Source: Getty

Polarization, Democracy, and Political Violence in the United States: What the Research Says

What can be done about polarization in the United States? Reviewing a decade of research reveals unexpected findings.

Published on September 5, 2023


The United States feels roiled by polarization, and the philanthropic world is seized with debates about what to do. Some scholars claim that Americans are so polarized they are on the brink of civil war. Other polls suggest that voters agree on plenty of policies and that polarization is an illusion. Some philanthropists call for pluralism and civility, while others lean into activism, believing polarization is a byproduct of change toward a more just world. So, is the United States polarized or not? If it is, what is causing the polarization and what are its consequences? Should polarization be solved or tolerated?

This paper is intended to answer these questions. It opens with five facts about polarization in the United States today and what those imply for possible interventions. A literature review follows, organized chronologically to explain the scholarly shift from thinking of polarization as an ideological, policy-based phenomenon to an issue of emotion, as well as the emerging understanding of polarization as both a social phenomenon and a political strategy. This paper is organized as follows.

Part I: Introduction

  • Five Facts About Polarization in the United States
  • What This Understanding Means for Interventions

Part II: The Literature on Polarization

  • First Generation Understanding: Elite Ideological Polarization
    • Polarization Is Policy Difference, and Congress Is the Problem
    • How Was America Polarized?
    • What Caused Elite Polarization?
    • Interventions to Reduce Policy-Based Polarization Among Political Elites
  • Second Generation Understanding: Mass Affective Polarization
    • Polarization Is Emotional Dislike Based on Identity That Affects Regular People
    • How Was America Polarized?
    • What Is Causing Affective Polarization?
    • Interventions to Reduce Affective Polarization
  • Third Generation Understanding: Cracks in the Foundations
    • Reducing Affective Polarization May Not Impact Violent or Antidemocratic Attitudes
    • Antidemocratic Attitudes
    • Political Violence
    • Political Structures Affect Incentives to Polarize

Part III: Conclusion

  • What We (Think We) Know in 2023
    • Ideological Polarization
    • Affective Polarization


Five Facts About Polarization in the United States

  1. American voters are less ideologically polarized than they think they are, and that misperception is greatest for the most politically engaged people. Americans across parties share many policy preferences. There is some overlap even on hot-button issues, such as abortion and guns, and more overlap on how to teach American history.1 It is important not to make too much of this overlap, however. For instance, a majority of Democrats as well as four in ten Republicans support banning high-capacity ammunition magazines and creating a federal database to track gun sales; nearly as many Republicans support banning assault-style weapons. But only 18 percent of Republicans and Republican-leaners feel gun violence is a major problem (versus 73 percent of Democrats and Democratic-leaners). So despite the significant policy overlap, only one side is motivated to put the issue on the agenda.2 Democrats have moved to the left on racial issues and some social issues over the last decade, and Republicans have moved to the right on immigration under Joe Biden’s administration, though there remains overlap on these issues as well.3 In some cases, Republicans appear to be slowly adopting more progressive views on some social issues, resulting in what looks like polarization but is perhaps better characterized as faster moves by the left.4

    However, most partisans hold major misbeliefs about the other party’s preferences that lead them to think there is far less shared policy belief. This perception gap is highest among progressive activists, followed closely by extreme conservatives: in other words, the people who are most involved in civic and political life hold the least accurate views of the other side’s beliefs. Figure 1 shows the gap between what groups of Americans think the other side believes and what they self-report believing, as of 2019.

  2. American politicians are highly ideologically polarized. In other words, they believe in and vote for different sets of policies, with little overlap. This trend has grown in a steady, unpunctuated manner for decades.5 One reason that the most highly politically engaged Americans may misunderstand the other side is that they correctly estimate the extreme ideological polarization among politicians.

    It is easy to assume that polarized voters are selecting more polarized leaders—and that theory may hold true for recent primary elections. However, that is not the main story. The process begins long before voters get a choice: more ideologically extreme politicians have been running for office since the 1980s.6 Among the pool of people wishing to run, party chairs more often select and support extreme candidates, especially on the right. (In 2013, Republican party chairs at the county level selected ten extreme candidates for every one moderate; the ratio was two to one for Democrats.) The increase in “safe” seats, in which one party is overwhelmingly likely to win, explains candidate and party preferences for more polarizing platforms, but it does not explain the depth of the Republican preference.7

    Parties and candidates clearly believe that more polarizing candidates are more likely to win elections. This may be a self-fulfilling prophecy: voters exposed to more polarizing rhetoric from leaders who share their partisan identity are likely to alter their preferences based on their understanding of what their group believes and has normalized—particularly among primary voters whose identity is more tied to their party. 8 However, only about 20 percent of each party votes in primaries, and 41 percent of Americans are independents who may not have strong party identity and are barred from voting in some states’ primaries.9 That leaves the majority of voters with a relatively low ability to pick a less polarizing candidate of their party. Philanthropists and prodemocracy organizations attempting to reduce polarization often assume that the problem they must grapple with is polarized voters, but their interventions should also take into account the fact that that some of the ideological extremism and polarization since the 1980s is candidate- and party-driven. While at this point, candidates and parties may be responding to polarized primary voters, candidates and parties have been driving the polarization, and not all voters are ideologically polarized.

    The disparity between where leaders are ideologically and where their voters are precludes legislative policy agreement on many issues. Average voters are not able to assert their (often weak) policy preferences because they do not have an effective way to vote out representatives who do not accurately represent their constituents’ views, particularly on the right where party chairs are likely to substitute one extreme candidate for another.

  3. Even though Americans are not as ideologically polarized as they believe themselves to be, they are emotionally polarized (known as “affective polarization”). In other words, they do not like members of the other party. Americans harbor strong dislike for members of the other party (though they also dislike their own parties, as well).10 While social media is often blamed for this phenomenon, affective polarization started growing before the internet: its onset more closely correlates with the rise of cable news and radio talk shows.11 It is also growing most swiftly among Americans over sixty-five years old, a demographic that uses the internet less, but watches television and listens to talk radio far more, than younger age groups who are less polarized.12 These findings and other studies about the effects of social media suggest that all media, not just social media, may be playing a role.

    Many studies have considered what could reduce these feelings of dislike for the other party. They have found that affective polarization appears to be driven largely by an individual’s misbeliefs about the policy beliefs of the other party, a sense that members of the other party dislike members of their party, a fear that the other party supports breaking democratic norms, and misunderstandings about the demographic composition of the other party (for example, believing that Republicans are old, wealthy, evangelical Christians, while Democrats are young, unionized minorities, when the median demographic characteristics of both parties are actually quite similar).13

    Correcting these misperceptions can help reduce affective polarization. Studies have found that telling people in a believable way that they share policy beliefs and similar demographics and creating a sense that there is a shared identity (though the latter is complicated for minorities who prefer dual identities) are interventions that can reduce affective polarization.14 Often, bringing people together across difference is used to accomplish these ends, and this contact between groups may reduce affective polarization.15

    There is a problem with engaging in such interventions to improve U.S. democracy, however. While affective polarization is growing quickly in the United States, it is actually not much higher there than in many European countries. In other words, many European countries show affective polarization at about the same level as that of the United States, but their democracies are not suffering as much, suggesting that something about the U.S. political system, media, campaigns, or social fabric is allowing Americans’ level of emotional polarization to be particularly harmful to U.S. democracy.16

  4. Affective polarization is unlikely to be causing democratic backsliding or political violence on its own. The problem is not polarized emotions alone but how those feelings interact with voting systems, candidate incentives, and personal relationships.

    A series of studies found that it was possible to reduce affective polarization in lab tests, though these effects were only tested after short periods. (The most successful intervention managed to reduce animosity to its 1980s level, but the effects crumbled after a week.)17 But even successful lab interventions had no effect on attitudes toward political violence and other antidemocratic behaviors. For example, experiments that reduced affective polarization among participants found no marked changes in respondents’ willingness to support their party if a candidate employed antidemocratic means to suppress the vote, redraw districts “even if it may be technically illegal,” suppress a protest by the other side, or engage in corruption.18 Reducing affective polarization also did not affect survey respondents’ loyalty to their members of Congress (who may have engaged in some of these behaviors) or increase their desire for bipartisanship. In other words, affective polarization levels describe how people feel interpersonally about the other party’s members. But those feelings do not directly influence how they may behave—they may feel more warmly toward another party but are still likely to vote the same way, support the same policy positions, and have the same levels of disrespect for democratic norms, among other behaviors.

    These studies are confounding: how can emotions change without people altering their attitudes or actions? It is more understandable if one considers a typical effort to reduce affective polarization, such as bringing a group of Democrats together with Republicans over a dinner, or a series of dinners, full of richly moderated discussions and interpersonal insights. That experience may lead the individuals involved to understand each other better and feel more warmly toward the entire groups each side represents. But even if their emotions have shifted by 20 percent—a huge effect—their voting cannot alter by 20 percent because there are only two parties to choose from. So for feelings to affect votes in a two-party system, the dinner guests must engage in a major shift of identity.

    Alternatively, they may feel more warmly toward the other party but vote as they did before, leaving them with some cognitive dissonance. Most are likely to take the latter path and resolve the dissonance by reducing their warm feelings over time. Similarly, after the dinner has deepened understanding and positive feelings, most people will return to their existing religious congregations, friendships, workplaces, and other institutional affiliations. If they have really had a deep moment of change, they may speak up in these settings for their newfound appreciation. But if they meet resistance after sharing their new emotions, they are likely to retreat back into polarization over time. The lab experiments (and findings from real-world examples such as decades of attempts to improve relationships between Israelis and Palestinians) suggest that short-term work to change individual feelings will decay over time in the face of these broader pressures unless immense care is taken to build deep, long-term friendships or create another set of relationships or institutions for people to spend time in that will allow them to act on their altered feelings.19

    In other words, it is not that affective polarization is unimportant: antidemocratic behavior is unlikely to change unless underlying emotions among the voting public change. But it is as if emotions are on a hypothetical dial that can be shifted by degrees, while most behavior is on a dial that has only a few preset settings. For emotional change to yield lasting impacts, people must have more behavioral options—groups of differently minded friends to shift to (which is particularly hard in sparsely populated rural areas), a broader set of political choices, and other options in their social and political spaces.

  5. Similarly, affective polarization is not causing political violence directly. It is probably contributing to an environment that allows politicians and opinion leaders to increase violence targeted at politicians, election officials, women, and many types of minorities. Affective polarization in the United States has been rising for decades, while political violence only increased sharply in 2016.20 Affective polarization is also quite symmetrical across parties, while political violence is overwhelmingly from the right.21 This suggests that emotional feelings of hatred toward members of the other party are not a primary cause of political violence. In fact, some scholars have found that affective polarization is not even correlated with political violence or justifications of such violence (though other surveys have found correlation between “strong Republicans” or “Make America Great Again (MAGA) Republicans” and support for violence, while the same does not hold true for “strong Democrats”).22

    It certainly feels as if the vitriol and violence-mongering in U.S. politics has something to do with the recent increase in political violence. How can this discrepancy be explained?

    People who commit political violence have aggressive personalities.23 Many who commit forms of political violence that do not include preplanning—such as most hate crimes, threats, and violence at rallies—also have poor self-control or have lowered their self-control temporarily (by drinking, for instance).24 These traits are shared across individuals who commit all forms of violence. What is unique about political violence is that it does not arise from interpersonal friction. Instead, for people with low self-control (a large pool that includes, for instance, teenage boys and anyone who has drunk in excess) and aggressive personalities (which limits that pool somewhat) to turn to violence, they need to be enraged and have that anger directed at a group of people they don’t know. They also need to believe that they will not face severe consequences or not care about consequences (because they are too impulsive to care or because they think the consequences are worth it).

    In moments of low political polarization, aggressive people whose impulse control is low are likely to focus most of their rage on personal interactions: domestic violence, road rage, violent crime, or even a school or workplace shooting. Some may commit hate crimes if they are prejudiced and come across a member of a minority group. Those experiencing psychosis or using particular drugs that create illusions of grandeur may direct anger toward political figures, but they are more likely to focus that desire for a sense of importance on a nearer target, such as their school or workplace. Only an insignificant number will act; most pose greater harm to themselves. In other words, in a low-polarization environment, aggressive, impulsive individuals who don’t fear consequences may commit violence, but the targets will largely be apolitical. Meanwhile, violence is likely to be lower because people are facing only self-generated rage, not an additional level of anger and grievance projected toward people outside their personal interactions.

    However, as partisan leaders and media personalities demonize the other party, they can create feelings of rage among followers who fear the consequences of the other party’s perceived actions. Dehumanizing and denigrating rhetoric that normalizes violence or threats against some groups turns that sense of fear and anger into a target by making certain groups appear to be both threatening and, at the same time, vulnerable.25 Finally, the normalization of violence by political leaders, in particular, may provide a sense that acting violently against those groups will be permitted, may not be punished, or could be lauded and turn one into a hero (such as how Kyle Rittenhouse was supported monetarily and publicly embraced after he traveled to Wisconsin to offer “protection” from a Black Lives Matter protest and shot and killed two people).26

    Thus, the individuals committing political violence may not be very polarized themselves; they may even be fairly apolitical. But in seeking to connect to and belong within a political community, they may find leaders who make violence seem normal or even laudable, build followers’ rage, and suggest a target for that anger in a political figure, government official, or minority scapegoat. Affective polarization within society gives them a way to cloak violent impulses in a greater cause that may even allow them to imagine themselves socially embraced or heroized (a potent draw for individuals who may feel ostracized because of their irritable, aggrieved personalities). And hints that political leaders accept such violence reduces concerns about the consequences of their violence, which might otherwise stop them from taking part in a rally or other event where their aggression could manifest.

    As political leaders gin up anger and reduce the sense of consequences, and as affective polarization creates a sense of community and belonging for aggressive, more authoritarian personalities, all types of targeted violence are increasing. Not only are American politicians (from school board members to representatives in Congress) receiving more threats, but also, threats against judges are up, hate crimes are at the highest recorded point in the twenty-first century, and mass shootings are spiking, with perpetrators adopting some political rhetoric into their manifestos or targeting scapegoated groups.27

    Conversely, if Americans had very high levels of affective polarization, but politicians, media leaders, and local leaders maintained a stance of equanimity and civility toward the other party, political violence would probably not increase much outside of particularly virulent subcommunities. This is what occurred in the 1980s when violence rose within the neo-Nazi skinhead movement and in the 1990s when violence grew within the militia movement; both movements were viewed as unacceptable by the vast majority of partisan leaders so violence did not spill into wider society. Affective polarization probably requires alterations to leader rhetoric, norms of violence acceptability within a community, and the perceived likelihood of punishment to increase violence across American society. These other variables may be amenable to faster alteration than levels of affective polarization across a society.

    Understanding the key role of political and opinion leaders in manipulating affective polarization has led to a newer way of looking at polarization that is just starting to spread to U.S. democracy experts but is more prevalent among scholars who compare different political systems and scholars who study civil wars. They consider the structure of politics, looking at whether a variety of identities have collapsed into just two opposing groups that see each other as threatening and even immoral. And they look at whether such polarization can be instrumentalized by politicians as a winning political strategy. Jennifer McCoy, who has pioneered looking at the ways in which emotions and political strategies intersect, finds that the United States has been more “perniciously polarized” for longer than any consolidated democracy has ever been.28

What This Understanding Means for Interventions

These findings suggest that efforts to tackle polarization must be carefully calibrated to have an effect on what is usually described as their ultimate goal: a resilient democracy with strengthened guardrails and less political violence.

  1. Pluralism programs to bridge differences are not enough to improve democracy. Instead, interventions to change individuals’ emotions must be paired with efforts that alter how politicians are incentivized to instrumentalize or amplify affective polarization as a strategy.

    Programs focused solely on engaging individuals from different groups to talk about their differences are based on intergroup contact theory. That literature suggests that these programs can, if well structured, successfully change the feelings of the people who participate (though they can create backlash and harden prejudice if poorly structured).29

    However, just as affective polarization can be reduced in lab settings without impacting views and behaviors regarding democracy and political violence, talking across difference can reduce individual affective polarization—but there is no evidence that it impacts those individuals’ antidemocratic behaviors, preferences for antidemocratic candidates, or support for political violence. Nor is there any evidence that changing feelings on an individual level will aggregate into democratic improvements.

    A literature review of peacebuilding programs in other countries concluded that “programming that focuses on change at the Individual/Personal level, but that never links or translates into action at the Socio-Political level has no discernible effect on peace.”30 Lab research suggests that similar results will hold in the United States. As Mina Cikara, a psychology professor at Harvard University, has explained, “we frequently see equivalent degrees of out-party dislike on both sides, but there’s only one party seeking to curb voting access and throw out election results.”31 Efforts to talk across divides and reduce emotional dislike can improve understanding and reduce prejudice, but these interventions will not result in social or political improvements unless the trusting relationships that are created eventually lead to collaboration that alters broader social norms or political incentives.32

  2. Interventions should aim to reduce feelings of threat, not just feelings of dislike. The most effective interventions for reducing both affective polarization and antidemocratic attitudes involve reducing fears that the other side is intent on breaking democratic norms. A number of studies have found that partisans overestimate the willingness of the other side to break democratic norms and that overestimation makes them more likely to support or even take antidemocratic practices for their side.33 The same holds for political violence.34 The only interventions that currently appear to have a valuable effect on antidemocratic attitudes as well as affective polarization focus on correcting the particular misbeliefs about the other side’s willingness to break democratic norms.35

    For instance, lab experiments in which participants watched a video of Utah’s gubernatorial candidates discussing how each would honor the results of the 2020 election reduced participants’ support for their own side taking antidemocratic action. Correcting misperceptions by simply presenting the data reduces support for antidemocratic action, but only marginally because people may find the data unbelievable. When best practices are used to increase believability—such as showing people talking—they can significantly reduce support for antidemocratic action, but the effects do not seem to last for long, given the onslaught of contrary news in real-life contexts.36

    Nevertheless, it is worth focusing less on the emotions of dislike that Americans may have for one another and focusing more on perceptions of threat—particularly specific fears that the other party will undermine democracy to gain power. Quelling these fears could reduce the arms race.

