It’s too early to say who will win the 2020 presidential election, but there’s a good chance that one loser will be faith in the electoral system. President Donald Trump is alleging—as he did four years ago, though sooner in the cycle this time and with greater vehemence but no more evidence—that the voting system is subject to widespread fraud. Should he win, it will give him another four years to undermine the system from within and assail voting-rights protections. Should he lose, his defeat is likely to reinforce the spurious claims of fraud he is spreading now.
According to new research, unfounded claims of fraud from Trump and his allies significantly undermine faith in the American election system, especially among voters who support him. Worse, the damage seems to be resistant to repair by fact-checking.
“It may be that confidence in the election system is a soft target,” Brendan Nyhan, a political-science professor at Dartmouth and one of the authors of the new paper, told me. “It’s complicated, hard to observe, unintuitive, and relies on trust. Trust in institutions seems to be easier to destroy than to build.”
Trump’s baseless warning of a “rigged” election was a hallmark of the closing weeks of the 2016 campaign, as polls pointed to a Hillary Clinton victory. The Republican candidate wouldn’t commit to accepting the results if Clinton won. This time around, Trump has taken up the cry sooner in the campaign. Once again, polls show him trailing; this time, he is furious over pandemic-driven efforts to expand access to voting by mail in states where Democrats are in control, or where he seems to believe higher turnout would hurt him in November. (Trump shows conspicuously less worry about vote-by-mail in GOP-friendly states.)
Trump has made some of his most incendiary claims on Twitter, and the social-media site drew his fury for appending to his tweets extremely mild language offering readers a fact-check. But that approach, the new research suggests, may not be much help here.
To test the effect of statements such as Trump’s, an interdisciplinary team of researchers showed research subjects statements from Trump and other GOP politicians and commentators alleging fraud in elections, either in small or great amounts. The results were distressing, if not altogether surprising: Republicans, as well as independents, saw their faith in the election system decrease. (Views among Democrats did not meaningfully change.) The effect was especially pronounced when subjects were split between approving or disapproving of Trump.
Then the researchers showed some subjects a different set of tweets that fact-checked the politicians’ claims. (They showed tweets, not articles, hoping to replicate the condition in which a voter might encounter claims in both directions on social media.) As fact-checking has expanded in the past few years, some research has shown that it can be effective in correcting voters’ misapprehensions. But in this case, the fact-checks didn’t create a measurable reduction in the damage to voter confidence. It’s not entirely clear why not, though Nyhan speculates that fact-checks, while effective at shifting voters’ beliefs about specific points, may be less effective at changing attitudes, such as confidence in elections.
At least two previous studies have examined the question of Republican faith in elections in the Trump era. A 2018 study found that, despite Trump’s rhetoric, GOP confidence didn’t decrease before the 2016 election—and Democratic faith actually increased in comparison with 2012, perhaps because Hillary Clinton and other prominent Democrats counteracted Trump’s sowing of doubt. Another study found that after his victory, Republican partisans felt better about election integrity: “Many Trump voters concluded that illegal voting was not only less consequential (Trump won, after all) but also less prevalent than they had thought.”
This is an established concept in political science: the winner effect. After a candidate wins, voters in his or her party tend to see an increase in confidence in elections (while those who support the loser may see a drop).
But the current moment in American politics makes the stakes higher than the normal seesaw pattern of partisan confidence. Following the hotly contested and flawed 2000 presidential election in Florida, Democratic confidence was lower than Republican. After George W. Bush won reelection in 2004, GOP voters “were already about 85% likely to be very confident in the accuracy of the election, and thus did not have a lot of room to increase,” Michael Sances and Charles Stewart found in 2014. But since then, a dramatic shift has taken place:
Across the decade, aggregate confidence in the country’s vote declined about 30 percentage points. However, this overall decline is due almost entirely to a forty-point decline among Republicans during this period. The biggest change occurred between 2004 and 2008, which saw a shift, from Republicans being more confident in the nationwide vote count, to Democrats expressing more confidence.
In part, this stretch saw a string of Democratic victories: 2006, 2008, and 2012 (interrupted by off-year Republican wins in 2010 and 2014). But it also saw a concerted effort by conservative activists to sow doubt about the integrity of election systems. Figures such as Kris Kobach, Hans von Spakovsky, and J. Christian Adams have alleged widespread fraud, in different forms: out-of-state voters, double voters, noncitizen voters. (They have repeatedly failed to produce evidence to back up their claims.)
Trump seized hold of this idea, elevating it to the center of Republican discourse. First, he cast doubt on the results before the election, when he was widely expected to lose. Then, even after winning, he claimed that there had been as many as 3 million illegal votes cast, and appointed Kobach to run a commission on election fraud. (The group quietly disbanded after turning up nothing.) If this is how Trump spoke in victory, one can only imagine how he might react if he loses in November. This study suggests that many Americans will take what he says to heart, too.
“In modern American political history, we’ve never had a major-party presidential nominee allege that the election was stolen from them, let alone an incumbent who has to leave office,” Nyhan said. “The conversation has focused too much on the ‘Will Trump step down?’ question. I’m much more worried about the damage to institutional legitimacy that he can do on the way out.”
Perhaps the most alarming finding of the study is how little effect fact-checking has on the claims. If people believe spurious allegations from Trump and others, and the damage can’t be mitigated by fact-checks, it’s difficult to see how confidence in elections—and by extension, democracy itself—can be rebuilt. The best hope might be to counteract the claims before they take root.
“One of the most effective approaches to countering voter-fraud allegations is to discourage people from making them at all,” Nyhan said. “If media coverage is negative enough, or the political costs are high enough, we know politicians will often avoid making controversial claims.”
Not every claim of problems in elections is baseless, of course. The June 9 primary in Georgia, for example, had extensive issues: There were hours-long lines in some precincts, polling places were delayed in opening for hours, and other problems were reported. Election officials say thousands of votes sent in by mail may not have been counted, thanks to issues with tabulation machines. These problems are real, and documentable (even if capturing the precise scale is impossible). Voters in places such as Fulton County, Georgia, have real reason to wonder whether their votes were counted—if they were able to cast them at all. But claims such as the ones Trump is making don’t have evidence to back them up.
This past Friday, some Trump supporters in Michigan burned absentee-ballot applications they received as voters in the state. The state attorney general had decided to mail the applications to every voter as a response to the coronavirus pandemic, and Trump had criticized the move as an invitation to fraud. Meanwhile, local GOP officials from across the country told Politico that—despite growing evidence that the president’s reelection campaign is in trouble—they expect a victory or even a “landslide” in November, just as he is promising. Gestures and sentiments like this may seem merely absurd or self-defeating now, but if Trump loses in November, they could drive a dangerous erosion of faith in American democracy.