Protesters gather at Boston’s Free Speech Rally on Boston Common on Nov. 18, 2017. (Mark Lorenz/The Boston Herald via AP)

For the past several years, the United States has been debating freedom of speech. Some argue that encouraging dissenting points of view to be expressed, even the most hateful, ensures a healthy democracy. Others argue that speech that is offensive or hateful limits room in the public square for historically marginalized voices, while allowing it helps odious political views gain respectability.

These debates for and against free speech focus largely on what citizens and policymakers ought to do. They take a normative stance on the question of freedom of speech. In recently published research, however, we tried to see whether there is an empirical link between a climate of support for freedom of speech and indicators of the frequency of debate and dissent. In doing so, our goal is to test whether freedom of speech does in fact help to bolster democratic vitality.

Here’s how we did our research

We begin by measuring support for the principle of freedom of speech in American communities. We used public opinion data from the Freedom and Tolerance surveys, which are nationally representative surveys of the U.S. public, conducted every year between 2007 and 2011. The relevant survey questions focus on the broader concept of “political tolerance” — the degree to which citizens will support the extension of civil liberties to all, including those who advocate viewpoints that are highly disagreeable to respondents.

The first set of questions focuses on people’s support for civil liberties in general: whether citizens should have to carry around a national identity card at all times, whether high school teachers should be required to defend U.S. policies, whether the government should be allowed to monitor telephone calls and emails to prevent terrorist or criminal acts, and whether law enforcement officials should be allowed to use racial profiling and investigate non-violent protesters. The second set of questions asks if a group that the respondent strongly dislikes should be allowed to make a speech in the community, run for office, or hold a public rally. We combine responses to these eight questions into a single index of political tolerance.

Although our pooled survey data set is fairly large, with around 4,000 respondents, it is not large enough to allow us to split the data by respondents’ communities of residence. Instead, we apply a method known as multilevel regression with poststratification to the pooled survey data. This enables us to reliably measure support for freedom of speech and other civil liberties in 365 U.S. metropolitan areas.

As you can see below, the map shows some surprising findings. Major metropolitan areas, such as New York/Northern New Jersey/Long Island and Chicago/Naperville/Elgin, are not particularly supportive of free speech and civil liberties. The most tolerant metro areas are instead college towns such as Boulder, Colo.; Ithaca, N.Y.; and Corvallis, Ore. As previous research on tolerance in the states has found, regional patterns are also evident, with tolerance predominating in the upper Midwest and Mountain West, and intolerance in the South.


Next, to measure dissent in these metropolitan areas, we used data on protest incidence from the GDELT project, which uses software to scan news media for stories on political events. We counted the number of protests recorded by GDELT in each metropolitan area between 2007 and 2011. We divided this by the population of each metro area, and used it as a measure of publicly expressed dissent in that area.

More support for freedom of speech is associated with more public protests. But which comes first?

As you can see in the figure below, the rate of protest increases in step with the level of a metropolitan areas’ political tolerance.


But this doesn’t tell us why they are associated with one another. Which comes first? Does support for freedom of speech and other civil liberties lead to more protest, or does protesting  increase political tolerance? Do they reinforce one another, or is there some other relationship between the two that makes the same area more likely to host both?

To help us find out, we use a regression model to adjust for other factors that might affect the rate of protest, such as the ideological leaning and proportion of students in the area. Even accounting for those factors, we found that tolerance and protest are closely related. We then examined whether this association between tolerance and protest holds when we take the rate of protest in the preceding years into account. It does. Ultimately, although our data do not allow us to firmly conclude that tolerant metropolitan areas cause increased rates of protest, higher political tolerance and higher rates of protest are closely linked.

These findings are consistent with earlier research showing that Americans living in politically tolerant neighborhoods feel freer to engage in public expression of politics. What’s especially interesting is that African Americans living in neighborhoods more tolerant of racist speech also feel greater freedom to express themselves politically.

What are the implications of these findings for democratic vitality?

These studies suggest that local support for freedom of speech emboldens citizens living in these areas to participate in political life. If this is true, it would support a classic argument for freedom of speech and political tolerance more generally: that it encourages all views to be expressed. Limiting the freedom to express unpleasant opinions, the argument goes, can create a creeping conformity that spreads, in a “spiral of silence,” as people learn that it is best to simply keep their mouths shut. While our findings do not suggest where exactly these limitations should lie, they do bolster the argument for putting up with a variety of unwelcome speech if one’s goal is to invigorate political engagement.

Christopher Claassen is a lecturer in politics at the University of Glasgow.

James L. Gibson is Sidney W. Souers Professor of Government at Washington University in St. Louis and Fellow, Centre for Comparative and International Politics and Professor Extraordinary in Political Science at Stellenbosch University (South Africa).

The data used in the reported research were funded by the National Science Foundation (SES 1228619) and the Weidenbaum Center on the Economy, Government, and Public Policy at Washington University in St. Louis.