    Unfortunately, much prodemocracy programming enhances fear that the other side poses an existential threat to democracy. The attempt to use fear to get voters to pay attention to serious threats to democracy is understandable, particularly raising alarms in certain states or about certain politicians given the degree to which the Republican Party is being taken over by an antidemocratic faction. However, the broad sweep of fear may encourage people to vote while also building support for antidemocratic behavior. This is a real problem the prodemocracy community must consider seriously, possibly by experimenting with more positive, aspirational mobilizing strategies rather than relying on threats. The effects on younger voters, who are already less attached to the democratic system than other demographics, may be particularly harmful over time.37

    At the same time, misinformation, disinformation, and malinformation regarding Democratic election malfeasance are creating the same real sense of threat and desire for undemocratic action to save democracy on the Republican side.

    This arms race is creating a self-fulfilling prophecy. This is the core issue of the affective polarization problem. But as the most recent understanding of polarization suggests, changing feelings alone is not enough. Altering the sense of threat also requires changing political incentives that allow politicians to use hatred, othering, and claims that the other side is immoral as a winning political strategy.

  3. Win-lose-style, adversarial advocacy on the right-left dimension might be dangerous. Based on the finding above, advocacy that amplifies the belief that members of the other party are bent on destroying democracy itself is likely to deepen polarization and support for antidemocratic action on one’s own side. (So might a similar style of advocacy on issues adjacent to democracy, such as reproductive rights on the left or gun rights on the right, that campaigns by hardening beliefs that the other side is going to do away with democracy to solidify their control over the policy). The more each side fears the other is going to subvert the rules of the game, the more willing voters seem to be to do it first to lock in their party’s advantage.

    Meanwhile, highly partisan ideological advocacy may be subverting democracy indirectly. Highly adversarial advocacy in which partisans on one side of an issue are amassed, angered, and organized to change a policy along strict partisan lines is very likely to fail, given the intense ideological polarization of U.S. leaders and the overuse of the filibuster such that supermajorities are required for legislation to pass. Frequent failure to achieve reforms may be hardening misperceptions among voters about the average voter in the other party. Advocacy efforts that organize people toward such failed efforts can leave people more cynical. Meanwhile, levels of affective polarization are such that highly publicized, adversarial advocacy attempts are likely to generate backlash. Thus, this form of adversarial advocacy is likely to achieve few victories at the national level (though it may be able to achieve state-level change in reliably red or blue states). In either case, such success will generate backlash in other states or at the national level, creating Pyrrhic victories that are quickly overturned at the national level or are met with opposing legislation in other states.

    The United States may reach a point where the best it can hope for is to shore up democracy and inclusion in some states while abandoning others to a less democratic and inclusive future. This is, after all, what occurred under Jim Crow for eight decades. But America is not at this point yet, and an advocacy style that pushes the country closer to such a dystopic fate is in no way helpful to the goals of social justice. Instead, intense mobilization efforts built around a positive future vision are needed to galvanize voters without exacerbating antidemocratic attitudes.

  4. A different form of pluralistic work that coordinates groups to act on shared goals, despite other differences of opinion, may work to bridge partisanship, build trust, and advance a more just democracy. Even in today’s ideologically polarized Congress, progress has been made toward a more inclusive and stronger democracy using a strategy of coordinated action across difference. In 2022, with current levels of affective polarization, this type of programming led to success in reforming the Electoral Count Act, a crucial node for shoring up the U.S. democracy’s guardrails. It achieved the country’s only lasting, national modern criminal justice reform in 2018 and enshrined a national legal right to gay marriage in 2022. Advocacy framed to welcome Republicans as well as Democrats has allowed reproductive rights to gain ground in multiple states since the Supreme Court overturned Roe vs. Wade. Such programming has also been the bedrock of effective peacebuilding efforts in countries such as Colombia and Sierra Leone.

    Programming to bridge differences, not simply to talk but to find common ground among unlikely allies in order to act on shared problems, has another benefit: it may build trust and a sense of agency. A majority of the voting public are disgusted by politics in general, which they feel is corrupt and rigged against people like themselves.38 While they report feelings of affective polarization against the other party, 40 percent of Republicans also did not like their own political side until the MAGA movement.39 The affective polarization conversation misses the reality that a portion of angry, low-trust Americans do not simply dislike the other party but distrust nearly every institution in American life: big business, schools, newspapers, television news, Congress, the criminal justice system, and organized religion, among others.40 In reality, they are polarized from a political and economic system that feels separate (hence “elite”) and insensitive to their needs. While polling geared toward affective polarization has found them disgusted with the other party, they in fact feel frustrated and hopeless about the entire U.S. political and economic system in general. Instead of focusing on polarization, the alienation they feel needs to be addressed by enabling agency around problems they—and the people they are often pitted against in more simplistic media accounts—both want solved. These issues may be highly local, for instance, building a community’s resilience against more frequent extreme weather. Or they may entail major national mobilization, such as bringing together conservative religious groups and gay activists to pass the Respect for Marriage Act, which enshrined gay marriage as a national law just as it was under threat from Supreme Court rulings. Understanding which problems are shared and solvable cannot be guessed beforehand: it requires discussion and trust-building.

    Empirical evidence from overseas and from the United States has found that assembling groups across difference to talk can make a significant difference in a highly polarized atmosphere, if the goal is to uncover shared elements of a common agenda on which groups can act.41 In these cases, talking is not an end in itself but is used to build trust so that groups can work on concrete advocacy that moves countries toward less-polarized politics. Such advocacy can be highly adversarial, but against new lines of polarization rather than the right-left dimension (for example, uniting minority and White working-class Americans for better working conditions or bringing together rural White and minority Americans for recognition of rural needs as opposed to urban assumptions). Ironically, since misbeliefs about the other side’s policy preferences are strongest among highly educated, partisan political activists (and highest on the left), programs that attempt to solve problems by gaining understanding and working with unlikely allies on shared issues of concern may face the greatest pushback by those who claim to be most politically knowledgeable. This may obstruct the ability to build trust and agency among less politically engaged minority and White Americans whose preferences are quite different from those of engaged political elites.

  5. Polarization is a highly nuanced field, and small assumptions can lead to big mistakes. Practitioners and philanthropists should be particularly careful about assumptions regarding moderation. People who poll as moderates may also be antidemocratic or supportive of political violence, especially on the right. On the left, support for democracy may coincide with support for violence.

    Many people think of Americans as arrayed along a straight line, with the far left on one side and the far right on the other. They assume that the people at the edges are the most polarized, the most partisan, hold the most extreme ideological views, and are the most supportive of antidemocratic actions and violence. This is not the case. Consistent conservatives and liberals who are more politically engaged are both more affectively and ideologically polarized and more prodemocracy than those in the middle.42

    A frequent misunderstanding is that people who hold ideological views that contain elements from both sides of the aisle are moderates because they fall in the middle of the left-right spectrum. In fact, a Democracy Fund survey in 2018 found that Americans who held the least polarized ideological beliefs were actually the voting cohort least in favor of democracy and most supportive of a “strong leader” who does not need to bother with Congress or elections.43

    It is a common assumption that people who hold views from both sides of the aisle are economically conservative and socially liberal—the profile of many in the upper-middle-class political elite trying to reduce polarization. In fact, a 2016 study showed that this type of moderate ideology was held by only 3.8 percent of the electorate. Instead, the preponderance of Americans who respond to ideological survey questions with answers on both sides of the aisle (28.9 percent of the electorate) tend to be pro–economic redistribution while also upholding the belief that American citizens should be White, Christian, and born in the United States.44 That mix of views led this group to be swing voters for many years, although since 2016 many have moved more decisively into the Republican Party.

    This group, many of whom support an authoritarian leader and some of whom justify political violence when necessary to maintain traditional gender roles and a White-dominated racial hierarchy, may be highly affectively polarized—but they also may not be, since they tend to distrust all American institutions and elites.45 Some view themselves as strong partisan Republicans, so surveys such as Bright Line Watch found that people who strongly identified as Republican a year after the contested 2020 election were the most supportive of political violence.46 But others feel little partisan affiliation. They have voted for both parties in the past and seek someone to represent their views; if a third party of a Trumpian orientation emerged, they might switch to it. While they were united by their votes for Donald Trump in 2020 and their belief that the 2020 presidential election was stolen, they may be willing to mobilize in the future behind another strong, extreme leader: many of the most radical are no longer willing to turn out violently for Trump rallies because they feel let down by his failure to provide monetary support for those arrested on January 6.47 They are probably best characterized simply as the antidemocratic right, a growing counterculture with high distrust, a low sense of agency, and strong feelings of grievance and victimhood.

    On the left, the most partisan Democrats are actually far more moderate ideologically than progressives. In other words, on the left, extreme partisanship coincides with moderate ideology. The problem for democracy on the left is not those who report high polarization (those individuals tend to be active and politically engaged in a positive way for democracy), but a very small but vocal group that the international anti-polarization organization More in Common terms “activist mavericks.”48 These individuals are younger and wealthier than average Americans and more often male. They are far-left in their ideology, but they dislike the Democratic Party and so are not partisan (though they may be affectively polarized and hate Republicans). They claim to be very strongly prodemocratic. But they do not believe America has achieved democracy, and so they are willing to support political violence to achieve greater racial and democratic representation. Six in ten support property crime (versus 6 percent of Americans overall), and 28 percent support violence against people (versus 4 percent of the population overall).

    These two groups are asymmetric: the antidemocratic right is a plurality of somewhere between a quarter and a third of the Republican Party.49 The proviolence left is tiny and composes an insignificant part of the vote share of Democrats, especially since many may vote for third parties. Both are surrounded, however, by a penumbra of apologists and soft supporters who normalize their behavior. This has allowed the antidemocratic faction of the right to achieve a nearly complete takeover of the Republican Party that is giving it significant political power. Maverick activists on the left hold virtually no political power at any level of government, but their views have achieved outsized cultural sway. Despite their asymmetry, the bogeyman of these two groups is fueling the other and is the main force tearing the country apart—not a more generic or symmetrical polarization.

    Thus, it may be worth testing interventions focused precisely on these groups, as well as supporting independent, high-quality research alongside these programs to track behavioral outcomes over the short, medium, and long term. Or it might be worth testing interventions to build the prodemocracy habits and norms of Americans who are supportive of but not quite within these more virulent subcultures. To be of any use, these interventions and research must be done in real-world conditions, not just quick lab experiments, and studied over time.

The Literature on Polarization

In the 1950s, some influential U.S. scholars argued that America needed more polarization, which they defined as members of different parties holding differing policy beliefs. At the time, each of the two dominant parties encompassed a breadth of often overlapping political views thanks to conservative Southern Democrats and liberal Rockefeller Republicans. A famous American Political Science Review study from 1950 concluded that more polarization would help voters differentiate between the parties.50 These beliefs persisted among some of the most highly regarded scholars in the polarization field, such as the University of Pennsylvania’s Matthew Levendusky, well into the 2000s.51

Yet other experts took into account massive sociopolitical changes that affected how they viewed polarization. First came the Civil Rights Movement. The Civil Rights Act of 1964 sent former Southern Democrats into the Republican Party, accelerating a process of party sorting by ideology that was largely influenced by racial prejudice.52 Subsequent cultural cleavages over women’s rights, gay rights, and the environment, combined with polarization over the Vietnam War and the Watergate scandal, continued the process of deepening ideological differences, while demographic sorting among the electorate put more liberals on the Democratic side and more conservatives on the Republican side of the aisle.53 A few crucial politicians, notably speaker of the U.S. House of Representatives Newt Gingrich, amplified these cleavages during the 1990s through a style of intensely partisan leadership that deepened the effects of these cultural and ideological differences on key institutions—in particular, Congress.54

First Generation Understanding: Elite Ideological Polarization

Polarization Is Policy Difference, Congress Is the Problem

By the early 2000s, members of Congress and people working on what was then called “good government” began to feel that polarization was a problem. They bemoaned a lack of compromise in Congress and a more poisonous atmosphere that appeared to be harming Congress’s ability to govern and undertake formerly uncontentious activities; for instance, the filibuster began to be routinely applied around 2010. They pointed to polarization as a core reason for the malfunctioning of Congress and other democratic institutions.

Political scientists began to look into the issue and brought with them the tools of their discipline, which conceived of polarization as ideological distance between policy preferences. To measure that ideological distance in Congress, they used a dataset called DW-Nominate that looked at how members voted.55 Three seminal studies published between 2006 and 2015, including a task force convened by the American Political Science Association, summarized scholarly opinion.56 They found that:

  1. Polarization had risen in the Senate since the mid-1950s and in the House since the mid-1970s (see figure 2). That finding ruled out event-driven polarization such as president Bill Clinton’s impeachment. It also ruled out developments that happened well-after polarization had become entrenched, such as the rise of the right-wing movement known as the Tea Party, which created its own House caucus.57 In the House, the seven states with a single representative (districts that clearly cannot be gerrymandered) also had quite partisan representatives.58  

  2. Ideological polarization was highly asymmetric. Republican members of Congress were moving to the right ideologically much faster than Democratic members were moving to the left (see figure 3).

  3. Findings on ordinary Americans’ polarization were disputed. Polarization among the public seemed to depend on what scholars were measuring. Pew Research Center found growing antipathy as well as ideological distance among political activists that was bleeding into regular voters, but it also found that a broad center still held.59

    Scholars who looked at self-identification found that conservatives were more likely to identify as Republicans and liberals as Democrats, suggesting that the parties were sorting by ideology much more than in past decades. Self-reporting also showed that ideological moderates had declined, and the self-described distance between mainstream Democrats and Republicans (not their respective activist wings) had doubled between 1972 and 2004.60 In fact, voters seemed to have the opposite asymmetry to their representatives: the share of Democrats who claimed to be liberal doubled between 1994 and 2014, and those who were consistently liberal quadrupled. Republican voters did not experience as much movement to the right or consolidation. Pew Research Center speculated that Democratic voters might be consolidating in reaction to Republican members of Congress moving to the right in a highly public way, such as with the release of Gingrich’s Contract With America.61 Scholars also found that voters were more likely to change their positions to match their party affiliation, rather than switching parties to support candidates who supported their positions.62 These findings suggested that the public was also polarizing over policy.

    However, two of the leading academics in the field of polarization studies, Morris Fiorina of Stanford University and Levendusky, argued strongly that the median voter remained a centrist on most issues.63 They found that Americans held a number of policy beliefs in common despite party affiliation, including on abortion, immigration, and gun rights.64 They also found that a broad group of voters held beliefs from both the left and right of the ideological spectrum and termed these individuals “moderates.” They concluded that Democrats and Republicans who were not political activists held far less unitary views than their own party elites, were more moderate ideologically, and therefore (since polarization was defined as ideological distance), Americans were mostly unpolarized. Their views held sway.

How Was America Polarized?

Scholars drew a number of conclusions from these findings about what was causing polarization and, therefore, what could be done to stop it.

  1. Ideological polarization was an elite issue. Polarization at the congressional level was worrisome but not deadly to democracy. It was causing difficulty with nominations to Congress and with legislation that broke along party lines. But it had not stopped major bipartisan legislation on important issues.
  2. Polarization among the American public was occurring, but it was largely among the politically engaged public, which was not the majority. There still seemed to be a broad middle, and there was disagreement about whether polarization even existed.65

What Caused Elite Polarization?

