Can Twitter track how politicians are affected by their constituents?

Prof. Pablo Barberá on assessing congressional responsiveness through tweets

America’s contemporary political scene is characterized by an open distrust of elected officials, with the common refrain, on both sides of the isle, being that politicians have forgotten their constituents. While there may be a disconnect between political action and public interaction, in the realm of social media, politicians seem to be heavily influenced by the sway of public opinion.

Professor Pablo Barberá is an Assistant Professor at the University of Southern California’s School of International Relations and a former Moore-Sloan Post-Doctoral Fellow at the Center for Data Science. While completing his PhD in Political Science at New York University, Barberá collaborated with NYU’s Social Media and Political Participation Lab, on a project titled “Leaders or Followers? Measuring Political Responsiveness in the U.S. Congress Using Social Media Data.” The project analyzed the Twitter accounts of U.S. Congress members to see if congressional tweets are affecting, or are affected by, tweets from constituents.

Barberá chose to analyze Twitter accounts because the platform’s interactive nature allowed him to understand how a “legislators’ expressed political agenda is affected by what their constituents discuss publicly on this same platform; and vice versa.” It was found that, far from dictating what people think, politicians — in the realm of social media — are significantly influenced by how constituents express their political beliefs.

Over a 15-month period, Barberá analyzed all of the tweets that were generated by members of congress. He used a Latent Dirichlet Allocation model — which assumes that a set of text is characterized by a topic — to see the relationship between topics discussed by politicians, and topics discussed by constituents. For example, on Twitter, the phrase “stopgovtabuse” is frequently used to discuss foreign policy, so a tweet that uses this phrase can be labeled as a discussion regarding foreign affairs.

By monitoring the distribution of a given topic, Barberá and his team found that, over time, both Democrat and Republican legislators tweeted more about a given topic if their Twitter followers began talking about a subject first.

However, the influence failed to go both ways. Barberá and his team also found that when a politician began tweeting about a given subject, it had almost no effect on what their followers tweeted.

Barbera argues that a politician’s tweets present a standardized picture of their political activity. He said that politicians use Twitter “as an additional platform to stay in touch with their constituents, to express their policy positions, and to broadcast political messages.” Now all that’s left to figure out is whether or not political tweets ever amount to political action.


Originally published at cds.nyu.edu on August 5, 2016.

Like what you read? Give NYU Center for Data Science a round of applause.

From a quick cheer to a standing ovation, clap to show how much you enjoyed this story.