The space between us threatens to swallow us whole

Social distance, affective polarization, and partisan mega-identities are helping to unravel the fabric of American democray

Americans are growing further apart, with no real stop in sight.

It is not so much our (asymmetric!) policy or ideological polarization that concerns me today. Much of the issue-based divide between parties has been overstated, at least among the voters themselves. Instead, political scientists seem to be worried about the consequences of rising sorting and polarization in our identities.

That’s because Americans’ overwhelmingly negative feelings towards members of the other party are clearly becoming more consequential. Evaluations of the out-party now shape many aspects of citizens’ day-to-day lives, the partisan gradient subsuming everything from their choices for medical care (eg during covid-19) to who parents will allow their kids to marry to belief in the core tenants of liberal democracy (eg fueling the January 6 riot). Our growing social distance is not alone in causing America’s problems, but it has laid the groundwork for these threats to democracy to exist and persist.

The resulting political factionalism, several political scientists suggest to me, in both institutions and among the masses, is the grave threat of our time.

My dispatch on social polarization from the annual meeting of the American Political Science Association held this month in Seattle is now online for The Economist and in print this weekend. I focus on several new papers and upcoming books that highlight just how dangerous our increasing political sectarianism is. One paper shows citizens were glad to divert resources meant to fight covid-19 from areas where their political opponents live. Another study suggested we might be able to decrease this distance by fostering communication between people.

But the biggest effect might come from elites. We have observed how socially sorted parties can be stoked to anger and violence by bad-faith leaders. The January 6 riot is an obvious example, but it is not the only one. Someone threw a molotov cocktail into the Travis County Democratic Headquarters in Texas last week. The Trump presidency saw multiple coordinated acts of violence against media outlets and plots to kidnap governors. But these are not automatic byproducts of social polarization. Evidence from an upcoming book written by Lilliana Mason and Nathan Kalmoe suggests the biggest chance to decrease mass support for violence might be for opinion leaders to tell their voters not to be violent.

If you have the time and access, please read the piece. It is a short Economist-y write-up of my major takeaway from the conference and a serious warning for anyone interested in the fabric of our democracy. But as it is short, much was cut out. I want to talk more on this for you all is a long subscribers-only post today, and cover a few papers I didn’t get to.

---

I. How far apart?

First, a short recap of the data and some definitions:

At APSA, scholars of American politics appeared to be dismayed most by rising levels of “affective polarization,” the political science term for the hostility one person feels towards members of the other party relative to feelings towards their own. Driven almost exclusively by a rise in “negative partisanship,” which quantifies the degree of dislike for the out-party, levels of affective polarization have risen over two-fold since the 1970s when the American National Election Studies, a quadrennial academic survey started at the University of Michigan, began asking how citizens to rate how they felt about members of either major party. In 1978, according to the survey, the difference between Americans’ ratings of members of their own and the other party on a 100-point “feeling thermometer” scale was 27 points. The gap had widened to 56 by 2020.

This increasing divide now constitutes what the political scientist Lilliana Mason has called “identity-based” polarization. In her 2018 book “Uncivil Agreement” Mason crunched a mass of survey data to reveal how ideological, religious, gender, and racial identities have become “sorted” into overlapping mega-identities associated with the words “Democrat” and “Republican.” She suggests that social polarization is on balance bad for democracy, as isolated and warring tribes have become “relatively unresponsive to changing information or real national problems.” As long as the parties remain separated conglomerates of different identities, she concludes her book, “the electorate will behave more like a pair of warring tribes than like the people of a single nation, caring for their shared future.”

Mason’s work on identity has turned out to be remarkably prescient. We can summarise that polarization poses at least two important threats to democracy today: biased information processing and violent, intolerant, warring political factions. The consequences of this dual threat have become increasingly visible to observers of American politics over the last few years. Take the two major events in US politics and society over the last year: covid-19 and the attempted right-wing insurrection against Congress on January 6th, 2021.

