intensities not seen for at least a century and not seen elsewhere in the western world. R’s and D’s have increasingly and heatedly differed on public issues and on private ones as well, such as who to marry, who to befriend, where to live, what to buy, and the nature of faith. It is the era of the great polarization.
My previous post, Part I, described how so many Americans became so divided over the last 40 or so years. In this post, I try to explain why, evaluate the consequences, and consider what might be done. The short answer to why is that the nation’s political leaders first sorted themselves into separate and increasingly well-defined camps. Politically-attuned Americans could more clearly see where the parties stood on their big concerns, first race and then later “culture war” arguments. Different sorts of people became D’s or R’s and identifying as a D or R made people different. Peer influences, mass media, and social media did not create polarization, but accelerated it, particularly raising the emotional heat. In several ways, these forces all fed one another. One can argue that by now the fervor of the rank and file drives leaders to further polarize.
Explanations for Why Democrats and Republicans Became So Different
People Sorting. If members of one voluntary association differ from members of another, a lot of self-sorting has been going on. This is a major reason that D’s and R’s differ ideologically, culturally, and behaviorally. A person who decides that her taxes are too high or that there are too many immigrants can become an R; a person who decides that his wages are too low or that there is too much religion in the public square can become a D. But much more than such straightforward politics explains how R’s and D’s diverged so much.
The prime reason an Americans is a D or an R is because their parents were D’s or R’s. People largely inherit their parents’ party allegiances (perhaps genetically, some argue). But this inheritance pattern weakened, at least for non-Hispanic Whites, during the last decades of the 20th century. Eighty percent of White Americans in the 1950s with D fathers had stayed D’s; only 15% had become R’s. By the 1980s, however, only 65% of Whites with D fathers had stayed D’s; 26% had become R’s.1 This change was notably greater in the South.2 Such numbers detail the oft-noted surge of White Southerners out of the D and into the R party in a backlash to the Civil Rights era. Further erosion of D party inheritance followed as deployment of racial and cultural issues attracted less-educated Whites into the R party, even at the cost of repelling some higher-educated ones. Donald (“I love the poorly educated”) Trump cashed in on this new sorting.
Part I described indirect sorting, particularly by geography. Few Americans move just to live near fellow R’s or D’s, but some people move for reasons–say, a taste for diversity or a focus on family–that bring them nearer to people who share their lifestyles and world-views. Given the tightening connection of styles and views to parties, the movers thus sort themselves into communities with neighbors of the same party. The same indirect sorting process no doubt occurs in choosing other social settings–churches, volunteer activities, colleges, and jobs, for example.
Changing People. Although self-sorting into parties is the major reason that D’s and R’s increasingly differed from each other, party allegiances have also changed Americans. Rather than picking a party that fit their existing views, some Americans have changed their views and practices to match their parties. So, for example, the massive shift on abortion displayed in Part I–from R’s and D’s generally agreeing on abortion in the 1970s to R’s and D’s disagreeing by huge margins today–is partly the result of people changing parties to fit their abortion views and partly of people changing their abortion views to fit their parties.
More dramatic still, some people have changed the specific identities they claim–religion, race, ethnicity, sexuality perhaps–to match their party identities. One analysis found that “White Americans who viewed Trump favorably and did not identify as evangelicals in 2016 were much more likely than White Trump skeptics to begin identifying as born-again or evangelical Protestants by 2020.” In another study some D’s shifted from describing themselves as non-Hispanic White to claiming membership in an ethnic minority.3
Forces of Change
D’s and R’s polarized as people changed parties to suit their views and changed views to suit their parties. The first change seems clear. But what has driven the second?
Peer Influence. One process compounds sorting. Even people who have not moved in order to find compatible neighbors are influenced by their neighbors; local residents converge politically to the left or the right.4 We can assume that similar polarizing dynamics occur in other social contexts, such as congregations and workplaces, as well. For example, in the 1970s and ‘80s, White D’s were almost as likely to be frequent church attenders as White R’s; by the 2010s they were half as likely to be. Liberal church attenders who would not be swayed by rising conservatism, especially in Protestant congregations, left.5
Mass Media. Many have blamed the mass media for floods of partisan news and opinion that drag Americans to one extreme or the other. Commentators have opined along these lines and the public tends to agree. Data journalist Kevin Drum’s most concise explanation (with data) for polarization is that “It’s All About Fox News.” Not only are Fox devotees distinctively conservative, even among Rs, availability of Fox News in an area seems to, by itself, bump up R turnout. Media on the left side of the dial, like MSNBC, are not nearly as potent.