  1. Competition over power was likely to be playing a significant role. Since the country’s founding, congressional power was usually held by one party for decades at a time. But since 1980, it had been more up for grabs than in any period since the Civil War and Reconstruction, leading to more bare-knuckled politics and uglier campaigns as each election could be decisive for congressional control. The Gilded Age was another period in which control of Congress changed parties frequently because each party was close to 50 percent of the voting public. Frances Lee of Princeton University was the main proponent of this theory.66
  2. One book-length study suggested the problem was that as the personal and financial costs of running for office increased and the benefits of holding office decreased, moderate candidates were disproportionately less likely to run, leaving the field to ideological extremes. The study, looking at all candidates who had run for House seats and not just the winners, found that all candidates had polarized significantly since the 1980s—so that whoever won would be more polarized, regardless of voter preferences.67 Further suggesting that the issue was structural and candidate-driven rather than voter-driven, another study found that in 2013, Republican party chairs at the county level preferred extreme candidates by a ten-to-one margin. Democratic party chairs also preferred extreme candidates, by a two-to-one margin. The safety of seats explains why both chairs preferred strong partisans but not why Republicans leaned so much more heavily in that direction.68
  3. The rise in economic inequality in the United States appeared to be causing congressional ideological polarization—but congressional ideological polarization was also leading to increases in inequality, so causality was a vicious circle.69 Nolan McCarty, Keith T. Poole, and Howard Rosenthal found in 2007 that inequality exacerbated ideological polarization, and ideological polarization led to policies that further increased inequality. In other words, they found that people with vastly differential wealth had different policy preferences. But ideological differences between Republican and Democratic partisans led to the failure of redistributive policies, thus exacerbating inequality.70
  4. Religious or evangelical voters were one of many factors—not the driving factor—behind polarization. Religiously driven political polarization was a concern at the time because of the organization of the Moral Majority and post-1980s evangelical movements.71
  5. Gerrymandering was not the problem (in terms of polarization) that democracy activists believed it to be. Since the Senate was also polarized (in fact, had been polarizing for decades before the House had been), as were members who held impossible-to-gerrymander, single-state House districts, gerrymandering could at best be a modest cause of polarization.72 Voter sorting by geography was the greater cause of what looked like geographic polarization.
  6. The abolishment of the Fairness Doctrine in 1987 may have caused polarization by allowing for greater media partisanship and self-selection into different news silos. When it was in effect, the Fairness Doctrine saw television news as a near monopoly because there were only three major channels, so it required news outlets to provide equal time to different ideological viewpoints; as the number of cable and radio news outlets grew, news appeared to be a more regular private market and president Ronald Reagan’s administration removed this requirement of equal time. That change meant partisans could self-select into media bubbles, while most Americans could opt out of news in favor of entertainment.73

    There were also concerns that the way the media had started to cover the horse race of politics rather than policy issues led people to be more angered about their losses and that incivility in argumentation in order to increase drama and entertainment value encouraged a more black-and-white view of the world.74
  7. Elite polarization among members of Congress was leading to greater partisan sorting of Americans into separate parties based on their policy preferences. Thus, solving the problems in Congress would help whatever polarization existed among the general public.75

Interventions to Reduce Policy-Based Polarization Among Political Elites

Based on this understanding of polarization, what was then best described as the “good government” field coalesced in the 2010s around the idea that the United States faced a problem of policy-based polarization among political elites. Therefore, many people working in the field of improving American governance concluded that the need was to find ways for Congress to compromise, negotiate, and break partisan incentive structures to win at all costs in order to win ultimate power.

  • Some efforts were technical alterations to encourage compromise and working together, such as the House Select Committee on the Modernization of Congress, a truly bipartisan congressional committee, assisted by outside help, that worked from 2019 to 2022 to improve congressional effectiveness and transparency, as well as other efforts to change congressional rules.
  • Others were based on altering political incentives to encourage bipartisan cooperation. For instance, the Democratic House majority proposed bringing back earmarks, which allowed legislation to direct congressionally appropriated funds to specific uses in districts. Earmarks had been implicated in corruption and were banned by the Republican House majority in 2011. Yet they also incentivized compromise by giving members something of real worth to trade—taking a vote that could be difficult ideologically or could risk bringing on attack ads by an opponent in the next campaign could be neutralized if a member could show that they had brought home something desired by their district, such as funding for a new health center or bridge. Moreover, by letting members bring valued goods back to their districts, earmarks could show that the government was functioning for the people.76
  • Still other interventions focused on members’ emotions, such as restoring a greater number of bipartisan congressional trips (CODELS) where members socialized and got to know one another across the aisle.
  • There was a lack of consensus and general skepticism that voting reforms, such as altering primaries, would reduce elite polarization. Fiorina and Levendusky argued that regular Americans were not very polarized. That accorded with findings in many other countries that partisan polarization tended to be an elite problem that had little effect on regular voters. If voters were more moderate than politicians, then electorates were not causing the polarization in Congress and electoral reform would not help.77 Few reformers paid attention to the findings on candidates and how candidate incentives might be influenced by beliefs about campaigning and fundraising.
  • The problem of growing inequality fit the data for when ideological polarization began in both chambers of Congress, but the good government field at the time generally paid little attention to economics.

Second Generation Understanding: Mass Affective Polarization

Polarization Is Emotional Dislike Based on Identity That Affects Regular People

By 2016, experts were struggling to describe a landscape that appeared to have high mass public polarization, despite the fact that many scholars continued to find a great deal of overlap among Americans’ policy views on contentious issues.78

Scholars from new disciplines, such as political psychology and comparative politics, started to formulate a new hypothesis about what was occurring in the United States and other polarized democracies. Working separately and writing in hundreds of papers and books, these scholars crafted what became known as the idea of affective polarization:emotional dislike and disgust between members of opposing parties based not on policies but on identity.79 Partisans who were affectively polarized might actually overlap in their policy views (without knowing it due to misperceptions), or they might differ; policy alignment was not a good measure of emotional feelings.80 As political psychologist Lilliana Mason argued, “identity-based ideology can drive affective ideological polarization even when individuals are naïve about policy. The passion and prejudice with which we approach politics is driven not only by what we think, but also powerfully by who we think we are.”81 In some cases, differences in policy beliefs even appeared to shrink as affective polarization increased.82 Using measures based not on what policies people supported but on feeling thermometers ranking “warmth” or “coolness” toward the other party, trust and trait measures (such as asking if the other party was more hypocritical or more trustworthy), and social distance measures (such as “would you mind if your child married a member of the opposite party” and spousal agreement on party),83 scholars attempted to understand the public’s emotional state.

Scholars claimed that affective polarization was more destabilizing than policy polarization because it wove itself into a country’s social fabric. Different studies drew implications for the economy,84 health,85 and democracy itself. In particular, affective polarization seemed to allow one’s own political side to get away with undemocratic behavior to prevent the other side from winning, enabling the election of leaders who harmed democracy with the support of their base voters.

  • The international peace-building organization Beyond Conflict found that Republicans and Democrats who overestimated the degree of dehumanization, dislike, and disagreement between parties were more likely to support putting their political party over the general good of the country. They also found lower levels of trust in the country’s civic institutions, such as local and state governments or the Supreme Court, among those who overestimated the affective polarization of the other side.86
  • A survey of American voters found that they were less willing to support same-day voter registration if they were told that it would primarily benefit the other party.87
  • Studies by Milan Svolik and coauthors found in lab experiments in Türkiye, the United States, and Venezuela that the most polarized voters were the most willing to turn a blind eye to antidemocratic behavior committed by their party in order to prevent opponents from gaining control because they feared subversion by the other side.88 Actual voting data showed that Americans were willing to trade democratic principles for partisan loyalty or policy preferences and that only a small portion of Americans would prioritize democracy when these goals came into conflict.89 An online experiment by Alia Braley and coauthors with nearly 2,000 Americans found that voters were particularly willing to allow democratic norm-breaking by their side if they were credibly concerned that the other side wanted to break the rules—similar to the security dilemmas faced by countries locked in potential future conflict, where the incentive is to attack first if one fears being attacked.90

How Was America Polarized?

  1. The American public is affectively polarized today, and affective polarization has been increasing steadily since the 1980s and possibly earlier, depending on the measurement used. (For one such measurement, see figure 4, based on data modeled by Massachusetts Institute of Technology election expert Charles Stewart III.) Since at least the mid-1990s, affective polarization had been growing for all age groups, but it had grown most quickly among people over sixty-five.91
  2. Affective polarization is being driven primarily by dislike of the other party rather than positive feelings toward one’s own, since positive feelings toward one’s own party has also dropped slightly since the late 1980s (see upper lines in figure 4). Republicans reported liking their own party more in 2016, in correlation with the MAGA movement. Affective polarization, however, is quite symmetric across the parties.92
  3. Affective polarization appears to be largely driven by misperceptions. Misperceptions about the actual demographic identities of the other party, the policy beliefs held by members of the other party, how much members of the other party disliked and disagreed with the respondents, and whether members of the other party supported democratic norm-breaking are all driving the problem.93
  4. While the public feels a great deal of emotional dislike and polarization between the parties, ideological divergence in the mass public has not increased greatly over time.94 It is increasing somewhat: liberals are becoming more liberal on social issues and Democrats have moved to the left on racial issues since 2010, while Republicans have largely remained in the same place. Meanwhile, under the Biden administration, Republicans have started to move right on immigration.95 Yet most scholarship continues to agree with earlier findings that there is significant ideological overlap and agreement on policies, including on contentious issues such as guns and abortion, in part because there is a lot of ideological inconsistency in the parties—particularly with the move of a group of conservative but economically redistributive former swing voters into the Republican Party in 2016.96 However, it is worth noting that most Americans do not hold very intense policy beliefs; for many years, scholars have found that American voters alter their policy preferences to match their partisan identities. In contrast to Congress, which appears to be ideologically polarized, the problem for the public seems to be emotional rather than ideological—though some of the emotional polarization is based on real and perceived ideological differences among the activist edges of each party.
  5. Affective polarization has been rising much more rapidly in the United States than in most other Western countries (see figure 5). Attitudes across countries are notoriously hard to measure and compare; different systems of evaluation have yielded different measurements even when using the same data. However, most scholars agreed that America’s overall level of affective polarization, though rising, was only middling when compared with other countries in the Organisation for Economic Co-operation and Development (OECD). A number of peer democracies with multiparty systems had higher levels of affective polarization but were showing much less democratic strain.97 That finding suggests that if affective polarization is harming U.S. democracy, it is not acting alone but is being exacerbated by other factors, such as the structure of the U.S. political system or the media environment.

    Trends in affective polarization also look very different across OECD democracies—rising in some, falling in others, and sometimes staying the same. That suggests that the United States’ swift rise in affective polarization has a domestic cause rather than being caused by an internationally systemic issue.

    Scholars such as McCoy and Murat Somer have very different findings based on a different way of looking at polarization. Their measurement of “pernicious polarization” not only includes affective polarization but also considers whether the process of polarization has collapsed a variety of identities into two opposing camps that view each other with so much distaste, othering, and aversion that each group sees the other as immoral. It further considers whether politicians are using polarization as an instrumental strategy. Viewing polarization as a process rather than just an amount of dislike provides a way of measuring how hardened polarization is likely to be and whether there are negative feedback loops between public emotions and political incentives that create a fundamentally different environment than can be gauged by a single snapshot of mass feeling. Using this definition and measures based on expert opinion rather than self-reported feelings, their research finds that the United States is polarized at extraordinary rates compared to other countries and that no other wealthy, consolidated democracy has been as perniciously polarized for as long as the United States.98 This observation may be incorrect, as the expert data is subject to a presentism bias (since the measure was created recently and coded for past years). But it may also be more accurate than self-reported feeling thermometers because different countries report their feelings differently (that is, what feels like intense polarization to a Swede might feel quite mild to an American) and because their underlying definition of polarization is a better measure of antidemocratic attitudes than affective polarization alone.

What Has Been Causing Affective Polarization?

Theories of what has been driving affective polarization draw from psychology, sociology, comparative democracy, and political science. Each discipline sees the problem through the lenses it brings to the world. The problem could have many causes—there is no reason to believe it is only one. There is also no reason to believe that the same cause affects both parties equally—they could each be affectively polarized for different reasons. A number of theories that are common among the public are not supported by research. Eight major theories or explanatory narratives have emerged in recent years.

  1. Stacked and sorted identities. People’s brains are wired to form in-groups for safety and then use shortcuts based on shared values, stories, symbols, and language to decide who is in and who is out. It is possible to prefer one’s own group without feeling much antagonism for an out-group (known as “in-group affinity”). But human brains are constantly scanning for threats to in-groups. As people affectively polarize, they appear to blow out-group threats out of proportion, exaggerating the out-group’s dislike and disgust for their own group and getting ready to defend their in-group, sometimes aggressively. Once polarization starts, it can become a vicious cycle because the human tendency toward motivated cognition seeks out information that proves these exaggerated beliefs and discounts information inconsistent with those beliefs.99

    Mason and other scholars claim that polarization entered this cycle in the United States because of increasingly well-sorted parties. Americans have sorted themselves into parties based on their ideological affinities (meaning there are fewer liberal Republicans and conservative Democrats) at the same time that other racial, religious, sexual, geographic, and cultural identities have been increasingly aligned with partisanship.100 This “stacking” of identities within each party has allowed partisanship to become a social identity, not just a marker of one’s policy preferences.101 People with stacked identities are quicker to take offense when any of their identities are perceived to be maligned. So these in- and out-group dynamics are operating at high sensitivity, in which a threat against any one of many stacked identities can be perceived as an attack against one’s group.102

  2. Geographic and social sorting. Other scholars point to people sorting into more homogenous groups of friends and geographic sorting,103 two factors that are leading people to socialize more with people they already agree with.104 Partisans may have incidentally self-sorted into different geographies because of preferences for urban or rural living or other factors. After that, they socialize with copartisans and their beliefs deepen as a result of that social homogeneity.

    This is not the same as self-sorting purposefully with the primary goal of living among copartisans, for which there are anecdotal claims but evidence is unclear. Mobility data and voter records suggest that people are moving because of other issues, such as jobs or the quality of public schools.105 However, these could be linked to partisan identity and ideology (that is, what one views as a quality school may differ based on ideology, or one could look for a job in another state because the laws in one’s current state feel threatening).

  3. Media bubbles. Another common claim is that Americans are sorting themselves into media echo chambers whose lack of balanced content further radicalizes and polarizes them.106 Social media is particularly singled out for its negative effects because its algorithms and business models have been shown to exacerbate outrage and anger.107 Social media has also been found to help recruit and provide platforms to extremists.108

    However, scholarship is extremely mixed regarding the effects of social media.109 Four recent landmark studies by dozens of researchers conducting tests on tens of thousands of Facebook and Instagram accounts found that social media algorithms did not affect users’ polarization.110 While scholarship points in multiple directions, those findings support the preponderance of empirical studies that suggest, insofar as social media may be exacerbating affective polarization, the problem is small. And, the problem is not social media alone but the country’s entire media system. A recent literature review into media and social media in the United States found that media in general was likely increasing ideological and affective polarization. (It also noted significant methodological issues in the literature, particularly an overreliance on Twitter to measure social media usage, which is a problem because Twitter’s user demographics are not those of the median voter).111

    Most studies do not find much correlation between internet news or social media use and affective polarization. That may be because most Americans do not use social media to discuss politics; politics is a very incidental part of life for most Americans, who largely avoid it.112 Scholars find that internet usage has increased similarly across Western Europe and in OECD countries, but affective polarization has skyrocketed in the United States while it has shown no consistent trend across OECD countries. Another cross-country study found that online news is slightly negatively correlated with affective polarization across countries—Americans get far less news online than Norwegians or Swedes, for instance, but affective polarization is rising in the United States while declining in the two other countries.113 Other studies argue that the internet only became a significant part of news in the 2000s, yet affective polarization has been growing since at least the 1980s and correlates more closely with the rise of cable news.114 (However, affective polarization from Democrats seems to have experienced a sharper rise since the early 2000s; see figure 6.) A large-scale experiment in which individuals deactivated their Facebook accounts for four weeks prior to the 2018 midterm elections did find that the deactivation reduced affective polarization, but only slightly.115

    Meanwhile, affective polarization is growing most quickly among Americans over sixty-five years old. Yet younger Americans use more social media than older adults, while those over sixty-five spend more than six and a half hours a day in front of their televisions.116

    Social media remains a problem for polarization, but in context. Multiple studies show that most Americans are not seeking out political news on social media, and so the loose connections they get through social media actually tend to bring them into connection with less polarizing news, rather than amplifying media bubbles. However, for the small segment of the public that is highly engaged with politics, social media is probably polarizing.117 Since nearly everyone running or funding efforts to reduce polarization is, by definition, highly engaged with politics, this group may overstate the effects of social media versus the entire media environment. Meanwhile, for violent individuals, particular sites are used for organizing, though the degree of radicalization occurring through those sites is unclear.

    In the United States, surveys (such as one by Public Religion Research Institute) find more intense and consistent negative democratic effects from far-right and right-wing cable news and radio companies and channels, such as Newsmax, One America News, and Fox News, rather than social media.118 Exposure to these offline forms of partisan news makes those with extreme attitudes even more extreme.119 Only one-tenth of insurrectionists arrested after the January 6 attack on the U.S. Capitol received most of their news from social media; most favored conservative television and radio.120 An experiment in which people were incentivized to shift from watching Fox News to CNN showed that further exposure to partisan outlets, such as Fox News, was radicalizing and that incentivizing a shift from Fox to CNN led to more accurate and less polarized information.121

    But causation is unclear: people who are more polarized might also be more motivated to watch partisan news. After Fox News made an early call during the November 2020 election that the majority of Arizonans had voted for Biden for president, 37 percent of the network’s prime-time viewers chose to move their viewership temporarily to more extreme outlets, such as Newsmax and One America News, that reported for months that a Trump victory was possible (or had actually occurred). Viewers were shaping the content, rather than being led by it.122 Similarly, after Fox News fired its most popular host, the right-wing personality Tucker Carlson, viewership plummeted—while viewership increased during the same period at Newsmax.123 More polarized people may also be more motivated to engage with polarizing social media. Studies of YouTube showed that most engagement with extremist content was largely confined to a small, concentrated group of people who had preexisting, negative views on gender and race.124 And while partisan traditional news sources do seem to exacerbate polarization in America, in Europe there is little evidence to suggest that exposure to partisan or populist traditional media causes widespread polarization.125

    It could be that social media and partisan traditional media are having more polarizing effects in America than elsewhere because of other factors. For instance, having a trusted media source—whether a national source of record or a strong local news source—might mitigate the polarizing effects of social and partisan media. In the United States, no single news source is trusted by more than 50 percent of Americans, while in the United Kingdom, for example, 86 percent of Britons say that they are satisfied with their state broadcasting service.126 Countries that spend more on public broadcasting have seen greater decreases in polarization, though the direction of causality is unclear.127 In the United States, it is possible that local news played the mediating, trusted role that a government broadcaster plays in some other democracies and that the decline of local news is enhancing affective polarization. One study found that after communities lost a local newspaper, voters were less likely to split their ticket in federal elections.128 The decline of local news has also been linked to lower voter turnout and citizen engagement in general.129

  4. Inequality and economics. While ideological polarization in Congress is correlated with inequality, inequality’s relationship with affective polarization is much less clear—in part because there are very few studies on it.

