Hundreds of thousands march to the White House during the Women’s March on Washington on Jan. 21. (Evelyn Hockstein for The Washington Post)

This is the second post in our series on what social science can tell us about the Women’s March. Find our first one here. More posts will appear in the coming days. — TMC editors

Millions of people across the United States and the rest of the world showed up at one of the Women’s Marches this past Saturday, rallying around a common theme of respect for women’s rights (and those of all marginalized groups) in the wake of Trump’s election.

In particular, the march listed specific policy goals such as reproductive rights, a livable minimum wage, paid family leave, and an Equal Rights Amendment along with broader social goals such as ending discrimination and violence against women, people with disabilities, LGBT people, Muslims, communities of color, immigrants, and many other groups.

Will they be able to achieve their goals?

It turns out that social science has a lot to say about which protests are likely to be effective. My research shows that social movements can indeed create long-lasting political change.

Using the U.S. civil rights movement as a case study, I assembled a data set of counties that did and did not experience peaceful civil rights protests before Congress passed the Voting Rights Act in 1965. To see whether these protests shifted these counties’ residents in a more liberal direction, especially on racial issues today, I also mapped contemporary public opinion data and voting rates for the Democratic Party onto the counties that did and did not experience civil rights protests.

I then compared whites from counties that had no protests with those that had peaceful civil rights protests 50 years ago. Here’s what we find among the white people from counties that did have peaceful protests 50 years ago:

  • less racism (as measured by standard racism scales used by political scientists and psychologists)
  • 2.5 percent more likelihood of supporting affirmative action
  • 3 percent more likelihood of identifying with the Democratic Party today.

Statistical analyses of these findings demonstrate that it is unlikely that these patterns exist as a result of chance alone.

Moreover, these patterns hold after adjusting for many alternative explanations, like that county’s median income, the proportion of the county that is black, and the Democratic Party vote share of that county before the civil rights movement.

Here’s what we find in two nearly identical counties

Take, for example, two counties that had identical average incomes, demographics and voting behavior before the civil rights movement. Whites today from the counties that did have protests show less racial prejudice toward blacks than their counterparts from counties that did not — but were otherwise similar.

Whites from the counties that did have protests, decades ago, are more likely today to identify as Democrats — the party associated with racial liberalism — compared with those from counties that had no such protests.

What’s more, they’re less likely to believe, for instance, that African Americans don’t work hard enough, and more likely to believe that historical discrimination against African Americans still matters today. And they’re also more likely to support affirmative action policies than whites from the non-protest counties.

My statistical analyses demonstrate that these patterns are not the result of random chance alone and are unlikely to be explained by factors that political scientists cannot measure, such as the unique cultures of different counties.

Existing research shows that these local protests did sway members of Congress at the time. But I find that it did more: It changed hearts and minds in ways that lasted more than 40 years after the movement ended.

These 3 factors make a difference in whether protests work

Obviously, no two protests are the same. Here’s what successful ones have in common.

The civil rights movement and more recent social movements, like the tea party, succeeded because they worked hard on three key factors: organization, messaging and nonviolence.

Organization. Protesting isn’t enough. Protesters have to be organized in a way that lets them keep pushing for the movement’s goals. In an article that received quite a bit of media attention, Andreas Madestam (Stockholm University), Daniel Shoag (Harvard Kennedy School), Stan Veuger (American Enterprise Institute), and David Yanagizawa-Drott (University of Zurich) showed that the tea party made significant political gains in large part because, after the first Tax Day protests in April 2009, local groups stayed involved. Theda Skocpol of Harvard University and Vanessa Williamson of the Brookings Institution found that frequent grass-roots meetings, funded by such organizations as the Koch-run Americans for Prosperity, helped the tea partiers push the Republican Party rightward.

Messaging. Second, does the message resonate for more people than just the core supporters? This is especially important for groups that seek to persuade either tepid supporters or those who have yet to make up their mind.

Movements can do this in various ways. Its core supporters could personally persuade others — maybe through their social networks — or they could frame their grievances and goals in a way that appeals to core values such as patriotism, equality and freedom, to name a few.

For instance, the tea party’s name and core message of limited government harked back to the founding of the United States and its principle of individual freedom.

Nonviolence. Third and last, is the movement nonviolent? Erica Chenoweth of the University of Denver and Maria J. Stephan of the United States Institute of Peace find that, across the world, nonviolent movements tend to be about twice as effective as violent ones.

For example, research by Omar Wasow at Princeton University suggests that violent riots after the Rev. Martin Luther King Jr.’s assassination dramatically reduced the vote for the Democratic Party in 1968. A recent working paper by Ryan Enos, Aaron Kaufman and Melissa Sands of Harvard University suggests that violent protests can still be effective as long as they prompt supporters to become politically mobilized.

To be successful, social movements need not do all three of the above. But consistently succeeding at least one of these features makes a big difference.

So can it succeed?

The Women’s March may have been the biggest one-day protest in U.S. history. But will it be as successful as the tea party or fizzle out like the Occupy Movement? Let’s look.

Organization: check. So far, Women’s March organizers seem intent on keeping newly activated marchers involved in their 10 Actions/100 Days campaign. Moreover, the movement’s ability to raise the funds to cover the March’s expenses suggests that at least the core organization can attract financial donors.

Messaging: too soon to tell. Will the movement’s principle, “Women’s Rights are Human Rights and Human Rights and Women’s Rights,” resonate? Initial reports suggest that women brought a significant number of men and that supporters are using their social networks to keep spreading the word.

Nonviolence: check. Every observer — even if they dislike the movement’s goals — appears to agree that the protests were nonviolent. What’s more, the movement explicitly states its commitment to nonviolent principles. That’s likely to continue.

No one expects Women’s Marchers to reach all their goals immediately. But the movement seems to be heading in the right direction.

Shom Mazumder is a PhD student in government at Harvard University, examining the historical origins of gender inequality in the United States. Find him on Twitter @shom_mazumder.