Nikolov, D., Oliveira, D., Flammini, A., & Menczer, F. (2015). Measuring online social bubbles. Peerj Computer Science, 1, e38. http://dx.doi.org/10.7717/peerj-cs.38

The purpose of this article is to quantitatively measure the phenomenon of social bubbles on social media. The researchers assert that there is a tendency for people to only interact with a narrow range of information on social media, mainly information or sources they agree with or have shown previous interest in. This phenomenon, they state is due to algorithmic personalization features on social media and search engines. In addition, it is due to the tendency for users to only expose themselves to the opinions of like-minded individuals.

In this research, the authors analyzed, coded and categorized anonymous web traffic from over 100,000 people over 41 months, 18 million link clicks from the AOL search engine log over 3 months and 1.3 billion public tweets from 89 million Twitter users over 13 months. The research results showed that targets reached via social media are significantly less diverse than targets reached via search engines for all forms of web traffic including news. This result suggested that users are exposed to less diverse information options on social media versus via search engines and other information seeking platforms.

The research contributes to knowledge and research of online information behaviour. The authors suggest that the technology is not necessarily the cause of social bubbles as people generally tend to prefer to access information they agree with and disregard information they disagree with. However, they state that the technology makes it easier for people to act on that social tendency. Therefore, I believe the sociotechnical theoretical framework followed by the authors is social determinism.

The authors articulated their idea comprehensively and used sound research methods and large sample datasets to test their hypothesis. The limitation of the research pointed out by the authors include the fact that the methods used in the research can be argued to prove social collective biases on social media but cannot prove individual biases. In addition, the results show that users are exposed to less diverse information on social media versus search engines, however, the results cannot prove causality.

Keywords: social media, search engines, self-filtering, filter bubble, social bubble

Page author: Salim Zubair