    • One multicountry study found that high inequality and unemployment were correlated with greater affective polarization; however, another study using the same data but a different evaluation method found that the role of inequality no longer held.130
    • Other studies looking across countries suggest that inequality is poorly correlated with affective polarization. For instance, affective polarization has been falling in Australia at the same time as inequality has been rising, while affective polarization is rising quickly in Denmark while income inequality is quite low. In the Netherlands, affective polarization is growing relatively swiftly while income inequality is also low (though wealth inequality is higher).131
    • In Belgium, affective polarization is distributed quite differently in different parts of the country. One study found that differences in economic performance over the past few years drove affective polarization, particularly if wealthier areas were located near poorer places. But the rural/urban divide also still mattered, separate from these dynamics.132
    • One theoretical model suggests that inequality could contribute to polarization by encouraging people who are more risk averse due to a harder economic environment to limit their interactions to in-group members.133 Low levels of social trust are associated with these people being more risk averse and retreating to doing business within their in-group.134 But if inequality is having these effects, its role needs to be teased out from other factors that could be lowering social trust.
  5. Economic shocks. Some research suggests that calamitous economic shocks—such as the Great Depression and the 2008–2009 financial crisis—increase support for populist parties that denigrate social out-groups and thus cause affective polarization.135 However, globally, populist parties began winning power in 2000, prior to the 2008–2009 financial crisis, though the crisis may have increased their success: populist leaders hit their modern peak in 2018. If economic shocks are exacerbating affective polarization by leading voters to elect populist candidates, then inequality is mediated through populist political leaders who interpret economic realities and is not a direct cause.

    • Michael Podhorzer’s analysis of majority-White congressional districts showed that after the 2008–2009 financial crisis, the majority-White districts that voted Republican (prior to or after the shock) had a much slower economic recovery than those that voted for Democrats (see figure 7).136

    These trends in recovery strongly correlate with rural-versus-urban geographies (see figure 8).137

    Affective polarization among the American public had started long before 2008. But these trends—given direction by a concerted effort at political organizing by deep-pocketed donors—could explain the rise of the Tea Party, which firmly moved affective polarization from a feeling into a political force that led to candidates getting elected. These are the ways in which emotion, mediated by political incentives and politicians, can create changes to democracy.

    Another financial shock that may have played a role is the rise in Chinese imports, which, according to one study, reduced U.S. manufacturing jobs and increased unemployment, particularly among non-college-educated groups.138 Meanwhile, the regions most affected by reduced employment due to Chinese imports shifted their media viewing toward Fox News. Majority-White regions most affected by the substitution of Chinese imports for local jobs were more likely to elect a Republican congressperson and majority-minority regions a Democratic one.139

  6. Racial resentment. People holding hostile attitudes toward racial minorities had been found within both parties throughout the twentieth century. They began moving to the Republican Party after the Voting Rights Act of 1965. The 2016 election was another punctuation point, when a new group of voters scoring high on racial resentment moved from being Democrats or swing voters into the Republican Party.140 A study of 5,000 Americans conducted throughout the 2016 election found that White Trump supporters adopted more racially hostile views toward Black Americans over the course of the campaign in line with their preferred candidate’s rhetoric.141 Nathan Kalmoe and Mason found that White Republicans with higher levels of minority resentment were also more likely to see Democrats as evil or subhuman—that is, to be affectively polarized.142 These alterations coincided with Republicans holding a more racialized view of the Democratic Party (thinking of the party as containing a larger number of minorities than was accurate).143 Meanwhile, for approximately a decade, Democrats had been moving to the left on racial issues.144
  7. Status anxiety. Economics and race may work together to create a sense of fear and othering. The idea that fear of economic scarcity exacerbates racism is corroborated by experiments showing that in a lab context, White Americans primed to be thinking about economic scarcity reduce resource allocations to Black Americans.145 The problem may be linked to the idea of relative deprivation: when individuals feel deprived of success they had anticipated achieving or felt they deserved, they can experience a sense that they deserve better than their current situation and that someone else is to blame.146 Usually these grievances must be articulated and exploited by a conflict entrepreneur or political leader, and until that person or movement comes along, structural inequality per se does not move people into extremism.

    Consider again the rise of the Tea Party. As voters face greater economic precarity in general, particularly following a shock, they become more attuned to further threats and are more sensitive to political appeals that out-group members are supposedly taking their rightful place. Affective polarization grows because of the confluence of economic precarity, status threat, and the existence of a leader or common narrative that blames the situation on particular people or causes.

    The need for other factors to interact with economics may be why other markers such as living in poverty, being working class, being unemployed, and fearing economic loss do not correlate with support for nativist or xenophobic parties in the United States or Europe.147 Some research suggests that nativism (which is not the same as affective polarization but is potentially linked) may be most attractive to those doing less well relative to others within a growing economy—even if those individuals are doing better than they previously had been. For instance, many of the poorest European countries have not seen a major rise in nativist parties, while countries with relatively high employment and prosperity (such as Austria and the Netherlands), as well as regions with expanding economies and employment (such as Saxony and Brandenburg in Germany), have seen rising nativism.148 Poverty also has no correlation with political violence.149 In fact, domestic terrorists are, if anything, slightly better educated and more wealthy than other Americans.150

  8. Politicians and political incentives. Other scholars pointed to the particular role of political leaders and structural incentives in stoking affective, identity-based polarization to win elections. The literature suggests three mechanisms.
    • Personality or individual-level explanations: The rise of populist leaders suggested some politicians’ personalities led them to use polarization as their means to create intense ties to their voting base and thus win elections.151 For instance, countries such as Brazil and Poland, which were previously fairly unpolarized, polarized quickly during the campaigns and under the governments of populist leaders. Meanwhile, in Hungary, a politician drew on old but long-submerged fissures to scapegoat portions of society (which did not exactly polarize because there were not many minorities to form the other “pole”).152
    • Structural incentives: Others pointed to political incentives in both places with close races where control over power was under dispute and in places with first-past-the-post, winner-take-all systems in which moving a tiny number of voters might give a politician a total win over a district or country.153 Majoritarian systems seem to favor affective polarization by creating a binary of political actors.154 For this reason, political scientists have recommended a broad array of structural changes to America’s two-party, winner-take-all system to affect these incentives.155
    • Campaign incentives: Negative campaigns also seemed to play a role. Across recent election cycles, voters were 50–150 percent more affectively polarized by Election Day than they were a year earlier. Political advertisements, especially negative advertising, had particularly strong effects on affective polarization.156

Interventions to Reduce Affective Polarization

These findings led to a diverse set of early experiments to bring down affective polarization. Four strands of such work have received the most scholarship.

  1. Correct misperceptions about the other party, one-on-one or through broader awareness campaigns. Since people who are affectively polarized are known to hold erroneous beliefs and perceptions about the other party, one path that initially appeared promising was correcting these misbeliefs.
    • Correct misperceptions based on identity. Democrats tend to think Republicans are wealthier, older, and more evangelical than they are (for instance, believing that over a third of Republicans earned over $250,000 when the reality is 2 percent), while Republicans think Democrats are more often members of unions and more likely to be atheists, sexual minorities, or people of color (for instance, believing that nearly a third of Democrats were members of the LGBT community while the reality is six percent). Like ideological misperceptions, these demographic misbeliefs are strongest among those who are more politically informed. They increase the feeling that the average member of the other group has little in common with themselves and increase animus.157

      In fact, the modal member of both parties is a middle-aged, White, nonevangelical Christian. In one set of lab experiments, scholars found that when respondents’ misperceptions about the demographics of the other party were corrected, affective polarization decreased.158 Because many people hold biases against the other party’s stereotypical groups, correcting these biases back to the modal member helps—though done without care, correcting misperceptions can further the implicit connection between Whiteness, Christianity, and heterosexuality with “Americanness” (see the discussion in point 2).159

    • Correct misperceptions based on policies. People who are affectively polarized may dislike one another and be unaware that their policy beliefs overlap. Multiple studies found that when partisans were made aware that they shared policy beliefs across parties, their affective polarization declined.160 This finding suggested that although affective polarization is emotional, it may still have a strong policy basis and thus could be changed by correcting misperceptions about policies supported by the other side. Some studies found that issue positions influenced feelings about the other party more than partisanship. If voters learned they shared the same policy beliefs, they rated each other more warmly. Those feelings cooled if they learned they were in different parties, but they were still warmer than feelings toward copartisans who held different political beliefs.161 Another study, part of Stanford University’s Strengthening Democracy Challenge, found that watching a short video in which Democrats and Republicans discovered they shared more common ground than they previously believed reduced affective polarization.162
  2. Increase the salience of a common identity. Multiple studies suggested that efforts to create a “superordinate identity,” or a common, overarching identity, could reduce affective polarization. The “one-group” model emphasizes a shared identity that encapsulates members of all groups. For instance, Levendusky used the Fourth of July to prime American identity and the killing of Osama bin Laden to prime a common threat. He found emphasizing these things that Americans share could reduce partisan animus.163 This was in keeping with other studies that have consistently found that emphasizing a shared, single group membership reduces intergroup bias.164

    However, America’s partisan polarization is interwoven with racial identity, creating an important caveat. Some scholarship has found that subordinating minority identity to a common identity (that is, “we are all Americans”) makes minorities less interested in engaging over time.165 Even if minority individuals are willing to engage in the intervention program, they tend to reassert their minority or subgroup identities, often with some intensity.166 As minority individuals reassert their minority identities (such as claiming that language like “we are all Americans” risks erasing the specific histories of their groups), other people in the intervention may respond with an increase in bias (for instance, feeling that the strong assertion of minority identity is the reason Americans cannot get along).167 Other scholars worry that insofar as creating a superordinate American identity does reduce polarization, it may do so by further associating American identity with Whiteness. Not only could such associations exclude minorities from American identity, but also, the implicit connection between race and American identity could enhance divides as the United States transitions to a country with an increasing proportion of minorities. In other words, so long as the modal member of both parties is White, increasing the salience of American identity reduces partisan prejudice. But efforts to create a single identity can implicitly engage race: for instance, asking people to recall the imagery of America’s founding may implicitly call to mind images of White people. If there is a mismatch between the concept of American identity as White and the actual composition of the opposing party, increasing the salience of American identity could actually correlate with higher levels of partisan hostility.168

    For these reasons, scholars have tried interventions that emphasize both a common American identity and other identities. A 2011 study found that this “dual identity” approach maintained the willingness of disadvantaged groups to engage.169 A 2013 study found that this dual identity approach enabled majority groups to recognize and decrease prejudice against minority groups in America, and studies of other countries showed it worked in reducing prejudice.170

  3. Bring groups together to have social relationships across difference.171 Often called “bridging programs,” these interventions are based on a tradition of intergroup contact theory that dates back to the 1960s and was developed predominantly in reaction to America’s racial divide. In the United States, programs usually bring people together who self-identify as Republican or Democratic to talk across that divide, often to talk about the divide or about divisions on policy issues. Years of scholarship show that programs like these can reduce prejudice if they are very well moderated.172

    However, a 2021 meta-analysis of the extensive literature on intergroup contact theory found serious limitations in applying this theory to political polarization.173 The majority of the studies were conducted with young adults and teenagers, whose identities are still in formation, but interventions are being applied to older adults whose identities are fully formed. Studies largely focused on overcoming prejudice based on characteristics that individuals are born into, such as race and religion, but there is no evidence that the theory continues to work when applied to political identities that people choose and that they may be held morally culpable for holding. Most studies were too small to know whether they could scale. Meanwhile, evidence suggests that these programs work only under tightly circumscribed conditions: for instance, participants across a divide must be of the same economic class. If these programs are run poorly or preconditions are not met, they can backfire and deepen bias.

  4. Make structural alterations to electoral systems. A number of scholars have suggested an array of structural alterations to electoral systems. These are sometimes framed as a way to reduce affective polarization but are more commonly discussed as a way to reduce political incentives to polarize the public or to reduce the likelihood of more extreme partisans or antidemocratic personalities getting elected. Intervention suggestions include forms of proportional representation, such as requiring states with more than a minimum number of districts to elect half of their congressional delegations by individual districts and the other half at large, whether from the entire state at large or from larger, nongerrymandered regions;174 abolishing primaries or making primaries nonpartisan, in which all candidates run and the top two vote getters qualify for the general election;175 instituting ranked choice voting, which is supported both for its possible effects in getting around polarizing primaries and because it may reduce negative campaigning based on a study by the advocacy organization FairVote;176 and finally, setting up proportional representation systems, which allows a broader spectrum of political options to gain representation and ensures that no significant percentage of the population fails to get represented because it represents a minority within a given geography in the way winner-take-all systems allow.177

Third Generation Understanding: Cracks in the Foundations

Reducing Affective Polarization May Not Impact Violent or Antidemocratic Attitudes

More recent data, however, has introduced some core challenges for any interventions that are intended to reduce affective polarization alone.

A number of recent studies have found that interventions to decrease affective polarization can change participants’ feelings toward members of the other party. But they don’t seem to affect the antidemocratic or violent attitudes and behaviors that were supposedly being caused by affective polarization.178

A study of the interventions in Stanford’s Strengthening Democracy Challenge (which tested short, online interventions on 30,000 participants) found that many of the lab interventions reduced partisan animosity immediately, and some effects even persisted two weeks later.179 But the interventions did not reduce antidemocratic attitudes or justifications for political violence. In fact, one intervention that was based on creating a shared identity as an “exhausted majority” actually increased support for antidemocratic practices. Meanwhile, a video that graphically depicted social instability and violence in collapsing democracies around the world reduced affective polarization while increasing conservative Republican support for political violence (possibly because it included footage of January 6).180 A separate study by an overlapping group of researchers used three completely different tests on the findings from the Strengthening Democracy Challenge. They also found that interventions that reduced affective polarization in lab environments did not seem to affect antidemocratic attitudes.181 A third study with different researchers conducted five different lab-based tests and agreed that affective polarization is linked to how people feel interpersonally about the other party’s members, but it is not related to their support for partisan views, policy positions, or democratic norms, nor is it related to how people react to actual information about their real representatives and other tested political effects.182

Stepping back from the interventions, these scholars also looked at whether affective polarization was even correlated with antidemocratic attitudes or political violence. They found that social distrust, partisan animosity, and biased evaluation of politicized facts (all hallmarks of affective polarization) did group together. But these factors were not correlated closely with support for undemocratic practices, political violence, and undemocratic candidates, and interventions that affected the first set of variables that characterize affective polarization did not necessarily affect the second set of variables, which were what researchers ultimately wanted to change.183 Subsequent testing by an overlapping group of researchers using different tests found that there was no reliable correlation between affective polarization and antidemocratic attitudes or political violence.184 These findings echo the literature review on intergroup contact theory, which showed that bridging programs can be successful at changing the feelings of the people who participate but provided no evidence that changing such feelings will affect individuals’ actual behaviors, such as voting for antidemocratic candidates or supporting political violence.185

Antidemocratic Attitudes

A few interventions within the Strengthening Democracy Challenge were found to reduce both affective polarization and antidemocratic attitudes. These were interventions that reassured participants that the other side was not going to undermine democracy itself.

One intervention that reduced antidemocratic attitudes was a video showing Utah’s gubernatorial candidates from both parties committing themselves to honor the results of the 2020 election. Another program corrected misperceptions about the other party’s willingness to engage in antidemocratic behaviors, such as reducing the number of polling stations, refusing to accept election results, or using violence to block laws.186

Learning that partisans on the other side were not so bent on playing dirty and destroying democracy reduced support for antidemocratic actions from one’s own side.187 That intervention found support from other scholars who showed that voter decisions to support an antidemocratic candidate seem sensible if they think the harm from their decision is small while the consequences of the opposition winning are catastrophic to democracy.188 These findings by scholars such as Svolik previously cited affective polarization as a cause, but in fact, they could be right about their empirical findings but wrong about the cause. Newer research suggests that the problem is not general feelings of dislike for the other party but instead specific fears that the other party will undermine democracy.

Political Violence

The data do not support the idea that affective polarization directly causes political violence. In fact, justifying political violence seems separate from both affective polarization and antidemocratic attitudes. That makes sense given what is known about violence; the literature suggests it has much more to do with personality and self-control than emotions or ideology.

