Beyond the bubble? Exploring the 2016 US election echo chambers
Immediately following the 2016 U.S. election, a range of journalists, academics and other commentators speculated on the media’s role in Trump’s victory. Alongside arguments that traditional media had failed to act as a watchdog against Trump’s lies (or ‘post-truths’), blame was assigned to social media in two ways. First, commentators argued that Trump’s success (and the disbelief on the part of Clinton supporters) was a product of the filter bubbles, or news echo chambers, that both sides inhabit. Second, a related argument was that these bubbles facilitated the spread of fake news, and more generally pushed news consumption further away from ideal political discourse and towards sensationalism, conspiracy theory and negatively charged emotions.
While the importance of the filter bubble has never been more apparent, the substance of these concerns are not new. Writing in 2001, Cass Sunstein argued that the Internet greatly expanded a trend in which media deregulation facilitated the growth of echo chambers. Within these echo chambers, moreover, extreme opinions tend to dominate. In 2011, this argument was revived through Eli Pariser’s concept of the filter bubble. Social media algorithms may do a wonderful job of recommending music and movies we like, Pariser notes, but this same mechanism applied to politics is dangerous.
Inspired by the concept of filter bubbles, various efforts have been made to reproduce an experience of them. The Washington Post’s Blue Feed, Red Feed project displays two Facebook feeds based on the news shared by known liberal and conservative Facebook pages. Similarly, The Guardian conducted an experiment asking conservatives and progressives to try living in the opposite camp’s bubble. This was operationalized through two fake Facebook accounts in which only partisan sources from one side were ‘liked.’
These projects seek to expand as well as challenge current efforts to map the election filter bubbles. How would the filter bubble look if we mapped it differently, for instance through the tweets of Trump and Clinton supporters? Might we have a more accurate view if we differentiate between ‘establishment’ and ‘fringe’ factions on both sides? What can we learn from changing our assumptions about the ‘content’ of filter bubbles? What if instead of news we took images, videos or emotions as our starting point for mapping these echo chambers? Finally, how can we repurpose research on filter bubbles to find their opposite, namely the common ground topics and interests that may offer opportunities for conversation and understanding across party lines?
These projects were facilitated by researchers Michael Stevenson and Alex Gekker. Data visualizations and facilitation provided by Gabriele Colombo. Custom tools and facilitation by Erik Borra and Emile den Tex.