First, responses to the coronavirus pandemic from both political leaders and the voters themselves have been polarised along partisan lines. In the beginning, lockdown phase of the pandemic, Democratic governors were much likelier to enact strict restrictions on their states’ residents than Republican governors were. A report conducted by researchers at the University of Washington and published in March 2020 found that Republican-led states with higher concentrations of votes for Donald Trump in the 2016 election delayed social distancing restrictions up to 3 days relative to predictions that factored in the severity of the virus in the state and demographics about its residents. In a follow-up study circulated in late September 2021, the researchers found that these redder states ended regulations related to covid-19, such as business restrictions and mask mandates, nearly two weeks earlier than expected — and largely as the more-deadly Delta variant of the virus began to circulate across the country.

Among the masses, masking and social distancing were also much more common for members of the left than the right. In a survey conducted by YouGov for The Economist in March 2020, adults aligned with the Democratic Party were much likelier to say they were worried about the virus, and about twice as likely to say they had canceled travel plans or were wearing face coverings. In the recent vaccination-and-reopening stage of the virus, conservatives remain much more likely than liberals to get jabbed; another study of YouGov’s in July data revealed that Trump’s 2020 voters were about 13 percentage points less likely to be vaccinated even after controlling for relevant demographic and political factors such as their age, education, self-described ideology, or race, among other things.

Of course, peoples’ pandemic behaviors are not caused entirely by their party affiliations. Mason’s work suggests that partisan differences in masking and distancing stem from the underlying identities that lead to Americazns’ partisanship and voting behavior, not strictly the other way around. A well-educated person is more likely to believe in the science of vaccines, for example, and a socially conservative person is less trusting of the government. But hardened identities make people much less likely to adopt the positions of the opposition party once they get subsumed under the partisan divide. 

In a paper Mason presented at the APSA conference alongside Nathan Kalmoe and Julie Wronski, two of her frequent co-authors, the researchers conducted a series of experiments to gauge the extent to which Americans thought certain groups deserved life-saving aid if they were infected with covid-19, if the economy should be reopened even if it would hurt certain groups, and whether disproportionate deaths among the opposing party were deserved or not. They found that respondents were much likelier to say disproportionate deaths among the out-party were more acceptable than deaths among their own. In this way, identity-based polarization is a direct threat to the health of the country’s people.

II. Democracy for me but not for thee

Social polarization is also a threat to the country’s institutions. In a 2019 paper, Kalmoe and Mason studied the extent to which voters rationalize partisan violence against their opponents, feel pleasure for such violence, or express outright support for it—what they call “moral disengagement” and “lethal partisanship.” They found that 60% of partisans said the opposition was “a serious threat to the United States,” and 40% said opponents were “downright evil.” Further, people with stronger attachments to either political party were more likely to exhibit levels of such lethality, and between 5 and 15% of Americans in either party endorse violence depending on the circumstances. 

Mason and Kalmoe find there have been roughly equal shares of radicals in both the Democratic and Republican camps. But the authors are quick to note that one side is pushing for an expansion of democracy and inclusion of multiple racial groups in the franchise, while another is motivated by a backlash to it. And only the members of one party, fringe as they were, invaded the US Capitol building on January 6th. 

Kalmoe and Mason are updating their work on partisan violence for an upcoming book, “Radical American Partisanship.” According to new data, Americans of both parties have become more morally disengaged and more likely to endorse violence since the authors’ first study in 2019. They also find Republicans have become disproportionately more likely to hold these views. And since the 2020 election, Mason and Kalmoe write, the GOP’s voters have also become much likelier than Democrats to endorse violence against political leaders. Their work also shows that higher levels of racism and sexism among Republicans are linked to higher probabilities of supporting violence.