In addition to the lab tests and survey responses already described that show a lack of correlation between attitudes supporting political violence and affective polarization, other facts do not match. For instance, Democrats are about equally affectively polarized as Republicans. In fact, from 2017 to 2020—the years for which Kalmoe and Mason published survey data—political violence justifications were nearly as common (and sometimes more prevalent) on the left as on the right.189 The Strengthening Democracy Challenge also found Democrats to be almost as supportive of political violence as Republicans and only slightly less tolerant of antidemocratic attitudes. 190 But actual political violence is vastly higher on the right. If emotions were driving violence, there would be more correlation between the depth of emotion and the intensity of violence.

Moreover, the timing of the rise in affective polarization does not match the rise in political violence in the United States. Affective polarization has been rising fairly steadily since the 1970s or around 2000, depending on which measure is being used, and took another leap around 2008, after which it continued to rise steadily (see earlier, figure 6).

Yet hate crimes in the United States spiked after the terrorist attacks on September 11, 2001, and then fell through 2014, when they started to rise to what is now the highest-recorded level in the twenty-first century (see figure 9).191

Premeditated violent attacks by the left and right started to rise around 2011–2012 and then spiked sharply in 2016 (see figure 10).

Threats against members of Congress remained fairly steady until 2016, when they rose tenfold in five years (then fell after 2020).192 Threats against local officials, electoral officials, and public health and school officials show even less correlation. While data are poor retrospectively, local elected officials and electoral officials seem to have faced increased threats since around 2018 or 2019, while health and school officials report greater threats starting in 2020.193 While these forms of political violence could be a lagging indicator (growing only a few years after affective polarization had risen), that would not explain the drop in total hate crimes and right- and left-wing domestic terrorism during years when affective polarization was rising after 2001. The time lines for various forms of political violence simply do not correlate with rising affective polarization.

  • Having an aggressive personality is the variable that most strongly predicts whether someone justifies political violence. It is a trait that is completely separate from ideology. People from both parties who hold more hostile sexist beliefs are also more likely to justify violence.194 For most politically violent individuals, ideology tends to be a rationale that is used to cloak an aggressive personality in a larger cause. Further demonstrating that violence is not directly related to affective polarization, hostility toward racial minorities is the variable most predictive of right-wing affective polarization. But aggression and hostility toward women is more predictive than racism for right-wing justifications for violence.195

    There is also some new research suggesting that an authoritarian personality may play a role in augmenting someone’s aggression in the case of political, but not interpersonal, violence. The traditional authoritarian scale was built in the 1930s to explain fascism and has questions that conflate authoritarianism with conservativism and with working-class child-rearing, which tends to emphasize obedience to authority. (Researchers have suggested that this may be because working-class parents are trying to prepare their children for jobs in which obeying a boss is more important for success than creativity or thinking for oneself.) The scale thus overpredicts for Republicans being authoritarian and is more of a measurement of ideology than personality. A new scale by Emory University researchers examines a constellation of cognitive and personality traits to create a definition of authoritarianism that works across ideologies by looking for a willingness to use coercion and punitive measures to enforce values in the political sphere. While the values being supported through coercion are polar opposites between the right and left, there is significant overlap in the personality traits, suggesting that ideology is secondary to a mindset that leads those with authoritarian personalities to advocate more autocratic, antidemocratic action, including violence, in support of their preferred world order.196

  • The strongest correlation with committing any form of violence is having poor self-control, which is correlated with childhood abuse, alcohol and drug abuse, untreated psychosis, being a young male, and other factors, none of which are ideological and all of which describe large groups of people, most of whom are not violent. Hate crimes, for instance, are more frequently committed by young, unemployed men with criminal histories.197 Other than premeditated domestic terrorism incidents, most violent acts that occur at political events or hate crimes are also likely spontaneous, impulsive decisions by individuals influenced by a crowd. Such spontaneous crowd dynamics are why placing violence instigators (agents provocateurs) within crowds is a long-standing tactic of authoritarian leaders.

    How do lifelong personality traits held by a relatively stable percent of a population transform into sudden changes in the rate of political violence? Increased fear and sense of threat as well as stress are likely to play a role. These feelings may be triggered by events (such as a pandemic or a rise in crime). But who aggressive people are violent toward, and how much they feel their violence is tolerated by state authorities, is likely to be affected by polarization. Aggression that may have been directed toward neighbors or intimate partners may be channeled toward other targets based on in-group cues of who is believed to be creating a threat. A sense that public aggression is less likely to be punished also appears to play a role in political violence—this variable is not related to polarization but to state response. In-group leaders, such as politicians and media personalities, play a particularly important role in normalizing violence; dehumanizing certain groups or individuals to make violence against them more likely; suggesting who or what group is a threat; and, by aggrandizing violent individuals, paying their legal fees, or offering them pardons, creating a sense of who is and is not likely to be punished.198 Government leaders also play a role based on whether they hold the violent individual accountable. A sense that the state will provide impunity increases the likelihood of violence and suggests another strong role for political structures and incentives.199

Political Structures Affect Incentives to Polarize

Increased understanding that reducing affective polarization does not reduce antidemocratic attitudes or political violence has made U.S.-focused democracy scholars more aware of the importance of the interplay between political incentives and mass emotions.

For instance, after an interview with Herbert P. Kitschelt, a political scientist at Duke University, New York Times columnist Thomas B. Edsall wrote that “too much of the debate over affective polarization and democratic backsliding has been restricted to the analysis of competing psychological pressures.” Instead of a “myopic and U.S.-centric” focus on psychology, Kitschelt said, it is important to acknowledge the incentives “for politicians to prime this polarization and stoke the divides, including fanning the flames of affective polarization.”200 Meanwhile, Brendan Nyhan, a political scientist at Dartmouth, explained, “we should renew our scrutiny of the role of elites and political systems in fomenting illiberal behavior.” To him, the problem “is not affective polarization as such; it’s a political system that is failing to contain significant democratic erosion and illiberalism being driven by G.O.P. elites (though affective polarization may help encourage and enable such tactics).”201

This understanding accords with the more recent scholarship among experts on comparative political systems and scholars of civil war, who have begun to look simultaneously at affective polarization among the public and at whether political structures incentivize political parties and elites to catalyze or amplify such emotional polarization. For instance, looking at a definition of “pernicious polarization” that includes both these structural and emotional factors, McCoy and Benjamin Press found that among less-developed, less-wealthy democracies that had become perniciously polarized, half of the cases had their democracy ratings downgraded in Varieties of Democracy data.202 The U.S. government–funded Political Instability Task Force, charged with finding the variables most correlated with a regime facing instability or war, found that factionalized political parties where multiple identities aligned with partisanship was the most correlated variable with future conflict, in part because of the ways that correlation incentivized politicians to amplify the fissures politically.203 Barbara F. Walter, a scholar of civil wars who served on that task force, claims that elite factionalization and affective polarization played a role in civil conflict in other countries.204


What We (Think We) Know in 2023

Ideological Polarization

  • Congress remains very ideologically polarized. Since the Tea Party Caucus entered in 2011, there has been almost no issue area overlap between members of Congress across the two parties (see figure 11).


  • Within the American public, Democrats have moved ideologically to the left and Republicans to the right (see figure 12). Both have also become more consistent in their beliefs (that is, liberals moved into the Democratic Party and conservatives into the Republican Party). But there remains overlap and policy agreement on many issues. As previously mentioned, because of the polarization among politically engaged partisans, this may be difficult for many people who form this paper’s audience to believe. But it is worth remembering that the overwhelming majority of Americans do not think much about politics, do not hold strong ideological views, and can remain quite inconsistent or apt to alter replies when survey wording changes because their views are so weakly held. What is held strongly is a sense of identity. However, political campaigns trigger identity, making it hard to disaggregate in the real world.


In 2016, a cohort of White swing voters who favor economic redistribution but also exhibit greater racial hostility moved into the Republican Party.205 Since then, the Republican Party has had greater ideological heterogeneity—and more demographic homogeneity (see figure 13). In other words, when more party members are White men but fewer hold the same views on low taxes and small government, the party has an incentive to use identity-based affective polarization as a strategy.

  • A consistent misunderstanding is that people who hold views from both sides of the aisle are moderates. In fact, survey findings from the Democracy Fund Voter Study Group found that Americans who held the least polarized ideological beliefs were actually the voting cohort least in favor of democracy and most supportive of a “strong leader” who did not need to bother with Congress or elections. The preponderance of Americans who respond to ideological survey questions with answers on both sides of the aisle tend to be pro–economic redistribution (for their group) while also upholding a White, Christian, U.S.-born norm of American citizenship.206 These voters were long (and correctly) classified as swing voters, although they have since 2016 moved more decisively into the Republican Party. But these former swing voters are often erroneously classified as centrist or moderates by surveyors because they hold cross-partisan beliefs. They are better viewed not as moderates but as disgruntled, not very intense on policy or partisanship, and only very loosely attached to democracy.

Affective Polarization

  • The American public feels affectively polarized largely because of misunderstandings about the other side (though the misunderstandings seem sensitive to actual ideological differences). Older Americans are polarizing more quickly than younger Americans. And the United States is polarizing much more rapidly than other Western democracies; partisans even more so. However, the United States may not be more affectively polarized than a number of other multiparty democracies whose systems are functioning much better. The rapidity of U.S. polarization compared to similar wealthy, consolidated democracies suggests that domestic issues in the United States are likely to be driving more of the country’s polarization than issues affecting many other countries.
  • A number of interventions have been shown in lab settings, games, and short experiments to reduce affective intervention in the short term. These include:
    • correcting misperceptions about the other party’s demographics;
    • correcting misperceptions about the other party’s ideological beliefs;
    • forming a dual identity that builds a shared common identity alongside existing minority or other identities; and
    • bringing people of similar economic class and age together in controlled, well-moderated circumstances to engage in a shared activity while building understanding.
  • Reducing affective polarization through these lab experiments and games has not been shown to affect regular Americans’ support for antidemocratic candidates, support for antidemocratic behaviors, voting behavior, or support for political violence.
  • People may be more at risk of affective polarization and more supportive of same-party antidemocratic breaches if they fear that the other party will gain power and use it to undermine democracy. Affective polarization is likely driven more by feelings of threat than simply feelings of dislike.
  • America’s entire media system—not just social media—may be playing a role in both ideological and affective polarization. Highly polarizing cable news and talk radio shows are probably more to blame than social media, and all of their polarizing effects are likely exacerbated because the United States lacks a single trusted media source or trusted local media organizations. Increasing the availability of trusted local media may be a helpful remedy but requires more study.
  • The interaction of economic precarity in rural areas facing a tougher recovery from the 2008–2009 financial crisis—at a moment when greater cultural attention was being directed at racial identity and politicians were exacerbating a sense of status anxiety—may be making rural Americans more vulnerable to racial resentment and affective polarization.
  • Politicians and political incentives are probably playing a larger role in driving affective polarization than structural issues such as inequality or geographic sorting.
  • Interventions to reduce affective polarization will be ineffective if they operate only at the individual, emotional level. Ignoring the role of polarizing politicians and political incentives to instrumentalize affective polarization for political gain will fail to generate change while enhancing cynicism when polite conversations among willing participants do not generate prodemocratic change.
  • Adversarial approaches that position other Americans as immoral and antidemocratic also ignore the understanding that affective polarization is a political strategy. By deepening the binary, us-versus-them frame, they offer a gift to polarizing politicians who wish to generate backlash and deepen polarization.

Finally, it’s appropriate to draw two takeaways from the decade’s worth of research into polarization in general, and recent findings regarding affective polarization.

First, researchers are drawing overly broad, even hyperbolic, conclusions in both positive and negative directions from short-term lab experiments. For example, a 2022 paper by Jan G. Voelkel and other researchers analyzing a “megastudy” of the dozens of Stanford anti-polarization interventions found that a number of “successful” interventions reduced affective polarization based on effects that were tested at most two weeks afterward.207 A second paper released later that year by Voelkel and an overlapping group of scholars declared, “our findings suggest that affective polarization may not be as problematic for democratic societies as is widely assumed.”208

Self-reported surveys based on short, online interventions, trust games, and other lab tests are not necessarily indicative of actual behavior in the real world, where people are subject to greater social cues, accountability, media bombardment, and other pressures pushing in multiple directions. All claims of success or failure should be taken with a grain of salt.

Second, it is not enough to focus only on interventions that reduce the emotions of affective polarization at the individual level. Interventions should consider the interplay between affective polarization and political structures, incentives, and strategies.

Interventions that focus on individuals’ emotions may not be useful at all because they tend to target willing and more moderate participants who are not driving current democratic problems. Or they may not be useful unless carried out in tandem with interventions that directly address antidemocratic behaviors and/or violence at the social and political level. That understanding would accord with the growing literature on comparative democracies and civil war scholarship. A metastudy of interventions to reduce conflict in other countries, for instance, found that “programming that focuses on change at the individual level that never links or translates into action at the Socio-Political level has no discernable effect on peace.”209 Dialogue can improve understanding and reduce prejudice, but it will not result in social or political improvements at the state or national levels unless those relationships are directed toward collaborative action that creates positive change in a broader cultural or political context, not just the feelings of individuals.210

At the same time, interventions focused on changing political or social structures that ignore how polarization is used as a political strategy are likely to generate backlash, deepen polarization, and exacerbate a sense of threat that enables antidemocratic behaviors. In a perniciously polarized environment like the United States, highly adversarial, polarizing advocacy—even to achieve prodemocratic ends—can be used instrumentally by polarizing politicians to cement their political success. Efforts at cultural and political change are essential, but in an environment in which affective polarization is rife and is being used strategically by politicians, advocacy for change must sidestep polarization by engaging unlikely allies who agree on the particular issue despite often being on opposite sides of polarized debates. Advocacy that attempts to overcome the other side by simply amassing on one side of a polarized divide will generally result either in failure or in cycles of Pyrrhic victory and cultural or political backlash that deepens the extremism of the other side.211

Finally, the third generation understanding of polarization, which is just emerging in the United States but is more established abroad, considers the political structures that incentivize politicians to exacerbate affective polarization for their own benefit. Winner-take-all political systems, where winning 50.1 percent of the vote wins the entire district even if 49.9 percent of the voters wanted someone else, are correlated with more affective polarization, as well as greater discontent with one’s own party.212 Altering U.S. political structures in order to change the political incentives that are exacerbating affective polarization is almost certainly part of the solution to ensuring a more cohesive citizenry that supports a stronger democracy in the United States.


The author of this paper is grateful to the William and Flora Hewlett Foundation for supporting this work and to the William and Flora Hewlett and David and Lucile Packard Foundations for asking the question that prompted the initial draft of this paper. The author is solely responsible for the paper’s contents.


1 Frank Newport, “U.S. Opinion and the Election: Guns, Immigration, Climate,” Gallup, October 21, 2022,; and Stephen Hawkins et al., “Defusing the History Wars,” More in Common, December 2022,

2 Katherine Schaeffer, “Key Facts About Americans and Guns,” Pew Research, September 13, 2021,

3 Robert Griffin et al., “Racing Apart: Partisan Shifts on Racial Attitudes Over the Last Decade,” Democracy Fund Voter Study Group, October 2021,; and Michael Tesler, “Republican Views on Immigration Are Shifting Even Further to the Right Under Biden,” FiveThirtyEight, August 17, 2021,

4 Delia Baldassarri and Barum Park, “Was There a Culture War? Partisan Polarization and Secular Trends in US Public Opinion,” Journal of Politics 82, no. 3 (2020):

5 Nolan McCarty, “What We Know and Don’t Know About Our Polarized Politics,” Washington Post, December 7, 2021,; Nolan McCarty, Keith T. Poole, and Howard Rosenthal, “Does Gerrymandering Cause Polarization?,” American Journal of Political Science 53, no. 3 (2009): 666–680,; Michael Barber and Nolan McCarty, “Causes and Consequences of Polarization,” in Political Negotiation: A Handbook, ed. Jane Mansbridge and Cathie J. Martin (Washington, DC: Brookings Institution Press, 2015), 37–90.

6 Andrew B. Hall, Who Wants to Run?: How the Devaluing of Political Office Drives Polarization (Chicago: University of Chicago Press, 2019).

