Social media networks shape consensus through individual posting decisions

What is it about?

This paper shows how people influence one another through their social media posts and explains why certain narratives spread and what different groups come to believe and support. We built an artificial social media environment – like a controlled playground – where volunteers read several passages about the 2011 Fukushima earthquake, and composed Tweet-like social media posts about the event under conditions controlled by the researchers. Using a purpose-built AI tool to measure people's beliefs about events through their online interactions, the researchers found that an online playground where everyone is connected to one another by, for example, using the same hashtags, supports the adoption of shared beliefs that closely match the real earthquake causes and effects they had read about. In contrast, when people find themselves in a playground where they are connected only to a small number of neighbors, shared beliefs do not emerge. Much like what happens in online echo chambers, people adopt idiosyncratic narratives that may not be faithful to the facts. These differences persisted in participants’ language even after social media interactions concluded, demonstrating how social media shapes people's offline understanding of real-world events.

Featured Image

Photo by Alina Grubnyak on Unsplash

Photo by Alina Grubnyak on Unsplash

Why is it important?

Social media helps people create and share narratives that can have either positive or negative effects on society. Understanding how people influence one another through their posts can help explain why certain narratives spread and what groups come to believe and support. However, research on how online narratives spread and shape beliefs is limited for two reasons. First, the complexity of real-world social media environments makes it difficult for research to get at the heart of the problem. Second, it has been challenging to precisely measure how people's interpretations of narratives change over time and across groups. We addressed these gaps by building an artificial social media environment where we could control the narratives people read and who they interacted with. Controlled group experiments help identify cognitive mechanisms that can be leveraged to improve narrative-based communications on the internet.