Universiteit Leiden

nl en

Rebekah Tromble to lead Twitter-funded research team on online discussions

In the context of growing political polarisation, the spread of misinformation, and increases in incivility and intolerance, how can Twitter assess and improve the quality of its conversations? To address this question, an interdisciplinary team of researchers from Leiden University, Syracuse University, Delft University of Technology, and Bocconi University was selected by Twitter to conduct research and develop metrics to help identify behaviours that are threatening to the quality of the discussions on the platform. The team was one of two chosen to receive the Twitter research grant. Over 230 proposals were reviewed in the selection process.

‘Academic researchers and tech companies need to work together’

Led by Rebekah Tromble, Assistant Professor of Political Science at Leiden University, the team will conduct research comparing discussions around polarised and non-polarised topics in the United States and United Kingdom to develop and implement four metrics that will provide a better understanding of how communities form around discussions on Twitter and will investigate the extent of certain problems that may develop  in those discussions. ‘The last few years have made it clear that if we are going to effectively identify, evaluate, and address some of the most difficult problems on social media, academic researchers and tech companies will need to work together much more closely. This initiative presents an important and promising opportunity for such collaboration,’ Tromble says.

Tweet announcing the 'health of public conversation' initiative
Tweet announcing the ‘health of public conversation’ initiative

Filter bubbles and echo chambers

The project focuses on two challenges faced by Twitter. The first is the presence of echo chambers—that is, the extent to which discussions are enclaved in homogeneous, and often polarized, groups. ‘When people are unaware of others’ points of view, they have informational blind spots. This problem is aggravated online, because people are more likely to access content that is tailored to their preferences and interactions. While we have some evidence that filter bubbles and echo chambers are a problem on Twitter, one of this project’s contributions will be to empirically understand the extent and severity of the problem,’ explains Nava Tintarev, Assistant Professor at Delft University of Technology (Netherlands), whose work will focus on developing and implementing computational techniques to assess the diversity of perspectives on Twitter.

Uncivil and intolerant discourse

The second area of research will explore uncivil and intolerant discourse. The research team will develop metrics examining the extent to which those who participate in Twitter discussions engage in toxic behaviours. ‘Political discussion online is often dismissed due to uncivil discourse, and because of that we are sometimes unable to understand its true value. Rather than lumping all problematic discourse into a single category, we distinguish between uncivil and intolerant statements. Incivility online might serve an important and valid purpose in discourse, while intolerance is, by its nature, threatening to democracy’, explains Patricia Rossini, postdoctoral researcher at Syracuse University (United States), whose work focuses on understanding such discourse online.

Bridging political science, communication, and computer science

The team brings together scholars with different backgrounds and expertise, bridging political science, communication, and computer science to develop metrics and conduct experiments aimed at identifying potentially problematic behaviours on Twitter. Dirk Hovy, Associate Professor at Bocconi University (Italy), an expert in computational sociolinguistics, notes the extraordinary opportunity afforded by this project: ‘I am extremely excited to be part of a such a unique collaboration between social and technical sciences on a project of this importance to society. Working across disciplines will enable unparalleled insights into healthy online conversations, and collaborating with Twitter means we can analyse authentic source data at an unprecedented scale.’

Rebekah Tromble
Rebekah Tromble (photo: Ruben Verheul)

The team also counts the expertise of Michael Meffert, Assistant Professor in Political Science at Leiden University, who will be responsible for leading experimental design, implementation, and analysis, and Jennifer Stromer-Galley, Professor in the School of Information Studies at Syracuse University, whose expertise in political communication and social media will provide support to the conceptualisation, development, and implementation of Twitter health metrics.

As part of the research grant, the team will work closely with Twitter. While scholars will have the autonomy to develop and publish research under this grant, the outcomes of this project will help inform Twitter’s future policies and practices to promote a healthy conversational environment.

Further reading

Vijaya Gadde and David Gasca, ‘Measuring healthy conversation’ (Twitter Blog)

This website uses cookies.  More information.