Dan Thompson of Canton, Mich., yells out against health-care reform during Rep. John D. Dingell’s town hall meeting in Romulus, Mich., in 2009.(Kimberly P. Mitchell/Detroit Free Press via AP)

The cynical view of political discourse is that, in these polarized times, members of Congress can’t really persuade their constituents. If talk does anything, leaders can only mobilize their supporters and tell them how they should think and vote. But in a new study, my collaborators (Kevin Esterling, David Lazer, William Minozzi and the Congressional Management Foundation) and I found something potentially much more encouraging about how political leaders and citizens can talk to each other.

We discovered that when members of Congress substantively engage constituents about controversial subjects, they can persuade them on policy, increase trust in their leadership and even garner more votes. More importantly, it doesn’t matter if they’re talking to someone from a different political party or not. The key is to reach out to regular citizens who may not have the passion of intense partisans, but who want to be involved in the democratic process nonetheless.

For our study, published today in the Proceedings of the National Academy of Sciences, we worked with 12 members of the U.S. House of Representatives (seven Democrat and five Republican) and one senator to set up 20 online town halls with their constituents. These town halls met in 2006 and 2008.

They weren’t the typical town hall shouting matches that you see covered in the media. In addition to being online rather than in an actual room, we recruited a random group of each member’s constituents to participate. That means there weren’t just the members’ biggest supporters or critics participating. In fact, our analysis showed that these citizens were actually more representative of eligible voters in their districts than were actual voters.

In each of the town halls, the members of Congress and citizen participants talked about a hot-button issue linked to recent or pending legislation. In the case of the representatives, they discussed whether there should be a path to citizenship for undocumented immigrants. The senator (Carl Levin, a Democrat from Michigan) discussed terrorist detainee policy. In all cases, the members of Congress clearly stated their views on the subjects during the town halls.

But because these town halls included randomly selected citizens – and not just strong partisans there to cheer or jeer the member – the discussions were remarkably civil and substantive. The members concentrated on arguing for their position on its merits and didn’t rely on the red-meat talking points that fire up strong partisan supporters.

The results were clear and striking. We found that the town-hall participants’ views on the issue actually moved closer to that of their member of Congress. But it went further than that. People rated their member of Congress as more trustworthy and were more likely to approve of his or her performance after being a part of the town hall.

In addition, participating in these town halls actually changed the constituents’ voting behavior. Attendance increased constituents’ intent to vote for their representative by nearly 14 percent after the town hall. Participants were surveyed again after the next November election, and we found a nearly 10 percent increase in the likelihood of these constituents actually voting for their representative in that election. This election occurred about four months after the town hall, suggesting that these sessions had a relatively long-lasting effect.

Moreover, these patterns scale up. Similar results were found for those who attended Levin’s town hall, even though almost 10 times the number of people participated. In fact, we estimate that if members of the House spent just two hours a week doing outreach like this, they could interact with almost half of the voters in their districts over the course of a few election cycles.

This study, then, gives us a plausible roadmap for how we can begin to use online communications to improve our political discourse, in particular by offering attractive forums for those citizens who may not be participating now. We found that 94 percent of the participants said they would participate again. These citizens told us they enjoyed talking about the issues with their representative without the shouting and the intense strife. We heard similar sentiments from the members of Congress who took part.

That’s something we need more of in representative democracy today: Leaders talking about real issues with their constituents, minus some of the intense partisan heat.

Michael A. Neblo is an associate professor of political science and (by courtesy) philosophy at Ohio State University.