The main reason news sources have different audiences, however, is not because they shape the audience, but, again, mainly because people sort themselves into particular audiences. Political scientist Matt Grossman put it this way: “Media choice has become more of a vehicle of political self-expression [for partisans] than it once was…. while most citizens tune out political news as best they can.” An interesting revelation from the Dominion Voting System’s defamation suit against Fox News is how much Fox leadership felt compelled to follow, not lead, their election-denying viewers.
Nonetheless, the news media do modestly change their audience–mainly by intensifying partisan anger, raising the motivation to vote. And one way that media charge up their viewers is by showcasing extreme politicians.
Social Media. Social media operate similarly, though modestly, to amplify partisan hostility. One experiment found that altering the news mix on Facebook feeds did not change users’ politics but did affect their political emotions. Another online study found that making people pay attention to opposing opinions leads some partisans to just dig in more. Thus, social media, like mass media, seem to facilitate and inflame polarization. But they do not appear to be key instigators of polarization.
Leaders. Some analysts credit political entrepreneurs with having destroyed the collegiality of earlier eras and fomented contentious politics, often pointing particularly to House Speaker Newt Gingrich in the 1990s and Donald Trump in the 2010s. (But, surely, Reagan and Obama also epitomized what it meant to be R or D, say, middle-Americana versus urbane.) People who care about politics take their cues from party leaders. We saw in Part I how American office-holder polarized among themselves before the wider public did; the interested public has followed.
Take, for example, the furor over “critical race theory.” It seems to have been sparked by the Trump administration latching on to an esoteric complaint made on Fox TV by a conservative think tank intellectual. In short order, an Ipsos poll showed that 64% of R’s agreed and 73% of D’s disagreed that “public schools are teaching children that all white people are racist, by teaching critical race theory” though most Americans don’t even know what CRT is.
Leaders’ initiatives on issues such as gendered bathrooms or reparations soon mobilize partisan rank and file who may have thought little about those topics before. In a pointed case, a recent study found that many attentive R voters falsely denied that they voted by mail, lying, it seems, to demonstrate partisan allegiance.
And why did American political leaders hive off into such homogeneous, opposing camps? That question would lead to a complex discussion including primaries, gerrymandering, voter targeting, campaign strategy, and much more that is beyond this essay.
Deeper Social Changes and Conditions? What about deeper social trends that may underlie the political maneuvering? Recent polarization seems entwined with concern about three such developments: Blacks’ advances and ongoing social justice demands (from Civil Rights to Obama to Black Lives Matter), accompanied by White resentment;6 women stepping out from the domestic sphere and making claims for further equality, also spurring reaction7; and young Americans challenging traditional social conventions, be they about premarital sex, accepting LGBT people, religion, abortion, or pot, yet again stirring backlash.8 Notably absent in this list are struggles over labor, communism, or war that flared in the 1950s and ‘60s. (Race was critical then as now, but then race is an eternal American obsession.)
At a yet deeper level are the distinctive foundations of the U.S. political system, notably those that cement in a two-party system, such as single-representative constituencies, winning office with only a plurality, and separate election of presidents. These features encourage face-offs between two directly opposed blocks and partly explain why America is more polarized than comparable nations.9
What Might be Done?
Can these trends go on indefinitely? Can the D’s and R’s who held similar views on abortion in 1980 and then divided on abortion by 30 points 40 years later eventually divide by 60 points in 2060? Can all the social sorting just continue? No social trend can go on forever, but how it ends matters. Might “affective polarization” escalate to what some alarm bell-ringers foresee as a civil war?
On the less alarmist side, we need to remember that most Americans–for better or for worse–are not caught up in politics. Even the nightly viewers of Fox News represent fewer than one percent of American adults.10 Another calming thought: the unusually close balance between the two parties that we have today, which contributes to the tension, might abate. The margins of victory in recent presidential elections have been unusually narrow.11 Returning to the more historically typical pattern of dominance by one party might lower the heat.