7 David E. Broockman et al., “Why Local Party Leaders Don’t Support Nominating Centrists,” British Journal of Political Science 51 (2021): 724–749,

8 “The Primary Problem,” Unite America, March 2021, Of the ten House Republicans who voted to impeach former president Trump, four retired, four were defeated in their primaries, and only two succeeded in a primary election. See Eugene Scott, “What Happened to the 10 Republicans Who Voted to Impeach Trump?,” Washington Post, November 23, 2022, Of the thirty-five Republicans who voted to create the January 6 Commission, those who ran for reelection were four times as likely to face a competitive primary as other House incumbents and were five times more likely to lose. Those who voted against the January 6 Commission were ten times more likely not to have a primary opponent. Michael Podhorzer, “The MAGA Congressional Purge,” Weekend Reading, newsletter, September 18, 2022, An alternative view suggests that Republican primary voters are fairly similar to all Republican voters, suggesting there would not be a great difference if primaries were abolished; see: Lee Drutman, “What We Know About Congressional Primaries and Congressional Primary Reform,” New America, July 1, 2021,

9 Drew Desilver, “Turnout Was High in the 2016 Primary Season, but Just Short of 2008 Record,” June 10, 2016, For primary turnout after 2016, see Joshua Ferrer and Michael Thorning, “2022 Primary Turnout,” Bipartisan Policy Center, March 6, 2023, The independent vote share began to rise around 2004, reaching 43 percent in 2014 and has since moved slightly up and down, landing at 41 percent in polling from January 1, 2022. See Jeffrey M. Jones, “U.S. Party Preferences Evenly Split in 2022 After Shift to GOP,” Gallup, January 12, 2023,

10 Shanto Iyengar, Gaurav Sood, and Yphtach Lelkes, “Affect, Not Ideology,” Public Opinion Quarterly 76 (2012):; and Levi Boxell, Matthew Gentzkow, and Jesse M. Shapiro, “Greater Internet Use Is Not Associated With Faster Growth in Political Polarization Among US Demographic Groups,” PNAS Proceedings of the National Academy of Sciences 114, no. 40 (2017): 1,0612–1,0617,

11 Levi Boxell, Matthew Gentzkow, and Jesse M. Shapiro, “Cross-Country Trends in Affective Polarization,” National Bureau of Economic Research, Working Paper no. 26669, January 2020,; and Jill Kimball, “U.S. Is Polarizing Faster than Other Democracies, Study Finds,” Brown University, January 21, 2020,

12 Boxell, Gentzkow, and Shapiro, “Greater Internet Use is Not

Associated With Faster Growth in Political Polarization among US Demographic Groups.” Nielsen statistics show that teens watch less than an hour of television a day and people between the ages of eighteen and thirty-four spend just over an hour a day watching television. However, Americans aged sixty-five and over watch over six and a half hours of television each day. See “The Nielsen Total Audience Report Q3 2018,” Nielsen, 2019, Why Radio shows that those over sixty-five comprise the greatest percentage of listeners to All Talk stations (40.4 percent), followed by fifty-five to sixty-four year olds (26.4 percent). MRI Simmons 2022 data for the Radio Advertising Bureau; see “Radio Listening by Format” at

13 Jon C. Rogowski and Joseph L. Sutherland, “How Ideology Fuels Affective Polarization,” Political Behavior 38 (2016): 485–508,; Lori D. Bougher, “The Correlates of Discord: Identity, Issue Alignment, and Political Hostility in Polarized America,” Political Behavior 39 (2017): 731–762,; and “America’s Divided Mind: Understanding the Psychology That Drives Us Apart,” Beyond Conflict, January 6, 2021,

14 Matthew S. Levendusky and Neil Malhotra, “(Mis)perceptions of Partisan Polarization in the American Public,” Public Opinion Quarterly 80 (2016): 378–391,; Sean J. Westwood and Erik Peterson, “The Inseparability of Race and Partisanship in the United States,” Political Behavior 44 (2022): 1,125–1,147,; Jan G. Voelkel et al., “Megastudy Identifying Effective Interventions to Strengthen Americans’ Democratic Attitudes,” Strengthening Democracy Challenge, Stanford University, 2022, Matthew Levendusky, Our Common Bonds: Using What Americans Share to Help Bridge the Partisan Divide (Chicago: University of Chicago Press, 2023); and John F. Dovidio, Samuel L. Gaertner, and Tamar Saguy, “Another View of ‘We’: Majority and Minority Group Perspectives on a Common Ingroup Identity,” European Review of Social Psychology 18 (2010):

15 Elizabeth L. Paluck et al., “Prejudice Reduction: Progress and Challenges,” Annual Review of Psychology 72 (2021):

16 Diego Garzia, Frederico Ferreira da Silva, and Simon Maye, “Affective Polarization in Comparative and Longitudinal Perspective,” Public Opinion Quarterly 87, no. 1 (2023):; Noam Gidron, James Adams, and Will Horne, “American Affective Polarization in Comparative Perspective,” (Cambridge: Cambridge University Press, 2020):; and Markus Wagner, “Affective Polarization in Multiparty Systems,” Electoral Studies 69 (2021):

17 Fred Duong et al., “Outgroup Testimonials and Ingroup Validation Strengthen the Effects of Perception Gap Interventions on Affective Polarization: Evidence from a Large-Scale Experiment,” More in Common, preprint, May 16, 2023,

18 Jan G. Voelkel et al., “Interventions Reducing Affective Polarization Do Not Necessarily Improve Anti-Democratic Attitudes,” Nature Human Behaviour 7 (2022):; and David Broockman, Joshua Kalla, and Sean Westwood, “Does Affective Polarization Undermine Democratic Norms or Accountability? Maybe Not,” under review for American Journal of Political Science, January 13, 2022,

19 In Israel, Seeds of Peace occasionally found ways to build deeper relationships that led to continued activism—but most programming to bridge the Israeli-Palestinian divide by building understanding was conducted with less care and has had shockingly poor results. See Matthew Kalman, “Will Seeds of Peace Ever Bloom?,” Haaretz, September 15, 2014,

20 Rachel Kleinfeld, “Five Strategies to Support U.S. Democracy,” Carnegie Endowment for International Peace, September 15, 2022,

21 Nathan P. Kalmoe and Lilliana Mason, Radical American Partisanship: Mapping Violent Hostility, Its Causes, and the Consequences for Democracy (Chicago: University of Chicago Press, 2022), 88.

22 Garen Wintemute, Sonia Robinson, and Elizabeth A. Tomsich, “MAGA Republicans’ Views of American Democracy and Society and Support for Political Violence: Findings From a Nationwide Population-Representative Survey,” preprint submitted to SocArXiv, October 21, 2022,; and “Tempered Expectations and Hardened Divisions a Year Into the Biden Presidency,” Bright Line Watch, November 2021 surveys,

23 Kalmoe and Mason, Radical American Partisanship.

24 Vanja Ljujic Pauwels and Ann De Buck, “Individual Differences in Political Aggression: The Role of Social Integration, Perceived Grievances and Low Self-Control,” European Journal of Criminology (2018): https://; and M. Gottfredson, “Self-Control and Criminal Violence,” in A. T. Vazsonyi, D. J. Flannery, and M. DeLisi (eds.), The Cambridge Handbook of Violent Behavior and Aggression, 2nd ed., (Cambridge: Cambridge University Press, 2018).

25 Nathan P. Kalmoe, “Fueling the Fire: Violent Metaphors, Trait Aggression, and Support for Political Violence,” Political Communication 31, no. 4 (2014):

26 Dehumanizing language likely makes it easier for violence to occur, but it does not act in a vacuum—the important factor is the creation of a sense that a group is outside the in-group and that other community members, particularly leaders, view violence as positive or acceptable against out-groups. See Jonathan Leader Maynard and Aliza Lift, “Humanizing Dehumanization Research,” Current Research in Ecological and Social Psychology 4 (2023):

27 On hate crimes, see “Federal Bureau of Investigation Crime Data Explorer,” Federal Bureau of Investigation, FBI hate crimes data is notoriously poor because it relies on police agencies to self-report, and many do not. However, more departments have failed to report more recently, so these trends are likely to be even worse than the FBI statistics suggest, not better. On premeditated violence, see Kleinfeld, “Five Strategies to Support U.S. Democracy.” For threats against local officials, see “New Report: Harassment, Threats and Violence Directed at Local Elected Officials Rising at an Alarming Rate,” National League of Cities, June 24, 2022, For threats against members of Congress, see “USCP Threat Assessment Cases for 2022,” United States Capitol Police, January 17, 2023,

28 Jennifer McCoy and Benjamin Press, “What Happens When Democracies Become Perniciously Polarized?,” Carnegie Endowment for International Peace, January 18, 2022,

29 Paluck et al., “Prejudice Reduction.”

30 CDA Collaborative Learning Projects, “Reflecting on Peace Practice (RPP) Basics: A Resource Manual,” CDA Collaborative, January 19, 2017,

31 Thomas Edsall, “How Much Does How Much We Hate Each Other Matter?,” New York Times, September 29, 2021,

32 Shamil Idriss and Rachel Kleinfeld, “No One Is Right in the Debate for and Against Philanthropic Pluralism,” Chronicle of Philanthropy, June 15, 2023,

33 M. H. Pasek et al., “Misperceptions About Out-Partisans’ Democratic Values May Erode Democracy,” OSF Preprints (February 2022):; and Alia Braley et al., “The Subversion Dilemma: Why Voters Who Cherish Democracy Participate in Democratic Backsliding,” Nature Human Behavior (2023):

34 Joseph Mernyk et al., “Correcting Inaccurate Metaperceptions Reduces Americans’ Support for Partisan Violence,” OSF Preprints (September 20, 2021):

35 Mernyk et al., “Correcting Inaccurate Metaperceptions Reduces Americans’ Support for Partisan Violence.”

36 Duong et al., “Outgroup Testimonials and Ingroup Validation Strengthen the Effects of Perception Gap Interventions on Affective Polarization.”

37 R. S. Foa et al., “Youth and Satisfaction With Democracy: Reversing the Democratic Disconnect?” Cambridge, United Kingdom: Centre for the Future of Democracy, 2020,

38 A June 2022 poll by the University of Chicago Institute of Politics found that 73 percent of Republicans, two-thirds of Republicans and independents, and 51 percent of Americans who described themselves as “very liberal” agreed that the government is “corrupt and rigged against everyday people like me.” While Republican distrust could be assumed to rise under a Democratic president, these findings accorded with a study asking a variation on this question in 2015 and 2019, both of which found 70 percent of Americans claiming they were angry “because our political system seems to only be working for the insiders with money and power, like those on Wall Street or in Washington.” Carrie Dann, “A Deep and Boiling Anger’: NBC/WSJ Poll Finds a Pessimistic America Despite Current Economic Satisfaction,” NBC News, August 25, 2019, These findings hold across multiple polls, see for instance: “Our Precarious Democracy: Extreme Polarization and Alienation in Our Politics,” Public Opinion Strategies, University of Chicago Institute of Politics, and Beneson Strategy Group, July 2022, From 2016 through 2022, “fear of corrupt government officials” ranked number one in the Chapman University Fear Index, which tracks what a small but statistically significant sample of Americans claim to fear. See “The Chapman University Survey on American Fears,” Voice of Wilkinson (blog), Chapman University, accessed July 10, 2021,

39 See Charles Stewart III’s data from the American National Election Survey, presented in figure 4 in this paper.

40 Gallup polling from June 1–20, 2022, found that a third or more of respondents held “very little” or “no” trust in big business, schools, newspapers, television news, Congress, the criminal justice system, and organized religion, while 25 percent of respondents distrusted organized labor, 24 percent distrusted the medical system, and 23 percent distrusted banks. See raw data, Jeffery M. Jones, “Confidence in U.S. Institutions Down; Average at New Low,” Gallup, July 5, 2022,

41 Sara Cobb, Jale Sultanli, and Alison Castel, “Collaborating Across Difference to Reduce Authoritarianism: A Literature Review,” Horizons Project, See also Rachel Kleinfeld, A Savage Order (New York: Random House, 2018).

42 Lee Drutman, Larry Diamond, and Joe Goldman, “Follow the Leader: Exploring American Support for Democracy and Authoritarianism,” Democracy Fund Voter Study Group, March 13, 2018,

43 Drutman, Diamond, and Goldman, “Follow the Leader.”

44 Drutman, Diamond, and Goldman, “Follow the Leader.”

45 More in Common’s research on hidden tribes captures a part of this group in its “embattled conservatives” grouping, of whom 42 percent do not identify as Republicans. They may or may not be affectively polarized against Democrats. See “It’s Complicated: People and Their Democracy in Germany, France, Britain, Poland, and the United States,” More in Common, 2020, 112–117,

46 “Tempered Expectations and Hardened Divisions a Year Into the Biden Presidency,” Bright Line Watch.

47 Jared Holt, “Three Narratives Dissuading Pro-Trump Communities From Organizing on Behalf of the Former President,” Institute for Strategic Dialogue, May 29, 2023,

48 “It’s Complicated: People and Their Democracy,” More in Common, 117–122.

49 David Siders and Meredith McGraw, “Just How Big Is the Always Trump Component of the Republican Party?,” Politico, February 28, 2023,

50 “Toward a More Responsible Two-Party System: A Report of the Committee on Political Parties,” American Political Science Review 44, no. 3 (1950):

51 Matthew S. Levendusky, “Clearer Cues, More Consistent Voters: A Benefit of Elite Polarization,” Political Behavior 32 (2010): 111–131,

52 Ilyana Kuziemko and Ebonya Washington, “Why Did Democrats Lose the South? Bringing New Data to an Old Debate,” American Economic Review 108, no. 10 (October 2018): 2,830–2,867,

53 Thomas Carothers, “The Long Path of Polarization in the United States,” in Democracies Divided: The Global Challenge of Political Polarization, eds. Thomas Caorthers and Andrew O’Donohue (Washington, DC: Brookings Institution Press, 2019), 65–92.

54 Julian E. Zelizer, Burning Down the House: Newt Gingrich, the Fall of a Speaker, and the Rise of the New Republican Party (Penguin Press, 2020).

55 This gave them something to count, but it is far from an ideal measure, as anyone who studies Congress knows that a vast amount of the politics and ideological decisionmaking regarding legislation occurs in the stages of bill creation and agenda setting.

56Red and Blue Nation?: Characteristics and Causes of America’s Polarized Politics, eds. Pietro S. Nivola and David W. Brady (Washington, DC: Brookings Institution Press and Hoover Institution, 2006); “Negotiating Agreement in Politics,” eds. Jane Mansbridge and Cathie J. Martin, American Political Science Association, Task Force Report, December 2013,; and Michael Barber and Nolan McCarty, “Causes and Consequences of Polarization,” in Political Negotiation: A Handbook, ed. Jane Mansbridge and Cathie J. Martin (Washington, DC: Brookings Institution Press, 2015), 37–90.

57 Michael Barber and Nolan McCarty, “Causes and Consequences of Polarization,” in “Negotiating Agreement in Politics,” eds. Mansbridge and Martin.

58 McCarty, Poole, and Rosenthal, “Does Gerrymandering Cause Polarization?”; and Barber and McCarty, “Causes and Consequences of Polarization,” in Political Negotiation: A Handbook, eds. Mansbridge and Martin.

59 “Political Polarization in the American Public,” Pew Research Center.

60 Alan I. Abramowitz and Kyle L. Saunders, “Is Polarization a Myth?,” Journal of Politics

70, no. 2 (April 2008): 542–555,

61 “Political Polarization in the American Public,” Pew Research Center, June 2024,

62 Peter K. Enns and Ashley Jardina, “Complicating the Role of White Racial Attitudes and Anti-Immigrant Sentiment in the 2016 US Presidential Election,” Public Opinion Quarterly 85, no. 2 (Summer 2021): 539–570,

63 Morris P. Fiorina, Samuel A. Abrams, and Jeremy C. Pope, “Polarization in the American

Public: Misconceptions and Misreadings,” Journal of Politics 70, no. 2 (April 2008): 556–560,

64 Newport, “U.S. Opinion and the Election.”

65 See for instance: Abramowitz and Saunders, “Is Polarization a Myth?”

66 Frances Lee, “American Politics Is More Competitive Than Ever. That’s Making Partisanship Worse,” Washington Post, January 9, 2014,; and Frances Lee, “How Party Polarization Affects Governance,” Annual Review of Political Science 18 (May 2015): 261–282,

67 Hall, Who Wants to Run?

68 Broockman et al., “Why Local Party Leaders Don’t Support Nominating Centrists.”

69 John V. Duca and Jason L. Saving, “Income Inequality and Political Polarization: Time Series Evidence Over Nine Decades,” Federal Reserve Bank of Dallas, Working Paper no. 1408, January 2014,; and Toben Iversen and David Soskice, “Information, Inequality, and Mass Polarization: Ideology in Advanced Democracies,” Comparative Political Studies 48, no. 13 (2015):

70 Nolan McCarty, Keith T. Poole, and Howard Rosenthal, Polarized America: The Dance of Ideology and Unequal Riches (Cambridge: MIT Press, 2006).

71 Alan I. Abramowitz, “Disconnected, or Joined at the Hip?” in Red and Blue Nation? ed. Nivola and Brady.

72 For more on this view, see the individual work of scholars who contributed to the task force report, such as: McCarty, Poole, and Rosenthal, “Does Gerrymandering Cause Polarization?”; and Nolan McCarty, “Hate Our Polarized Politics? Why You Can’t Blame Gerrymandering,” Washington Post, October 26, 2012,

73 The Fairness Doctrine was in place from 1949 to 1987 and required the three major television broadcasters to air discussions of important and controversial public policy topics and provide equal time to contrasting views. Its abolishment allowed the rise of highly partisan talk radio and cable shows that did not attempt to offer equal time to contrasting views.

