Facebook's Data Team Challenges Fears of an Online Echo Chamber

August 24, 2012 Paul M. Davis

Online pundits and media critics warn that as social media increasingly becomes a dominant source of news, and aggregators like Google News develop algorithms to surface stories that are presumably more interesting to users, we’re participating in an echo chamber where self-selected social groups and online habits reinforce our existing beliefs. Online activist Eli Pariser named the phenomenon “The Filter Bubble” in his recent book, arguing that “left to their own devices, personalization filters serve up a kind of invisible autopropaganda, indoctrinating us with our own ideas.” The idea makes a fair amount of intuitive sense, and similar concerns have been raised by web luminaries such as Clay Shirky and Tim Berners-Lee. But a recently-released internal study of the behavior of 250 million Facebook users suggests such fears may be overblown.

Eytan Bakshy, a Ph.D. in Information Studies at the University of Michigan, who is now a member of Facebook’s data team, has been studying the spread of information on the service since 2010. The goal of Bakshy and his colleagues is to “rigorously quantify influence at a mass scale.” Working from the basis of economic sociologist Mark Granovetter’s theories of the strong and weak social ties that socially bind individuals, the researchers observed and experimented to see how information was shared on Facebook’s News Feed between users. They were particularly interested in “The Collective Influence of Weak Ties — how these network effects shape information spread as a whole,” and how interest and novelty leads to the amplification of certain information.

As Slate’s Farhad Manjoo explains, “Although we’re more likely to share information from our close friends, we still share stuff from our weak ties—and the links from those weak ties are the most novel links on the network. Those links from our weak ties, that is, are most likely to point to information that you would not have shared if you hadn’t seen it on Facebook.” Since the majority of users’ social ties on Facebook would qualify as weak, the research suggests that in aggregate, the News Feed surfaces information that users would not have otherwise seen.

While internally-conducted research that yields a positive verdict on Facebook’s societal impact demands a certain level of skepticism, (though Bakshy tells Manjoo that the study will be submitted to peer-reviewed academic journals). As it stands, the research does not necessarily disprove the notion of an online echo chamber, but it certainly challenges it. In addition, the research illuminates the promise and potential of applying data science to big datasets — in this case 75 million URLS shared among 253 million users — to address social science questions.

About the Author


CSS :first-child, :nth-child, and :last-child are not like :eq
CSS :first-child, :nth-child, and :last-child are not like :eq

One mistake I've seen made a few times is the notion that CSS's nth-child pseudoselector acts like jQuery's...

Announcing Pivotal Tracker for iOS 1.5
Announcing Pivotal Tracker for iOS 1.5

Today we released the latest Pivotal Tracker iOS app to the App Store. This is a major release that inclu...

SpringOne 2021

Register Now