Digital Data 2019, Week 04

Week 04: The Shape of Privacy

While the extent to which Facebook and other social media networks are participatory may be up for debate, the opportunities for new and expanded research provided by these platforms is not. Earlier research around audiences of corporate media channels primarily used focus groups or polling of a carefully-selected sample, but now every interaction of every single user can be collected, processed, and analyzed (Fuchs, 2014, p. 132). Studies might focus on what social media users have directly composed and knowingly shared; for example, a recent content analysis focused on using Facebook group posts to determine whether and how depressed teenagers were seeking out support (Lerman et al., 2017). Social media sites also allow us to access new information provided by users only indirectly; a network analysis provided researchers with the structure of Brexit-related “echo chambers” on Facebook, allowing them to evaluate the differing emotional responses and reactions within those groups (Del Vicario, Zollo, Caldarelli, Scala, & Quattrociocchi, 2017). These same questions might have been possible to investigate without using Facebook data — for example, using surveys, interviews, and the like — but they certainly would have been more time-consuming, possibly prohibitively so.

A panel mapping the emotional response of Facebook users to posts on controversial topics (Del Vicario, Zollo, Caldarelli, Scala, & Quattrociocchi, 2017).

Accessing and using data from Facebook presents a series of new challenges. Facebook’s business model relies on using its data to sell advertising; the better the data they collect, the more money corporations will pay in order to target their demographic of interest. As a result, Facebook has a strong incentive to protect their data, turning as little of it over to researchers as possible. While the company does provide an API, that access includes restrictions on what kind and how much data can be gathered at any one moment, whether and how that can be shared with other researchers for the purposes of peer review and replication, and can (and often does) change at a moment’s notice. This requires researchers to understand not only the requirements and expectations of their discipline, but to keep up with a constantly-changing set of technical requirements and policies from each individual data source.

Using Facebook data also raises new ethical challenges for researchers. As Fuchs points out, “social media do not automatically constitute a public sphere” (2014, p. 127) in terms of ownership and participation and also in terms of user expectations. What is said on Facebook is not necessarily intended for public consumption and users may not be fully aware that their statements are being used for research purposes. In an interview situation, research subjects can often be protected by removing identifying information from their statements, but the digital and indexed nature of social media means authorship can often be reconstructed via search engine. Many researchers end up taking note of these ethical dilemmas but not changing course. Fuchs takes an even more radical perspective, saying that “overdoing Internet research ethics […] results in a de facto censorship” (2014, p. 60). Many sociologists are looking for consensus on how the discipline will handle new and emerging privacy concerns, but this raises an interesting question: what is the shape of conversations around privacy in sociology as a whole?

Del Vicario, M., Zollo, F., Caldarelli, G., Scala, A., & Quattrociocchi, W. (2017). Mapping social dynamics on Facebook: The Brexit debate. Social Networks; Amsterdam, 50, 6. https://doi.org/10.1016/j.socnet.2017.02.002

Fuchs, C. (2014). Social Media: A Critical Introduction. London. https://doi.org/10.4135/9781446270066

Lerman, B. I., Lewis, S. P., Lumley, M., Grogan, G. J., Hudson, C. C., & Johnson, E. (2017). Teen Depression Groups on Facebook: A Content Analysis. Journal of Adolescent Research; Thousand Oaks, 32(6), 719–741. http://dx.doi.org.proxy.library.vcu.edu/10.1177/0743558416673717

Tagged