Article Preview
TopIntroduction
Online social networks (OSNs) have become a popular way for users to interact with their friends via the display of a user’s personal details, friendship connections and interactions with other users on OSN profiles. In December 2012, Facebook, the leading OSN, had more than one billion active users (Facebook, 2013).
Boyd and Ellison (2008) have described three ingredients that constitute an OSN:
- 1.
Allowing a user to make a public or semi-public profile inside a system that is bounded. A bounded system represents a set of activities that are interrelated and come together to make a single entity, which in this case is the OSN;
- 2.
Bringing together a list of other users with whom they share a connection and allowing the user to view these;
- 3.
Allowing users to travel along their set of connections and the connections made by other users within the system.
Depending on the users’ attitudes on privacy, data on a user’s OSN profile, which includes personal details, friendship connections and interactions with other users, can provide a rich source of useful information for extraction for a variety of researchers, from different domains that range from computer science to sociology.
Various approaches are available for extracting publically available OSN profile data. Extraction approaches can be split into two separate categories: non-automated and automated.
Non-automated approaches utilise surveys and interviews. They have been used in research studies such as the ones by Dwyer, Hiltz and Passerini (2007), Gibson (2007), Govani and Pashley (2005) and Strater and Richter (2007). Non-automated approaches include manual scrapping of OSN profile data, having conversations with people, joining forums to discuss issues of interest and listening in on conversations between other people.
During the last couple of years, more studies have been done with data extracted using automated extraction approaches, e.g. by Catanese, De Meo, Ferrara, Fiumara and Provetti (2012); Caverlee and Webb (2008); Pfeil, Arjan and Zaphiris (2009).
At present, there is significant debate as illustrated by BBC News (2011); Giles, Sun and Councill (2010); Zimmer (2010) about the issue of ethics in the automated extraction of publically available OSN profile data by researchers. This is due to the desire of some OSN users for control over their personal details, their curiosity about who can access their profile data and need to know what their personal details are used for. In the past, researchers could use automated means to extract public OSN profile data without informing the profile user.
Previous research that was carried out in the area of ethics and accessing OSN profile data by Wilkinson and Thelwall (2011); Wilson, Gosling and Graham (2012) ; Zimmer (2010) has discussed the ethical issues of extracting publically available profile data, but did not explore the unique ethical challenges that automated extraction from OSNs brings in comparison to non-automated extraction.
Wilson et al. (2012) explored automated extraction via data crawling and ethics in their overview of Facebook, whilst Wilkinson and Thelwall (2011) described techniques to access personal information including automated data extraction via specialist software. The Wilkinson and Thelwall (2011) study surveyed areas such as anonymity and informed consent with regards to extracting publically available OSN data and using it for research in an ethical manner. However, the study did not focus on automated data extraction or explore issues such as the storage of extracted data, OSN policies and the position of OSNs with regards to automated data extraction and the distribution of datasets. Zimmer (2010) explored the mistakes made regarding ethics when the Tastes, Ties and Time research study took place using non-automated approaches to extract OSN profile data.