74 Diana C. Mutz, “How the Mass Media Divides Us,” in Red and Blue Nation? ed. Nivola and Brady.

75 Morris P. Fiorina and Samuel J. Abrams, “Political Polarization in the American Public,” Annual Review of Political Science 11 (June 2008): 563–588,; and James N. Druckman, Erik Peterson, and Rune Slothuus, “How Elite Partisan Polarization Affects Public Opinion Formation,” American Political Science Review 107, no. 1 (February 2013): 57–79,

76 John Hudak, “Congress in 2019: Why the First Branch Should Bring Back Earmarks,” Brookings Institution, December 27, 2018,

77 Fiorina, Abrams, and Pope, “Polarization in the American

Public”; and Pietro S. Nivola and David W. Brady, “Disconnected: The Political Class Versus the People,” in Red and Blue Nation?, ed. Nivola and Brady.

78 Morris P. Fiorina, Samuel J. Abrams, and Jeremy C. Pope, Culture War? The Myth of a

Polarized America (New York: Pearson-Longman, 2010). Also see a constant stream of surveys such as: Newport, “U.S. Opinion and the Election.”

79 Lilliana Mason, “A Cross-Cutting Calm: How Social Sorting Drives Affective Polarization,” Public Opinion Quarterly 80, no. S1 (2016): 351–377,; Lilliana Mason, Uncivil Agreement:How Politics Became Our Identity (Chicago: University of Chicago Press, 2018); Murat Somer and Jennifer McCoy, “Déjà Vu? Polarization and Endangered Democracies in the 21st Century,” American Behavioral Scientist 62, no. 1 (2018):; Jennifer McCoy, Tahmina Rahman, and Murat Somer, “Polarization and the Global Crisis of Democracy: Common Patterns, Dynamics, and Pernicious Consequences for Democratic Polities,” American Behavioral Scientist 62, no. 1 (2018):; Jennifer McCoy and Murat Somer, “Toward a Theory of Pernicious Polarization and How It Harms Democracies: Comparative Evidence and Possible Remedies,” The ANNALS of the American Academy of Political and Social Science 681, no. 1 (2018):; Shanto Iyengar et al., “The Origins and Consequences of Affective Polarization in the United States,” Annual Review of Political Science 22 (2019): 129­–146,; and Eli Finkel et al., “Political Sectarianism in America,” Science 370, no. 6516 (2020): 533–536,

80 Iyengar, Sood, and Lelkes, “Affect, Not Ideology”; Liliana Mason, “I Disrespectfully Agree: The Differential Effects of Partisan Sorting on Social and Issue Polarization,” American Journal of Political Science 59 (March 2014):; and Lilliana Mason, “Ideologues Without Issues: The Polarizing Consequences of Ideological Identities,” Public Opinion Quarterly 82, no. S1 (2018): 866–887,

81 Mason, “Ideologues Without Issues.”

82 Matthew Levendusky and Neil Malhotra, “Does Media Coverage of Partisan Polarization Affect Political Attitudes?,” Political Communication 33, no. 2 (2016):

83 Eitan Hersh and Yair Ghitza, “Mixed-Partisan Households and Electoral Participation in the United States,” PLoS ONE 13, no. 10 (2018):

84 Studies found that both employers and employees would discriminate in favor of copartisans. See Karen Gift and Thomas Gift, “Does Politics Influence Hiring? Evidence from a

Randomized Experiment,” Political Behavior 37 (2015): 653–675,; and Christopher McConnell et al., “The Economic Consequences of Partisanship in a Polarized Era,” American Journal of Political Science 62, no. 1 (January 2018): 5–18,

85 Eitan D. Hersh and Matthew N. Goldenberg, “Democratic and Republican Physicians Provide Different Care on Politicized Health Issues,” Proceedings of the National Academy of Sciences 113, no. 42 (2016): 11,811–11,816,; and James N. Druckman et al., “Affective Polarization, Local Contexts and Public Opinion in America,” Nature Human Behaviour 5 (2020): 28–38,

86 “America’s Divided Mind: Understanding the Psychology That Drives Us Apart,” Beyond Conflict.

87 Devin McCarthy, “Partisanship vs. Principle: Understanding Public Opinion on Same-Day Registration,” Public Opinion Quarterly 83, no. 3 (Fall 2019): 568–583,

88 Milan W. Svolik, “Polarization Versus Democracy,” Journal of Democracy 30, no. 3 (2019): 20–32,; and Milan W. Svolik, “When Polarization Trumps Civic Virtue: Partisan Conflict and the Subversion of Democracy by Incumbents,” Quarterly Journal of Political Science 15, no. 1 (2020): 3–31,

89 Matthew H. Graham and Milan W. Svolik, “Democracy in America? Partisanship, Polarization, and the Robustness of Support for Democracy in the United States,” American Political Science Review 114, no. 2 (2020): 392–409,; Braley et al., “The Subversion Dilemma.”

90 Braley et al., “The Subversion Dilemma.”

91 Boxell, Getzkow, and Shapiro, “Cross-Country Trends in Affective Polarization.”

92 Iyengar, Sood, and Lelkes, “Affect, Not Ideology”; and Boxell, Gentzkow, and Shapiro, “Greater Internet Use Is Not Associated With Faster Growth in Political Polarization Among US Demographic Groups.”

93 Shanto Iyengar et al., “The Origins and Consequences of Affective Polarization in the United States,” Annual Review of Political Science 22 (2019): 129­–146,

94 Yphtach Lelkes, “Mass Polarization: Manifestations and Measurements,” Public Opinion Quarterly 80, no. S1 (2016): 392–410,

95 Delia Baldassarri and Barum Park, “Was There a Culture War?” Journal of Politics 82 (July 2020):; Griffin et al., “Racing Apart: Partisan Shifts on Racial Attitudes Over the Last Decade”; and Tesler, “Republican Views On Immigration Are Shifting Even Further To The Right Under Biden.”

96 Anthony Fowler et al., “Moderates,” American Political Science Review 117, no. 2 (2022):

97 Iyengar, Sood, and Lelkes, “Affect, Not Ideology”; Noam Gidron, James Adams, and Will Horne, “Toward a Comparative Research Agenda on Affective Polarization in Mass Publics,” APSA Comparative Politics Newsletter 29 (2019): 30–36,; Erik Knudsen, “Affective Polarization in Multiparty Systems? Comparing Affective Polarization Towards Voters and Parties in Norway and the United States,” Scandinavian Political Studies 44, no. 1 (2021): 34–44,; Garzia, Ferreira da Silva, and Maye, “Affective Polarization in Comparative and Longitudinal Perspective”; Gidron, Adams, and Horne, “American Affective Polarization in Comparative Perspective”; Andres Reiljan, “’Fear and Loathing Across Party Lines’ (Also) In Europe: Affective Polarisation in European Party Systems,” European Journal of Political Research 59, no. 2 (May 2020): 376–396,; Boxell, Getzkow, and Shapiro, “Cross-Country Trends in Affective Polarization”; Diego Garzia and Frederico Ferreira da Silva, “Negativity and Political Behavior: A Theoretical Framework for the Analysis of Negative Noting in Contemporary Democracies,” Political Studies Review 20, no. 2 (2022): 282–291,; and Wagner, “Affective Polarization in Multiparty Systems.”

98 McCoy and Somer, “Toward a Theory of Pernicious Polarization and How It Harms Democracies.” Unlike the scholars using self-reported feelings of warmth or coldness toward the other party, McCoy and her coauthor use Varieties of Democracy (V-Dem) data, which is coded by country experts, rather than survey instruments that directly measure reported emotions. V-Dem data may be subject to a presentism bias because the measure for affective polarization was only created a few years ago, after which experts were asked to use that criteria to backdate each country: that is, “Is society polarized into antagonistic political camps in 1950? In 1951?” But it is also quite possible that McCoy’s findings are correct, as they are based in both emotions and political incentives, and common scholarly wisdom based only on emotions is wrong, as it has been frequently in this field.

99 This large literature is summarized in Mason, Uncivil Agreement.

100 Mason, Uncivil Agreement; and Lelkes, “Mass Polarization: Manifestations and Measurements.”

101 Mason, Uncivil Agreement; Matthew Levendusky, The Partisan Sort: How Liberals Became Democrats and Conservatives Became Republicans (Chicago: University of Chicago Press, 2009).

102 Mason, Uncivil Agreement; and Lelkes, “Mass Polarization: Manifestations and Measurements.”

103 James G. Gimpel and Iris S. Hui, “Seeking Politically Compatible Neighbors? The

Role of Neighborhood Partisan Composition in Residential Sorting,” Political Geography 48 (September 2015): 130–142,

104 Bill Bishop, The Big Sort: Why the Clustering of Like-Minded America is Tearing Us Apart (New York: Houghton Mifflin Harcourt, 2009).

105 Philip A. Klinkner, “Red and Blue Scare: The Continuing Diversity of the American Electoral Landscape,” The Forum 2 (2004):; and Jonathan Mummolo and Clayton Nall, “Why Partisans Do Not Sort: The Constraints on Political Segregation,” Journal of Politics 79, no. 1 (January 2017):

106 Yphtach Lelkes, Gaurav Sood, and Shanto Iyengar, “The Hostile Audience: The

Effect of Access to Broadband Internet on Partisan Affect,” American Journal of Political

Science 61, no. 1 (January 2017): 5–20,; Matthew Levendusky, “Partisan Media Exposure and Attitudes Toward the Opposition,” Political Communication 30, no. 4 (2013):; Natalie J. Stroud, “Polarization and Partisan Selective Exposure,” Journal of Communication 60, no. 3 (September 2010):; Petter Törnberg et al., “Modeling the Emergence of Affective Polarization in the Social Media Society,” PLoS ONE 16 (2021):; and Jonathan Haidt, “Why the Past 10 Years of American Life Have Been Uniquely Stupid,” The Atlantic, May 2022 issue, April 11, 2022,

107 Charlotte Hu, “Social Media Really Is Making Us More Morally Outraged,” Popular Science, August 13, 2021,; and Jeremy B. Merrill and Will Oremus, “Five Points for Anger, One for a ‘Like’: How Facebook’s Formula Fostered Rage and Misinformation,” Washington Post, October 26, 2021,

108 Oussema Othmeni and Lei Nguyen, “The Rise of Digital Extremism: How Social Media Eroded America’s Political Stability,” IVolunteer International, June 25, 2021,

109 For a comprehensive review of this entire literature, see: Jonathan Haidt and Chris Bail, “Social Media and Political Dysfunction: A Collaborative Review,” unpublished manuscript,

110Jeff Tolldefson, “Tweaking Facebook Feeds Is No Easy Fix for Polarization, Studies Find,” Nature, July 27, 2023, For underlying sources, see Sandra González-Bailón et al., “Asymmetric Ideological Segregation in Exposure to Political News on Facebook,” Science 381, no. 6656 (2023): 392–398,; Andrew M. Guess et al, “How Do Social Media Feed Algorithms Affect Attitudes and Behavior in an Election Campaign?,” Science 381, no. 6656 (July 2023): 398–404,; Brendan Nyhan et al., “Like-Minded Sources on Facebook Are Prevalent but Not Polarizing,” Nature 620 (July 2023): 137–144,; and Andrew M. Guess, “Reshares on Social Media Amplify Political News but Do Not Detectably Affect Beliefs or Opinions," Science 381, no. 6656 (2023): 404–408,

111 Emily Kubin and Christian Von Sikorski, “The Role of (Social) Media in Political Polarization: A Systematic Review,” Annals of the International Communication Association 45, no. 3 (2021): 188–206, Twitter users are more educated and politically engaged; moreover, the median American is not on Twitter nor is it the platform of choice for organizing more violent Americans.

112 Yanna Krupnikov and John Barry Ryan, The Other Divide: Polarization and Disengagement in American Politics, (Cambridge: Cambridge University Press, 2022).

113 Boxell, Gentzkow, and Shapiro, “Cross-Country Trends in Affective Polarization”; and Kimball, “U.S. Is Polarizing Faster than Other Democracies, Study Finds.”

114 Boxell, Gentzkow, and Shapiro, “Cross-Country Trends in Affective Polarization.”

115 Hunt Alcott et al., “The Welfare Effects of Social Media,” American Economic Review 110, no. 3 (March 2020): 629–676,

116 Boxell, Gentzkow, and Shapiro, “Greater Internet Use Is Not

Associated With Faster Growth in Political Polarization Among US Demographic Groups”; and

“The Nielsen Total Audience Report Q3 2018,” Nielsen, 2019,

117 Andrew Guess, Brendan Nyhan, and Jason Reifler, “Selective Exposure to Misinformation: Evidence From the Consumption of Fake News During the 2016 Presidential Campaign,” European Research Council, January 9, 2018,

118 “Ahead of Anniversary of 1/6 Insurrection, Republicans Remain Entangled in the Big Lie, Qanon, and Temptations Toward Political Violence,” Public Religion Research Institute, January 5, 2022, See also Gregory J. Martin and Ali Yurukoglu, “Bias in Cable News: Persuasion and Polarization,” American Economic Review 107, no. 9 (2017): 2,565–2,599,

119 Levendusky, “Partisan Media Exposure and Attitudes Toward the Opposition.”

120 Robert A. Pape, “The Jan. 6 Insurrectionists Aren’t Who You Think They Are,” Foreign Policy, January 6, 2022,

121 David E. Broockman and Joshua L. Kalla, “The Impacts of Selective Partisan Media Exposure: A Field Experiment With Fox News Viewers,” OSF Preprints, September 1, 2022, While not about polarization per se, a 2007 study found that the emergence of Fox News (which launched in 1996, long after American polarization had begun) influenced between 3 and 8 percent of non-Republicans to vote for Republican candidates. Stefano DellaVigna and Ethan Kaplan, “The Fox News Effect: Media Bias and Voting,” Quarterly Journal of Economics 122, no. 3 (2007): 1,187–1,234,

122 Kevin Arceneaux and Martin Johnson, Changing Minds or Changing Channels?: Partisan News in an Age of Choice (Chicago: University of Chicago Press, 2013).

123 “Sacking Tucker Carlson Has Put a Dent in Fox News Ratings,” Economist, May 16, 2023,

124 Annie Y. Chen et al., “Subscriptions and External Links Help Drive Resentful Users to Alternative and Extremist YouTube Channels,” manuscript submitted to arXiv, revised April 2, 2023,

125 Richard Fletcher and Joy Jenkins, “Polarisation and the News Media in Europe,” European Parliamentary Research Service, March 2019,

126 “The BBC Has Bigger Problems Than a Misbehaving Interviewer,” Economist, May 27, 2021,; and Mark Jurowitz et al., “U.S. Media Polarization and the 2020 Election: A Nation Divided,” Pew Research Center, January 24, 2020,

127 Kimball, “U.S. Is Polarizing Faster than Other Democracies, Study Finds.”

128 Joshua P. Darr, Matthew P. Hitt, and Johanna L. Dunaway, “Newspaper Closures Polarize Voting Behavior,” Journal of Communication 68 (2018): 1,007–1,028,

129 Danny Hayes and Jennifer L. Lawless, “The Decline of Local News and Its Effects: New Evidence From Longitudinal Data,” Journal of Politics 80, no. 1 (January 2018):

130 Gidron, Adams, and Horne, “American Affective Polarization in Comparative Perspective”; and Knudsen, “Affective Polarization in Multiparty Systems?”

131 Garzia, Ferreira da Silva, and Maye, “Affective Polarization in Comparative and Longitudinal Perspective”; and “The Wealth Inequality Pandemic: COVID and Wealth Inequality. Build Back Fairer, Report 4,” Australian Council of Social Service, July 2022,

132 Luca Bettarelli and Emilie Van Haute, “Regional Inequalities As Drivers of Affective Polarization,” Regional Studies, Regional Science 9, no. 1 (2022): 549–570,

133 Alexander J. Stewart, Nolan McCarty, and Joanna J. Bryson, “Polarization Under Rising Inequality and Economic Decline,” Science Advances 6, no. 50 (December 11, 2020): Two of these scholars joined a third to publish very similar data in a different article; see Alexander J. Stewart, Joshua B. Plotkin, and Nolan McCarty, “Inequality, Identity, and Partisanship: How Redistribution Can Stem the Tide of Mass Polarization,” Proceedings of the National Academy of Sciences 118, no. 50 (2021):

134 Ira Kalish, Michael Wolf, and Jonathan Holdowsky, “The Link Between Trust and Economic Prosperity,” Deloitte, May 20, 2021,

135 Atif Mian, Amir Sufi, and Francesco Trebbi, “Resolving Debt Overhang: Political Constraints in the Aftermath of Financial Crises,” American Economic Journal: Macroeconomics 6, no. 2 (2014): 1–28,

136 Michael Podhorzer, “Socioeconomic Polarization,” Weekend Reading, May 15, 2022, private subscriber-based newsletter.

137 “U.S. Employment in Metro and Nonmetro Areas, 2007-21,” USDA, Economic Research Service, using data from the Department of Labor, Bureau of Labor Statistics, Local Area Unemployment Statistics Program, released April 15, 2022,

138 David Autor, David Dorn, and Gordon H. Hanson, “On the Persistence of the China Shock,” National Bureau of Economic Research, Working Paper no. 29401, October 2021,

139 David Autor et al., “Importing Political Polarization? The Electoral Consequences of Rising Trade Exposure,” American Economic Review 110, no. 10 (October 2020): 3,139–3,183,

140 Kirill Zhirkov, “The Origins and Consequences of Racialized Schemas About US Parties,” Journal of Race, Ethnicity, and Politics 7, no. 3 (November 2022): 484–504,; and Peter K. Enns and Ashley Jardina, “Complicating the Role of White Racial Attitudes and Anti-Immigrant Sentiment in the 2016 US Presidential Election,” Public Opinion Quarterly 85, no. 2 (2021): 539–570,

141 Enns and Jardina, “Complicating the Role of White Racial Attitudes and Anti-Immigrant Sentiment in the 2016 US Presidential Election.”