The book constitutes, in the authors’ words, “a vital warning that we face immense threats” for democracy in America.  If social polarization and the strength of mega-identities continue to increase, according to the theories, the efficacy of government policy that has been consumed by partisanship will decrease and the chance of violence will rise. Risks are heightened when political elites—such as presidents—call for or even vaguely threaten violence to get what they want. But the reverse is true as well; Mason and Kalmoe find, in a final set of experimental studies, that priming Republicans with the message that Trump had asked partisans to abstain from violence decreased the share of them that endorsed it, particularly as it related to violence from losing elections. Such is cause for hope among political scientists that a sustained campaign for liberal democracy, preferably from both sides of the aisle, could dampen tempers and remove electoral legitimacy from the realm of social polarization.

III. If you can keep it

How can leaders reduce affective polarization in America? Will that even solve the country’s problems? One idea is to correct the obvious potential cause of the problem—isolation from the out-party. At APSA, Erin Rossiter, a political scientist at Notre Dame, presented an analysis of hundreds of transcripts of online conversations between partisans. She found that the presence of empathetic language in these conversations caused an improvement in participants’ post-treatment evaluations of the other party. Earlier work from Rossiter has found that even imagining a conversation with an opponent could cause at least a temporary reduction in affective polarization.

But political scientists are less sure if reducing affective polarization, even on the wide scale Rossiter’s work suggests could be helpful, would directly reduce the levels of moral disengagement, partisan schadenfreude, or violent tendencies in the electorate. In a study conducted by David Broockman, Joshua Kalla, and Sean Westwood, all political scientists, survey respondents were randomly invited to participate in a “trust game” intended to decrease their feelings of negative partisanship and increase feelings of social trust, such as whether they’d be comfortable having neighbors of the opposing political party. They observed such a decrease.

But when the authors looked for effects of manipulating a decrease in social polarization on support for key democratic norms, such as investigations against corruption or stopgaps against leaders attempting to override election outcomes, they found none. Broockman et al suggest that decreasing the strength of partisan identity (how staunchly on a 3-part scale from “weak” to “strong” one associate themselves with their tribe), regardless of how they feel about the other side, could help. But they have yet to conduct a study on that. 

Others are rightly pessimistic about potential solutions. “It is hard to imagine fixing these problems because they are all so multifaceted,” Steven Webster, a professor at Indiana University who studies anger and partisanship, said, pointing to the backlash to multiracial democracy and America’s two-party system as underlying causes of strife. He joined a chorus of other scholars, including Mason and Broockman and their co-authors, in saying political leaders and partisan elites might be able to reduce support for democracy-eroding attitudes and behaviors by purging such rhetoric from their own speeches. 

Such requests are liable to fall on deaf ears, especially for those at the very top of the Republican Party. The byproduct of increasing social polarization, Mason and Kalmoe write, is that leaders will do anything they can get away with in order to win—and voters will give them wide latitude to do it.

So it is possible that decreasing social polarization among the masses could correspondingly diminish the desire for leaders to adopt anti-democratic tactics. “We do know that when people dislike the other side more, they cordon themselves off from the other side,“ said Yphtach Lelkes, a political scientist who published a paper earlier this year that found people tell survey researchers they have personality traits they associate as being traditionally Democratic or Republican. He continued:

“So I’m not ready to rule out a link between affective polarization and illiberal attitudes. I think the relationship is more circuitous than some have hypothesized, however. Social distance leads to a one-sided information environment. When that side consists of politicians and pundits who want you to support illiberal actions, such as denying that Joe Biden won the 2020 election, the result is likely increased support for such policies.”

I think that just about sums it up.

. . .

Recapping what I learned at APSA and discussing some of these papers online, I was struck by just how much research has been done on the distance between us — and how little we know about how to or the effects of shrinking it. America’s political scientists have amassed an abundance of studies on potential causes and correlations, but appear sadly short on surefire solutions.

That is OK, for now. We cannot expect social scientists to have all the answers, and high levels of partisan violence are only recent phenomena in the last fifty years of US politics, when behavioral political science and the widespread availability of scientific survey data really took off. Still, there is one obvious strategy worth trying, however: While decreasing social polarization alone will not save democracy, we probably can’t save democracy without decreasing social polarization.