Support for Political Violence Is Higher Than It Seems
But we won't know just how high until it starts.
If you think the chances of violence if Donald Trump loses the election are substantial — which I do — what would that actually look like? Another January 6 siege of the U.S. Capitol? Similar events at state capitols, but this time with more use of arms? Local officials being assaulted or murdered? Stochastic outbursts of interpersonal violence? Right-wing militias becoming larger and more visible, terrorizing people in towns across the country? Or all of the above?
We don’t know, but let’s consider this finding from a new PBS NewsHour/NPR/Marist poll:
This isn’t that far out of line with what other surveys have found on similar questions recently; for instance, this Public Religion Research Institute poll last year found that 33 percent of Republicans agreed that “because things have gotten so far off track, true American patriots may have to resort to violence in order to save our country” (13 percent of Democrats agreed). Among those who believed the 2020 election was stolen from Trump, the figure was 46 percent.
It’s even worse than it looks
One common response to these results will be, “It’s not so bad — most Americans still reject violence!” That’s the wrong way to look at it, for a couple of key reasons.
The first is any poll results on questions like this one almost certainly understate the support for violence. That’s because of social desirability bias, the tendency of poll respondents to give socially acceptable answers to certain questions rather than revealing their true feelings. Survey researchers have known about this problem for decades, and while there are sometimes ways you can mitigate it in the way you word questions, you can’t get rid of it entirely. For instance, if you ask people whether they voted in the last election, you’ll often find that about 20 percent more people say they voted than actually did vote.
That’s on a question with low moral stakes; it’s not as though admitting you didn’t bother to vote will cause anyone to gasp in horror at what a reprobate you are. On the question of violence, on the other hand, you’d expect social desirability bias to operate much more powerfully. How many people want to tell a stranger they’re talking to on the phone that they favor violence?
Since we can’t read minds, we have no idea how many people who say “No” to these questions are actually thinking “Kinda maybe, yeah.” Not only that, we have no idea how many people are sincere today when they say they wouldn’t favor violence, but if it actually started happening — and it was being undertaken by people on their side, everyone they trusted in the media was defending it, and their party’s leader was cheering it on — they would come around to a different view.
That’s why I’m not all that encouraged by some other studies that have shown that support for political violence is very low overall. Let’s take this example, a recent study testing support for various antidemocratic ideas, including the use of violence. This was one of the survey questions they used:
“(name) was convicted of murder. He was arrested by police after surveillance footage was found showing him stabbing a prominent (outparty) to death. (name) targeted the victim because he believed the victim had prevented him from voting in the last election as part of a conspiracy to stop (inparty) voters.
Do you support or oppose (name)’s actions?”
This study was carried out by a group of respected and experienced scholars; I’m sure it was done with the utmost care. You can look at the result and say, “Great, only 1.7 percent of respondents answered yes to this question.” But who in their right mind is going to tell a pollster they think stabbing someone to death is acceptable? That may be a particularly vivid example, but it illustrates the problem with using polls to assess this question.
Support for violence goes up when it actually happens, at least among some people
In the real world, there are a number of recent cases that suggest something different, even if they aren’t perfectly analogous. Kyle Rittenhouse, a teenager who took an AR-15 to Kenosha in 2020 looking for trouble at a protest against police violence, shot three people, killing two of them. The response of the right could have been “Look, he’s a dumb kid and it was a tragedy, but he was acting in self-defense.” But that was not how they responded. Instead, they celebrated him as a hero. He got a fawning interview from Tucker Carlson, an invitation to Mar-a-Lago from Donald Trump, and standing ovations at conservative conferences where he’s an invited speaker. Not in spite of the fact that he killed two people, but because he killed two people.
That uniform reaction from the Republican elite isn’t just about one case. It sends a message to the rank and file that violence is an acceptable response to your enemies, whoever they might be. Which of course brings us to Donald Trump, who has more influence over the conservative masses than anyone else. Not only has he long characterized violence against his political opponents as something desirable and thrilling (“I’d like to punch him in the face, I’ll tell ya”), he is currently attempting a rewriting of the January 6 insurrection so those who rampaged through the halls of the Capitol will be understood as noble patriots rather than the violent thugs they were.
Here are some relevant poll results on how Republicans now view January 6, which I gathered together in a recent MSNBC column:
Trump may not yet have convinced his entire party to view those criminals as warriors for justice, but he doesn’t have far to go. YouGov polls, for instance, have consistently shown at least a quarter of Republicans saying that they approve of “Trump supporters taking over the Capitol building in Washington, D.C.” on Jan. 6; at one point, it was more than a third. By a year after the insurrection, three-quarters of Republicans said they didn’t consider the events of that day “a threat to democracy.” One year ago — the last time YouGov asked this question — 60% of those who voted for Trump in 2020 said they considered the insurrection “legitimate political discourse.” Nine months ago, Monmouth University found that 51% of Republicans said Jan. 6 was “a legitimate protest.”
And in the most recent poll on these questions, CBS News found that 30% of Republicans and 43% of those who identify with the MAGA movement approved of “the actions of those who forced their way into the Capitol.” A full 66% of Republicans favored pardons for these people explicitly identified by their violent acts.
Trump rallies now begin with an announcer saying “Ladies and gentlemen, please rise for the horribly and unfairly treated January 6th hostages,” after which they play a rendition of the Star Spangled Banner sung in the D.C. jail by Trump supporters arrested on January 6. The clear message is that the insurrection was glorious and patriotic, and if you were to participate in something similar, Donald Trump would praise you, too.
What could happen in November
It’s important to keep in mind that any outbreak of violence that occurs around the election will happen in a context of intense partisan conflict in which people will be highly motivated to defend their side. If it’s produced by Trump losing the election, everyone on the right will be deluged with messages telling them that not only is a horrible crime (the theft of the election from Trump) in progress, but the result if Joe Biden stays in office will be an absolute cataclysm for them personally and everything they hold dear, the literal destruction of America in which they will quite possibly be arrested for their beliefs. They’ll be hearing that from conservative TV, conservative radio, conservative websites, Republican politicians, and of course Trump himself.
If they take those messages seriously, violence will seem like an absolutely reasonable reaction, if not the only reasonable reaction. And it won’t be necessary for the 75 million or so Trump voters to all take up arms. If even one-tenth of one percent did, you’d have an army of 75,000, which might not be enough to successfully overthrow the federal government but could certainly be enough to plunge us into something that looks a lot like civil war.
Among the other 99.9 percent of Trump voters, there would undoubtedly be many offering passive support to the violence even if they weren’t joining in themselves. “I’m not saying I was happy when they burned down the state capitol,” they might say to their buddies, “but something had to be done.”
So when you see majorities of respondents saying they don’t think violence is necessary, don’t be too comforted. The danger hasn’t gone anywhere, and it may only be growing stronger.
All it takes to start large scale mayhem is a few violent people. Why? Human psychology. The mob mentality begins when people see the opportunity to behave with total abandon. This, not huge levels of anger, is what has caused riots throughout history. It's proven by psychologists and by history itself. People behave in large groups in more extreme ways than when they're alone.