How else might we lower the heat? But, first, do we want to lower the heat?
While most observers see polarization as a problem, it does clarify party positions and in that way promotes political interest and democracy. Indeed, voter turnout in presidential elections has grown along with polarization in the last 40 years, reversing a long decline over the preceding 100 years.12 Clarifying polarization was what the American Political Science Association recommended in 1950, when it argued that the Tweedledee and Tweedledum parties of that era should move apart and become defined ideological opponents.
Most observers, however, believe they have moved too far apart. Polarization makes effective governing harder as parties devote themselves to winning every possible substantive or symbolic issue. The result, concludes one review, is “drift.” “Even if party polarization was the norm throughout the nineteenth and early twentieth centuries, the responsibilities of the federal government today are so much broader that policy stalemate has far more wide-ranging consequences.” Similarly, polarization may be leading many Americans on both sides to place winning power over preserving a liberal (in the classic sense) democracy, with intense partisans willing to override due process, minority rights, and even majority votes.13
Polarization, the product of social sorting, generates yet further and more bitter social sorting. Survey respondents who see high levels of polarization tend to express distrust in people generally. Conversations are stifled so as to avoid tension. (I’ve been on a couple of group tours where the announced rule was “no politics!”) Relationships have been sundered or avoided because of politics. Religious congregations are fracturing.
Speaking in Gettysburg, PA, in 2020, then-candidate Biden warned, “The country is in a dangerous place.” Too many Americans are engaged in “total, unrelenting, partisan warfare.” Forty percent of Biden voters and half of Trump voters agree that the “situation in America is such” that they would favor blue or red states “seceding… to form their own country.” In a 2022 survey, a fifth of Americans seemed to be OK with political violence and in another one most R’s and D’s said they expected political violence to increase. Two in five said that a “civil war” was at least “somewhat likely.” In yet another 2022 survey, half agreed that there would soon be a civil war.
Perhaps these survey respondents were just posturing or blowing off steam, but even that is bothersome; it helps generate an atmosphere of doom. The hostile polarization of recent years may have contributed to a broader demoralization, such as growing distrust in key American institutions and the modest but real jump since 1990 in the percentage of Americans willing to tell pollsters that they are unhappy.14 Of course, a lot else has happened since 1990, including 9/11, two wars in the Mideast, and the Great Recession, so one cannot claim that polarization is the cause of that jump, but polarization could not have helped.
If we agree that polarization is a problem, what might be done?
Many organizations, civic leaders, and philanthropists have stepped up. (See here for a comprehensive overview.) Some public experiments to reduce polarization have gotten wide attention.15 For the most part, the experiments–be they real-world interventions or academic studies–have focused on changing individuals’ views, especially their images of the other side, by fostering face-to-face conversation between D’s and R’s in one-on-one or small group meetings. The academic experiments have had some successes, some failures, and some cases of aggravating polarization.16
Even with successes, we can ask whether such experiments can scale up from influencing hundreds to influencing millions of people. We can also ask how effective they can be given the broader historical, structural, and institutional sources of our contemporary circumstances. What might be done structurally to abate polarization? Some point to reforming American politics.
Ezra Klein suggests getting more Americans to vote thereby diluting the power of the hyper-partisans. But experience suggests that marginally interested voters are mobilized precisely by more polarization. The 2020 election had a huge turnout surge, to the highest level in 60 years, only to be followed by unusually intense bitterness.
Other suggestions are to change our political rules and procedures–say, moving toward proportional representation, ranked-choice voting, open primaries, or combined (“jungle”) primaries. Based on my quick scan of the literature, the evidence for changing primary rules to foster moderation seems mixed.17 As to ranked-choice voting, again my impression is that the evidence is mixed, perhaps because that procedure is so new to most Americans.18
Even with progress along such lines, can an America that has been so sharply sorted by party in so many ways unsort itself? Can Americans who have made party their “mega-identity” return to other identities? It is hard to imagine today’s partisans (which includes me and probably you) coming out of their corners. Perhaps the let-us-reason-together solutions that are being tested can scale up. Alternatively, as I mentioned earlier, a clear victory and long-term prominence of one side (my side, I hope) can dampen down the fighting. And there are always new voter generations coming; their formative experiences–for example, being confronted by partisan excesses–may turn them away from polarization. Hopefully.
