Apostolos Filippas, Ph.D., On Social Media Echo Chambers
Faculty | Aug 05, 2021 | Michael Benigno
By Chelsee Pengal
Do you ever scroll through Facebook or Twitter and see the same tweets touting dubious information repeatedly shared on your feed? If so, you’re not alone.
According to Apostolos Filippas, Ph.D., assistant professor of information, technology, and operations at the Gabelli School of Business, social media echo chambers, where people share and consume pools of similar information, are common. In fact, news stories and documentaries about the way social media platforms can facilitate the spread of untrue or harmful information have become increasingly prevalent.
The cause of these echo chambers is the act of reciprocal following or what Filippas calls “influence trading.” In a recent paper, “Influence Trading on Social Media,” Filippas and his coauthor developed an economic model of social media platforms that describes the phenomenon. In modeling Twitter users’ abilities, followers, and influence, they found that influence trading greatly affects the production and consumption of information on social media. They also found it to be “natural human behavior.”
When you follow another user on social media just so they will follow you or vice versa, you engage in reciprocal following. In doing so, the overall quality of the tweets you consume decreases. “These are people you wouldn’t have followed before, so…on average, what they’re posting isn’t good enough to have on your feed,” Filippas explained. However, your influence increases with your rising follower count—a growing audience for the tweets you produce.
Filippas’s model illustrates that as users continue to follow one another back and forth, they form “filter bubbles” made up of people following the same subset of users. These groups continue to grow and eventually contain many users that are all part of the same bubble.
“What happens is your feed starts being filled with content from these reciprocal relationships,” said Filippas.
There is a potential downside. In a filter bubble with hundreds of users tweeting and reading the same content, the possibility of spreading misinformation increases. Filippas hopes his paper casts light on how these relationships occur and helps platforms consider ways to shift behavior.
“What you can do is start imposing costs upon these reciprocal relationships,” he pointed out. For example, platforms might hide the number of followers a user has or make it harder for a user to see who is following back.
On the other hand, Filippas cautioned, if attempts such as these are made to limit reciprocal following, it may reduce the enjoyment—and eventually the usage—of the platform.
Some people blame platform executives for allowing echo chambers to form. However, Filippas said the model shows this result is not dependent on how a platform is structured. “It’s an outcome of the fact that people now are all producers and consumers at the same time,” he said.
Despite the uncertainties, Filippas believes the filter bubble problem will eventually be solved. After all, he said, the industry is still relatively new. “I think we’re going to see social media as a positive development after these growing pains.”