The Core-Periphery Structure of Online Amplification
Dissertation proposal
Ryan Gallagher
Network Science PhD Student
Past Talk
Thursday
Jan 28, 2021
Watch video
3:00 pm
Virtual
177 Huntington Ave.
11th floor
Online
Register here

Social media relies on amplification. It is at the heart of how marginalized communities voice injustices, how information operations stoke long-standing racial divisions, how elected officials communicate public health guidance, how misinformation proliferates through vulnerable populations, how destined friends find each other online, and how abusers perpetrate harassment at scale. In all of its uses, good and bad, amplification is the process through which individuals construct online communication networks. Users create ties with others by amplifying them: any share, retweet, or crosspost forms a connection between them. As they amplify many different memes, stories, and videos, they begin to create a network with a dense core. This core consists of all those who receive the most amplification, the most shares, the most visibility. And around that core radiates a periphery, all those who shared content from the core, giving it its visibility. These two parts of the network depend on one another: without the periphery’s amplification, there is no core; and without the core’s content, there is no periphery. So it is only by understanding these two components together---the core and the periphery---that we can fully understand the structure of amplification.

Here, I propose three projects for understanding the core-periphery structure of online amplification. First, I demonstrate that there are multiple ways of envisioning core-periphery structure in networks, and that these different characterizations can be formulated as distinct statistical models. Next, I describe how I will apply these statistical models to a number of online communication networks, ranging across Twitter, Reddit, and TikTok, and demonstrate why amplification needs to be understood, both theoretically and methodologically through the lens of core-periphery structure. Finally, I propose to couple my core-periphery models with techniques from computational text analysis and measure how affect flows within and between polarized communities as they engage around contentious issues. By bringing together network science with theories of online communication, I intend to advance our understanding of amplification and how it reverberates through the networked public sphere.

Dissertation Committee:
Brooke Foucault Welles (Chair)
Network Science Institute, Northeastern University

David Lazer
Network Science Institute, Northeastern University

David Smith
Khoury College of Computer Sciences, Northeastern University

Sandra González-Bailón
Annenberg School for Communication, University of Pennsylvania

About the speaker
About the speaker
Ryan is a fourth-year PhD student working with Professor Brooke Foucault Welles as a part of the Communication Media and Marginalization (CoMM) Lab. At the Network Science Institute, he studies how individuals use online communication networks to amplify their voices, and how that amplification resonates through online media ecologies. To do so, his research makes advances in network science and text-as-data methodology to develop new approaches for measuring the complexities of polarization, misinformation, and the networked public sphere. Ryan holds an MS in mathematics from the University of Vermont, where he worked with the Computational Story Lab at the Vermont Complex Systems Center, and a BA in mathematics from the University of Connecticut.
Ryan is a fourth-year PhD student working with Professor Brooke Foucault Welles as a part of the Communication Media and Marginalization (CoMM) Lab. At the Network Science Institute, he studies how individuals use online communication networks to amplify their voices, and how that amplification resonates through online media ecologies. To do so, his research makes advances in network science and text-as-data methodology to develop new approaches for measuring the complexities of polarization, misinformation, and the networked public sphere. Ryan holds an MS in mathematics from the University of Vermont, where he worked with the Computational Story Lab at the Vermont Complex Systems Center, and a BA in mathematics from the University of Connecticut.