——————————————————————————————
Endnotes
- I analyzed the American National Election Survey [ANES]. In the second half of the 20th century, about 70% of White respondents with D (or R) fathers became D’s (or R’s) themselves (ANES, 1952-1992, variables VCF0303 , VCF0306, VCF0106). Blacks with D fathers were likely to stay D–90%; those few Blacks with R fathers were likely, at 65%, to become D’s. But this changed for the Whites over time. In the 1950s, 80% of White respondents with D fathers were D’s and 15% had defected to become R’s; from 1980 to 1992 (after which the ANES stopped asking about fathers’ parties), 65% of White respondents with D fathers were D’s and 26% had defected to R.
- Further ANES analysis: In the South from 1950 to 1969 11% of D children defected to the R party (and 32% of the relatively few R children became D adults). From 1970 to 1992, 20% of D children defected to the R party (and only 22% of R children became D adults). Outside the South, there was less D party loyalty to start with but it did not change much between 1950-69 and 1970-92 (from 20% to 24% D defection; from 23% to 22% R defection).
- See also studies one, two, three, four, five, six, and this earlier post.
- See one, two, three.
- Calculated from the General Social Survey (GSS), “ATTEND.” In the 1970s, D’s were about 10% less likely than R’s to attend church “nearly every week” or more; in the 2010s, they were half as likely. On liberals leaving churches, see these earlier posts: 2014, 2018.
- See, for example, one, two, three.
- On the advances women have made, see these prior posts: 2011(a), 2011(b), 2021. On the political ramifications, for example, here and here.
- On the cultural changes, see this post and this Gallup report. On the political responses: one, two, three, four, five.
- Klein discusses some of these structures.
- About 1.5 million viewers among about 260 million adults.
- From 1900 to 1984 (Reagan’s re-election), the average winning margin in a presidential race was 13 points; in the three elections from 1988 to 1996, it was 7 points, and in the six elections since then, 2.5 points (1.6 points if one sets aside Obama’s 2008 Lehman Brothers collapse victory). The recent 2.5 average includes two elections in which the popular vote loser won the electoral college. Source: here.
- Calculated from UCSB, updating 2020 from Wikipedia. Also, a study shows that the more distance R’s and D’s see between themselves and the other party, the more they participate and vote.
- One call of alarm was a collective article in Science in 2020. Americans increasingly express frustration with the two-party system. Various studies show that strong partisans (on both sides) say they would break principles such as free speech or nonviolence “if necessary.” See, e.g., one, two, three, four–but not all studies, five. In one survey, 40% of Americans agreed that “having a strong leader for America is more important than having a democracy.” In another, most D’s and a large majority of R’s agreed that “the only way our country can solve its current problems is by supporting tough leaders who will crack down on those who undermine American values.”
- In the GSS, the percentage of respondents answering that they were “not too happy” rose from a low of about 9% c. 1990 to a high of about 13% in the mid-2010s (my analysis of “HAPPY”). In the Gallup Poll, “not too happy” was at 10% in 1992 and at 14% in 2019. The Gallup Poll also shows reports of “satisfaction with the way things are going in the U.S.” peaking around 2000 and then dropping by two-thirds to 2010, staying about there since. Ron Brownstein on close elections creating stress, here.
- For example, Brave Angels, America in One Room, the Institute for Civility and the Listen First Project. News coverage here.
- See overview cite above. Some of the very many studies are: one, two, three, four, five, six, seven. (A recent survey experiment suggested that accurately informing partisans about the other party on some topics might lower hostility.)
- See studies published in 2003, 2013, 2016, and 2017.
- See this 2021 review: “The boldest claim that RCV advocates make is that it can reduce partisan polarization …. The experiences of other democracies that use RCV (most notably Australia) offers strong support for the moderating, compromise-inducing influence of RCV. But so far, in the United States, we have very little data to back up any claims.” Also: 2018(a), 2018(b); 2023.