Towards expanding the filter bubble theory: an exploratory study into the relation between user behavior and content diversity on Facebook

Abstract

A lot has been written on algorithms and content diversity. The most notable contribution arguably came from Eli Pariser in 2011, when he coined the ‘filter bubble theory’. The theory was critically acclaimed, but drew criticism from many academics. A lot has been written on the phenomenon since, but little attention so far has been paid to the composition of Facebook’s timeline and how user behavior, like scrolling and interacting with politically charged content, affects the composition of Facebook’s timeline. In this master’s thesis, the central question is therefore: “How does a sample of Facebook users interact with politically charged content on Facebook, and how does liking and scrolling specifically affect the distribution of worldview challenging and worldview affirming content on Facebook timelines?” Relying on new media studies and political communication theory, this question was researched through a combination of online ethnographic approaches and experimental digital methods, like surveys, interviews and analyzing the behavior of Facebook’s algorithm. This resulted in answers that can be divided into two categories: user behavior and the composition of Facebook’s timeline. The surveys and interviews show that Facebook users in this sample consider themselves quite open to other people’s opinions, have a rich information diet, do not rely solely on Facebook for their news consumption, and the average Facebook user in this sample often sees content appearing in their timelines that they don’t agree with. As emerged from the interviews, most selected users tend to ignore worldview challenging content, but some also admitted to ‘lurking’, or checking out comments. By doing this, they deliberately expose themselves to ideological content that they do not agree with, which could possibly be done to self-assess. Finally: most Facebook users surveyed log in to Facebook daily and usually spend more time per session and keep scrolling. An analysis of the experimental digital method shows that, despite only interacting with worldview affirming content, the timeline is not a strict filter bubble. The more the user scrolls, the more diverse content appears in the timeline. Although this research has serious limitations, primarily because of its sample size and the position of the researcher, it is fair to say that the filter bubble needs expansion and further critical research. The filter bubble is more like a bottleneck: the first couple of posts show a lot of worldview affirming content, but after that the content diversifies quickly.