New research sheds light on how the structure of online social networks causes misinformation to go viral on the internet. The findings, published in PLOS One, indicate that social “echo chambers” act like kindling that gives misinformation the initial flare up it needs to quickly spread.
“While the link between ‘echo chambers’ and ‘fake news’ has been both widely discussed and been the subject of significant research effort, this discussion has largely left out the impact of the complex network dynamics that are so quintessential to online social media,” said study author Petter Törnberg, a sociologist from the University of Amsterdam.
“Such dynamics are central to understanding new social media, since what receives attention on social media is largely the result of what we call ’emergent processes’, that is, the unintended and unexpected macro consequences of the micro-interactions of millions of users. This is producing certain dynamics that, while having become central to current societal trends, are exceptionally difficult to study using our traditional research tools.”
Törnberg used a computer model to simulate misinformation spreading via shares in online social networks. They found that misinformation tended to “go viral” after first resonating within a cluster of like-minded individuals.
“The study finds that the presence of a group of like-minded users will tend to increase the virality of news and rumors that resonate with the views of this cluster,” Törnberg told PsyPost.
“This seems unintuitive: one would think that if a group of users are less connected to others, they will have a harder time spreading their views. As I write in the article, the nature of this link is perhaps most clear through the lens of a metaphor: if we think of the viral spread of misinformation in a social network as akin to a wildfire, an echo chamber has the same effect as a dry pile of tinder in the forest; it provides the fuel for an initial small flame, that can spread to larger sticks, branches, trees, to finally engulf the forest.”
“This essentially means that social media may, for structural reasons, give precedence to fringe ideas,” Törnberg remarked.
To validate that these dynamics reflected what was actually happening on the internet, Törnberg plugged data from Twitter into his model to generate empirically grounded networks. But the study — like all research — has limitations.
“Simulations are exceptionally powerful for identifying the link between underlying mechanisms and macro-dynamics, as they allow the isolated study of dynamic phenomena. However, like any approach based on abstraction, they are also always limited since we cannot be sure that the causal mechanism that they identify is in fact important in real-world dynamics,” Törnberg explained.
“The effects may be crowded out by factors that are not included in the simulation. Additional empirical study is therefore necessary to assess the impact of these emergent dynamics.”
“This paper is part of the EU research project ODYCCEUS, which focuses on using methods from the hard sciences to study societal polarization and cultural conflicts. I’m currently pursuing multiple ideas as part of this research project, some of which continues on the ideas of this paper, by modeling the dynamics of online debate,” Törnberg added.
“To follow my publications, follow me on Twitter (@pettertornberg) or on ResearchGate.”