The attempted insurrection Jan. 6 at the U.S. Capitol was a shocking but arguably predictable culmination of years of increasingly incendiary rhetoric, attitudes and actions by many Republican leaders and voters. Even as the backlash grows against President Trump and the Republican members of Congress who led the effort to contest Joe Biden’s presidential victory, there are real concerns that this violent episode will not be the last — even after Trump leaves office.

Our research on what we call “radical partisanship” suggests reason to be concerned. We began this research in 2017, after violent episodes during the 2016 campaign and a documented rise in right-wing violence. We were worried that there could be more support for this violence among the public than many people realized. As political violence continued to rise, we conducted a dozen surveys mapping what Americans think about partisan violence.

Here is what we learned.

What do Americans think about violence?

The most basic finding is that a significant minority of Americans will not reject violence outright. Several of our surveys asked respondents if they believed “it is justified for [their own party] to use violence in advancing their political goals these days.” In a 2017 Cooperative Congressional Election Study, 8 percent of partisans agreed that violence is at least “a little bit” justified. In 2018, it rose to 15 percent and has hovered around there since.

We also asked whether violence would be okay if their party lost the 2020 presidential election. Across nine surveys in 2017-2019, about 20 percent said that it would be at least a little bit okay. In fall 2020, the political science project Bright Line Watch asked: Would violence be justified if opponents acted violently first? Forty percent of partisans said yes, at least a little.

These numbers represent millions of Americans. Many are thinking of punches, threats or property damage, and not just killing, but the risks remain grave. As the U.S. Capitol attack demonstrates, relatively few violent people can generate enormous danger.

These attitudes have real behavioral consequences. In a Cooperative Congressional Election Study conducted before and after the 2020 election, only 1.5 percent of respondents said they had ever hit, pushed or grabbed people in a conflict over politics. But among those who said that violence was at least a little bit okay, 9 percent reported this kind of physical confrontation. That connection from violent expressions to violent acts manifested in the U.S. Capitol assault.

What kinds of people approve of violence?

We have also asked many questions about demographics, political attitudes and psychological attributes to learn who holds these views. The best predictor of pro-violence attitudes is an aggressive personality. People who report getting into fights, arguing frequently with others and getting angry easily in everyday life were likely to approve of partisan violence. The second-best predictor was strong identification with a party.

Another predictor of pro-violent attitudes is “partisan moral disengagement,” which entails seeing the other party as evil, less than human and a serious threat to the nation. This is consistent with psychology research that has found moral disengagement to be a precursor to harming behavior.

Notably, Democrats and Republicans did not much differ on whether they viewed violence as justified. But they did differ on the reasons for supporting violence. These partisan differences reflect the importance of social and especially racial hierarchies in the United States’ long history of partisan violence. Democrats tend to support violence on behalf of inclusive democracy and civic equality, while Republicans support violence in defense of the traditional social hierarchy, in which White men retain disproportionate status and power.

Evidence for this difference in motives is visible in the factors linked to partisan moral disengagement — and how those linkages differ for Democrats and Republicans. Partisan moral disengagement is particularly prevalent among Republicans who express more resentment toward African Americans and more sexist attitudes toward women.

This is not true for Democrats who morally distance themselves from Republicans. For Democrats, expressing lower levels of racism is associated with moral disengagement from Republicans. Sexism has no relationship with partisan moral disengagement among Democrats.

Notably, partisan opponents are not the only target for violence. In our 2017 survey, we found that strongly attached partisans are more violent toward partisan opponents and to apostates within their party. This helps explain why the U.S. Capitol attackers included “disloyal” Republicans as targets of wrath, and some Republicans on social media called for Vice President Pence to be executed for not supporting Trump’s election rejection strongly enough.

Support for violence is linked to belief in conspiracy theories

In a survey we conducted after the election, Republican falsehoods about the election were also linked to support for post-election violence. After Biden was announced president-elect in November, we asked Republicans whether they would support various ways of rejecting Biden’s presidency. Fourteen percent of Republicans said governors should call up the National Guard to resist federal orders. Nine percent said the military should overthrow the new president. And an astounding 25 percent said citizens should prepare weapons to resist the federal government.

Most alarming, Republicans who believe Democrats cheated in the election (83 percent in our study) were far likelier to endorse post-election violence, even after accounting for traits such as partisanship and aggression. Although we can’t be certain about cause and effect, this establishes a clear link. As Voltaire said, “Those who can make you believe absurdities can make you commit atrocities.”

What can we do about it?

To reduce the threat of violence and accompanying risks to democracy, political leaders have an important role. Historical and cross-national evidence shows leaders are key to regulating violence in their followers. In the wake of last week’s violence at the Capitol, the onus falls on Republican leaders in particular to renounce violence and the false conspiracy theories about the 2020 election.

Ordinary citizens have an important role, too. Interpersonal conversations are among the strongest influences on political views and actions. Thus, the more Americans renounce violence in conversations with family members, friends and neighbors, the more likely people will consider violence an illegitimate political strategy.

Most broadly, the United States is at an inflection point in a long conflict between democratization and authoritarianism. The insurrection on Jan. 6 was a backlash against the forces of democratization, organized along partisan lines. Ensuring robust American democracy requires all political leaders — and at this moment, especially Republican leaders — to renew their commitment to peaceful resolution of political differences.

Lilliana Mason (@LilyMasonPhD) is associate professor of government and politics at the University of Maryland and author of “Uncivil Agreement: How Politics Became our Identity” (University of Chicago Press, 2018).

Nathan P. Kalmoe (@NathanKalmoe) is associate professor of political communication at Louisiana State University and author of “With Ballots & Bullets: Partisanship & Violence in the American Civil War” (Cambridge University Press, 2020).