Protests against President Trump and his executive actions have been erupting across the country. As they began, immediately after the Jan. 22 women’s marches, Trump tweeted:

The tweet implied that people who protest do not vote. Research suggests that the opposite is true: People who protest are no less likely to vote than non-protesters. If anything, protesters vote at higher rates.

Here’s the research on whether protesters vote

Although we lack information about this year’s protesters, the American National Election Study (ANES) is full of information about whether and how Americans got politically involved in previous years. Here’s what they reported.

Among those who said they had been in at least one protest in the previous four years, 85 percent said they voted in the 2012 election, compared with 78 percent of those who did not go to protests. In the 2008 ANES, 89 percent of protesters reported voting in that year’s presidential election, compared with 79 percent of non-protesters.

Those numbers aren’t entirely accurate. People who don’t vote often say they did, because voting is generally considered the right thing to do. Only 57 percent of Americans actually voted in 2008, but 78 percent told the ANES that they did; in 2012, 55 percent of Americans voted, but 80 percent told ANES that they did). That makes surveys of voting behavior a bit unreliable.

With this problem in mind, we analyzed the British Election Study (BES), which uses verified turnout: It gathers identifying information about the survey’s respondents and matches this information against public election records. In the 2005 and 2010 BES, conducted during years that the United Kingdom held general elections, survey respondents were asked how likely they were to protest. During the first decade of this century, Britain saw frequent mass protests, such as the 2003 anti-Iraq war mobilizations and, in 2010, protests against education spending cuts and university fee increases.

U.K. respondents who described themselves as likely to protest were also more likely to vote. In 2005, 70 percent of the most-likely protesters said they voted in the previous election, compared with 62 percent of least-likely protesters. In the 2010 survey, most- and least-likely protesters alike said they voted at a rate of 70 percent.

What does this mean about whether recent U.S. protesters are also voters?

The demographic breakdown of the women’s marches, assessed from media reports and observations, suggests that those crowds were likely voters. Women outnumbered men at the protests, and for the past 30 years, women have generally voted at higher rates.

That’s not as true when it comes to age. Young people everywhere vote at lower rates in elections and are often overrepresented in protests. We do not yet have accurate evidence about whether the women’s marches and travel-ban protesters skewed particularly young.

However, while young people in general are less likely than their elders to vote, that’s not as true for those who protest. Young protesters are also, often, young voters.

Consider the evidence from Chile, a stable democracy that faced widespread student protests from 2011 to 2013. In two 2014 surveys of Chilean citizens, people who said they had demonstrated during the previous 12 months were more likely to report that they voted in the 2013 presidential elections. In one of the surveys, conducted by the Latin American Public Opinion Project (LAPOP), protest participation alone is not a statistically significant predictor of voting, but it has a strong positive effect once we control for age.

In other words, young people who protested voted at higher rates than their peers who did not.

Getting involved politically — by voting or by protesting — has costs

Trump’s claim that protesters are mainly non-voters also runs up against the logic of political participation. Participation in protests is more costly and onerous than voting. Protesters who made their way to the Mall or other locations on Jan. 21, or massed at airports on Jan. 28, had to make travel plans and coordinate schedules with others. They might have paused to consider the risk of being roughed up by police or counter-demonstrators. The costs and inconveniences of voting, by contrast, are minor.

Therefore, we expect those who are willing to get out to march are often quite willing to get out to vote.

Of course, just as there are costs to participate, there are also costs to staying away. Often people become politically active when friends, co-workers and family members expect them to; when they believe that they have a social responsibility to be civicly engaged; and when they care enough about collective outcomes to feel uncomfortable staying at home.

Nearly 140 million Americans cared enough to vote on Nov. 8. Between 3 million and 5 million cared enough to protest on Jan. 21, and tens of thousands were angered enough by the refugee and immigrant bans to gather at airports the following weekend. There is little doubt that — Trump’s tweet notwithstanding — most of the several million protesters in January had also been voters in November. And in fact because these protests, so early in Trump’s administration, have been historically large, they may portend an uptick in voting in the usually quieter midterm congressional elections in 2018.

Erdem Aytaç is assistant professor of political science at Koç University in Turkey.

Susan Stokes is John S. Saden professor of political science at Yale.

Eli Rau is a PhD student in political science at Yale.

Aytaç and Stokes are writing a book about popular participation in elections and protests.