142 Kalmoe and Mason, Radical American Partisanship, 105, 109.

143 Tyler T. Reny, Loren Collingwood, and Ali Valenzuela, “Vote Switching in the 2016 Election: How Racial and Immigration Attitudes, Not Economics, Explain Shifts in White Voting,” Public Opinion Quarterly 83, no. 1 (2019): 91–113,; John Sides, Michael Tesler, and Lynn Vavreck, Identity Crisis: The 2016 Presidential Campaign and the Battle for the Meaning of America (Princeton: Princeton University Press, 2018); Lee Drutman, “How Race and Identity Became the Central Dividing Line in American Politics,” Vox, August 30, 2016,; German Lopez, “The Past Year of Research Has Made it Very Clear: Trump Won Because of Racial Resentment,” Vox, December 15, 2017,; and Thomas Wood, “Racism Motivated Trump Voters More Than Authoritarianism,” Washington Post, April 17, 2017,

144 Matthew Yglesias, “The Great Awokening,” Vox, April 1, 2019,; Griffin et al., “Racing Apart: Partisan Shifts on Racial Attitudes Over the Last Decade”; and “Deep Divisions in Americans’ View of Nation’s Racial History—And How to Address It,” Pew Research Center, August 12, 2021,

145 Amy R. Krosch and David M. Amodio, “Economic Scarcity Alters the Perception of Race,” PNAS 111, no. 25 (June 9, 2014):; and Amy R. Krosch, Tom R. Yyler, and David M Amodio, “Race and Recession: Effects of Economic Scarcity on Racial Discrimination,” Journal of Personal Social Psychology 113, no. 6 (December 2017):

146 Christopher Vito, Amanda Admire, and Elizabeth Hughes, “Masculinity, Aggrieved Entitlement, and Violence: Considering the Isla Vista Mass Shooting,” International Journal for Masculinity Studies 13, no. 2 (2018): 86–102,

147 Marcel Lubbers, Mérove Gijsberts, and Peer Scheepers, “Extreme Right-Wing Voting in Western Europe,” European Journal of Political Research 41 (2002): 345–378,; Duane H. Swank and Hank-Georg Betz, “Globalization, the Welfare State and Right-Wing Populism in Western Europe,” Socio-Economic Review 1 (2003): 215–245,; Kai Arzheimer, “Contextual Factors and the Extreme Right Vote in Western Europe, 1980–2002,” American Journal of Political Science 53, no. 2 (2009): 259–275,; Matt Golder, “Explaining Variation in the Success of Extreme Right Parties in Western Europe,” Comparative Political Studies 36, no. 4 (2003):; and Nonna Mayer, “The Political Impact of Social Insecurity in France,” Partecipazione e conflitto 11 (2019):

148 Matthijs Rooduijn and Brian Burgoon, “The Paradox of Well-Being: Do Unfavorable Socioeconomic and Sociocultural Contexts Deepen or Dampen Radical Left and Right Voting Among the Less Well-Off?,” Comparative Political Studies 51, no. 13 (2017):; and Cas Mudde, On Extremism and Democracy in Europe (New York: Routledge, 2017). For instance, in Eastern Germany where the right-wing Alternative for Germany party has been growing, wages have risen from 1991 levels of 50 percent of wages in the West to 90 percent when adjusted for cost of living. So people in Eastern Germany are doing far better than in the past but still not as well as their relatively new countrymen to the West. See also Beverely Crawford, “It’s the German Economy, Stupid! Economic Inequality, Not Immigration, Explains Far Right Rise in Germany,” UC Berkeley Politics and Law Blog, September 27, 2017,; and “Deutschland AG Worries About the Rise of the Hard-Right AFD,” Economist, September 5, 2019, For employment and economic growth statistics in these countries and regions, see the European Commission’s EURES Labor Market Information statistics: “Labour Market Information,” EURES, European Commission, updated March 29, 2023,

149 Arie Perliger, “Challengers From the Sidelines: Understanding America’s Violent Far-Right,” Combating Terrorism Center at West Point, November 2012,

150 Gary LaFree and Bianca E. Bersani, “County-Level Correlates of Terrorist Attacks in the United States,” Criminology & Public Policy 13, no. 3 (August 2014): 455–481,

151 Limor Glutchin and Jordan Kyle, “Populists in Power Around the World,” Tony Blair Institute, November 7, 2018,

152 Thomas Carothers and Benjamin Press, “Understanding and Responding to Global Democratic Backsliding,” Carnegie Endowment for International Peace, October 20, 2022,

153 Bingham G. Powell, “Party Systems and Political System Performance: Voting Participation, Government Stability and Mass Violence in Contemporary Democracies,” American Political Science Review 75, no. 4 (1981): 861–879,; Hanne Fjelde and Kristine Höglund, “Electoral Institutions and Electoral Violence in Sub-Saharan Africa,” British Journal of Political Science 46, no. 2 (2016): 297–320,; and Lee Drutman, Breaking the Two-Party Doom Loop: The Case for Multiparty Democracy in America (New York: Oxford University Press, 2020).

154 Gidron, Adams, and Horne, “American Affective Polarization in Comparative Perspective”; and Matthew S. Levedunsky, “Americans, Not Partisans: Can Priming American National Identity Reduce Affective Polarization?,” Journal of Politics 80, no. 1 (January 2018):

155 For instance, see Drutman, Breaking the Two-Party Doom Loop.

156 Gaurav Sood and Shanto Iyengar, “Coming to Dislike Your Opponents: The Polarizing Impact of Political Campaigns,” SSRN Electronic Journal, September 17, 2016,

157 Kirill Zhirkov and Nicholas A. Valentino, “The Origins and Consequences of Racialized Schemas About US Parties,” Journal of Race, Ethnicity, and Politics 7, no. 3 (2022):; and Douglas J. Ahler and Gaurav Sood, “The Parties in Our Heads: Misperceptions About Party Composition and Their Consequences,” Journal of Politics 80, no. 3 (July 2018):

158 Douglas J. Ahler and Gaurav Sood, “The Parties in Our Heads: Misperceptions About Party Composition and Their Consequences,” Journal of Politics 80, no. 3 (July 2018):

159 Levendusky and Malhotra, “(Mis)perceptions of Partisan Polarization in the American Public”; and Westwood and Peterson, “The Inseparability of Race and Partisanship in the United States.”

160 Lilla V. Orr and Gregory A. Huber, “The Policy Basis of Measured Partisan Animosity in the United States,” American Journal of Political Science 64, no. 3 (2019): 569–586,

161 Orr and Huber, “The Policy Basis of Measured Partisan Animosity in the United States.”

162 Voelkel et al., “Megastudy Identifying Effective Interventions to Strengthen Americans’ Democratic Attitudes.”

163 Levendusky, “Americans, Not Partisans”; Levendusky, Our Common Bonds; and Ryan E. Carlin and Gregory Love, “Political Competition, Partisanship, and Interpersonal Trust in Electoral Democracies,” British Journal of Political Science 48, no. 1 (2018): 115–139,

164 Mark Levine et al., “Identity and Emergency Intervention: How Social Group Membership and Inclusiveness of Group Boundaries Shape Helping Behavior,” Personality and Social Psychology Bulletin 31, no. 4 (2005):; Jason A. Nier et al., “Changing Interracial Evaluations and Behavior: The Effects of a Common Group Identity,” Group Processes & Intergroup Relations 4, no. 4 (2001):; and Muzafer Sherif, “Superordinate Goals in the Reduction of Intergroup Conflict,” American Journal of Sociology 63, no. 4 (January 1958): 349–356,

165 Demis E. Glasford and John F. Dovidio, “E Pluribus Unum: Dual Identity and Minority Group Members’ Motivation to Engage in Contact, as Well as Social Change,Journal of Experimental Social Psychology 47, no. 5 (September 2011): 1,021–1,024,

166 Dean Scheepers et al., “A Shared Dual Identity Promotes a Cardiovascular Challenge Response During Interethnic Interactions,” Group Processes & Intergroup Relations 17, no. 3 (2014):

167 John F. Dovidio, Samuel L. Gaertner, and Tamar Saguy, “Another View of ‘We’: Majority and Minority Group Perspectives on a Common Ingroup Identity,” European Review of Social Psychology 18, no. 1 (2007):

168 Ryan Dawkins and Abigail Hanson, “‘American’ Is the Eye of the Beholder: American Identity, Racial Sorting, and Affective Polarization Among White Americans,” Political Behavior (2022):

169 Glasford and Dovidio, “E Pluribus Unum.”

170 Jillian C. Banfield and John F. Dovidio, “Whites’ Perceptions of Discrimination Against Blacks: The Influence of Common Identity,” Journal of Experimental Social Psychology 49, no. 5 (September 2013): 833–841,; Yuanyuan Shi et al., “Dual Identity and Prejudice: The Moderating Role of Group Boundary Permeability,” Frontiers in Psychology 8 (2017):; and Rita Guerra et al., “How Should Intergroup Contact Be Structured to Reduce Bias Among Majority and Minority Group Children?,” Group Processes & Intergroup Relations 13, no. 4 (2010):

171 Community resilience programs can look similar to bridging programs but are fundamentally different. They bring people together to do something, not to talk about something. They create relationships that neither are specifically about the divide nor ignore the divide; they are not about simply showing up in one’s identity on either side of the divide. Instead, individuals show up as people with agency in a variety of ways who can move portions of their community. They create a shared identity, which can be dual (see earlier discussion). If these programs succeed in creating authentic relationships across groups (known as “bridging social capital”), they can be useful for tamping down tensions that arise within a community. For instance, if a politician is attempting to gin up anti-immigrant feeling, or if a hate crime has been committed, groups that build community resilience can work across very different parts of the community quickly to reduce anger, devise common methods of showing solidarity, and keep problems from spinning out of control.

172 Paluck et al., “Prejudice Reduction.”

173 Paluck et al., “Prejudice Reduction.”

174 Michael Tomasky, If We Can Keep It: How the Republic Collapsed and How It Might Be Saved (New York: Liveright Publishing Corporation, 2019), 200.

175 Elaine C. Kamarck, “Solutions to Polarization,” in Solutions to Political Polarization in America, ed. Nathaniel Persily (Cambridge: Cambridge University Press, 2015), 96–103; Nathaniel Persily, “Stronger Parties as a Solution to Polarization,” in Solutions to Political Polarization in America, ed. Persily, 123–135; and Steven S. Smith, “Partisan Polarization and the Senate Syndrome,” in Solutions to Political Polarization in America, ed. Persily, 218–229.

176 Arend Lijphart, “Polarization and Democratization,” in Solutions to Political Polarization in America, ed. Persily, 73–82; and Todd Donovan, Caroline Tolbert, and Kellen Gracey, “Campaign Civility Under Preferential and Plurality Voting,” Electoral Studies 42 (June 2016): 157–163,

177 Rachel Kleinfeld, “The Rise of Political Violence in the United States,” Journal of Democracy 32, no. 4 (2021): 160–176,; and Lee Drutman, “The Case for Proportional Voting,” National Affairs 56 (2018):

178 Voelkel et al., “Interventions Reducing Affective Polarization Do Not Necessarily Improve Anti-Democratic Attitudes”; and Broockman, Kalla, and Westwood, “Does Affective Polarization Undermine Democratic Norms or Accountability? Maybe Not.”

179 The Strengthening Democracy Challenge brought together cross-disciplinary researchers across top-tier universities to evaluate 250 crowdsourced proposals that were supposed to reduce antidemocratic attitudes, political violence, or affective polarization. They looked only at short, scalable interventions and, after culling to a top twenty-five, tested them on more than 30,000 online participants who leaned toward one or the other party. See Voelkel et al., “Megastudy Identifying Effective Interventions to Strengthen Americans’ Democratic Attitudes.”

180 Voelkel et al., “Megastudy Identifying Effective Interventions to Strengthen Americans’ Democratic Attitudes.”.

181 Voelkel et al., “Interventions Reducing Affective Polarization Do Not Necessarily Improve Anti-Democratic Attitudes.”

182 Broockman, Kalla, and Westwood, “Does Affective Polarization Undermine Democratic Norms or Accountability? Maybe Not.”

183 Voelkel et al., “Megastudy Identifying Effective Interventions to Strengthen Americans’ Democratic Attitudes.”

184 Voelkel et al., “Interventions Reducing Affective Polarization Do Not Necessarily Improve Anti-Democratic Attitudes.”

185 In the case of intergroup contact bridging programs, it is unlikely to be reaching individuals vulnerable to messages of violence or promoting such measures, who are unlikely to self-select into programs to learn more about the so-called other side.

186 Voelkel et al., “Megastudy Identifying Effective Interventions to Strengthen Americans’ Democratic Attitudes”; and Braley et al., “The Subversion Dilemma.”

187 Voelkel et al., “Megastudy Identifying Effective Interventions to Strengthen Americans’ Democratic Attitudes”; and Braley et al., “The Subversion Dilemma.”

188 Graham and Svolik, “Democracy in America? Partisanship, Polarization, and the Robustness of Support for Democracy in the United States.”

189 Kalmoe and Mason, Radical American Partisanship, 88.

190 Voelkel et al., “Megastudy Identifying Effective Interventions to Strengthen Americans’ Democratic Attitudes”; and Kalmoe and Mason, Radical American Partisanship.

191 “Federal Bureau of Investigation Crime Data Explorer,” Federal Bureau of Investigation, accessed July 11, 2023, The FBI’s hate crimes data are notoriously poor because they rely on police agencies to self-report, and many do not. However, more departments have failed to report more recently, so these trends are likely to be even worse than the FBI statistics suggest, not better.

192 “USCP Threat Assessment Cases for 2022,” press release, United States Capitol Police, January 17, 2023,

193 “New Report: Harassment, Threats and Violence Directed at Local Elected Officials Rising at an Alarming Rate,” press release, National League of Cities, June 24, 2022,; Linda So and Jason Szep, “Reuters Unmasks Trump Supporters Terrifying U.S. Election Officials,” Reuters, November 9, 2021,; Lawrence Norden et al., “Local Election Officials Survey (March 2022),” Brennan Center for Justice, March 10, 2023.; and Paul Gronke et al., “Understanding the Career Journeys of Today’s Local Election Officials and Anticipating Tomorrow’s Potential Shortage,” Democracy Fund, April 20, 2021, Bridging Divides Initiative/CivicPulse data will eventually provide the first truly comparable data on threats and harassment against election officials and local officials, but the data are not retrospective so at present only a snapshot of this problem exists for the first wave of their study, released on January 2023. See

194 Kalmoe and Mason, Radical American Partisanship.

195 Kalmoe and Mason, Radical American Partisanship.

196 Thomas H. Costello et al., “Clarifying the Structure and Nature of Left-Wing Authoritarianism,” Journal of Personality and Social Psychology 122, no. 1 (January 2022): 135–170, Costello et al. do find differences between left- and right- leaning authoritarians. Notably, on the right, authoritarians tend to display more cognitive rigidity and less openness to new experiences. On the left, authoritarians are more cognitively flexible and open to experiences, but they are less good at processing stress and tend to overreact to danger (neuroticism). These different traits lead to the same place in terms of support for coercion and violence to enforce political values.

197 Michael Jensen, Elizabeth Yates, and Sheehan Kane, “Motivations and Characteristics of Hate Crime Offenders,” START, October 2020,; also see dataset “Profiles of Individual Radicalization in the United States (PIRUS),” START,; and “Countering Domestic Terrorism: Examining the Evolving Threat,” testimony of William Braniff, the director of the National Consortium for the Study of Terrorism and Responses to Terrorism, to the Committee on Homeland Security and Governmental Affairs, 116th U.S. Congress, September 25, 2019,

198 Kalmoe, “Fueling the Fire”; Donald R. Kinder and Cindy D. Kam, Us Against Them: Ethnocentric Foundations of American Opinion (Chicago: University of Chicago Press, 2010); and Brian F. Schaffner, “Follow the Racist? The Consequences of Trump’s Expressions of Prejudice for Mass Rhetoric,” unpublished manuscript, Tufts University, 2018,

199 Sreeradha Datta, “Political Violence in Bangladesh: Trends and Causes,” Strategic Analysis 29, no. 3 (2005): 427–447,; Severine Autesserre, The Trouble With the Congo: Local Violence and the Failure of International Peacebuilding (Cambridge: Cambridge University Press, 2010); Roland Paris, At War’s End: Building Peace After Conflict (Cambridge: Cambridge University Press, 2010); Mohammad Moniruzzaman, “Party Politics and Political Violence in Bangladesh: Issues, Manifestation and Consequences,” South Asian Survey 16, no. 1 (2009): 81–99,; and Matthew Ellman and Leonard Wantcheko, “Electoral Competition Under the Threat of Political Unrest,” Quarterly Journal of Economics 115, no. 2 (2000): 499–531,

200 Thomas B. Edsall, “Meet the People Working on Getting Us to Hate One Another Less,” New York Times, February 8, 2023,