Dealing Ethically with Social Media Data

The next speaker in this ANZCA 2017 session is Kim Barbour, whose focus is on ethical engagement with research participants in social media research. Social media research can be understood as human subjects research, yet we often do not have direct contact with the people whom we study: their communicative activities are being gathered through automated means, and the subjects are not usually even aware of this fact.

Yet this approach is often poorly addressed by conventional ethics approval processes at universities, which assume such direct contact and default to a historically informed preference for the anonymity of research subjects (e.g. in interviews or focus groups). In researching public online engagement through social media this preference may no longer be entirely appropriate; there is also a strong argument that can be made in favour of revealing the identity of participants in order to give them credit for their contributions.

If we conceptualise social media posts as somewhere between 'published' and 'semi-published', they are also covered by copyright, and could be considered from a legal perspective as literary or artistic works. This creates new obligations for the researcher: most importantly, to cite correctly and acknowledge the original author, rather than to change texts and obfuscate authorial identities in order to maintain anonymity. These diverging requirements are not easily reconciled.

It may be possible to offer participants the choice to appear in eponymous, pseudonymous, or anonymous form in published research; researchers also need to be aware of the various implications of naming participants in these ways, as well as of the impact of such naming on what is able to be reported in the research – naming participants also confers on them a greater sense of ownership of what is reported, after all. Such complications are not inherently negative, but they do need to be addressed carefully.

On the other hand, the removal of social media handles while reproducing the messages themselves treats messages as public utterances in a public space, but complicates the tracing of such messages back to the original authors while not rendering such a trace-back entirely impossible; this mitigates risk without avoiding it altogether, while simultaneously also removing credit while not making it entirely impossible for authors to claim such credit.

The overall challenge, then, is to balance risk and credit, and protection and empowerment; this remains a difficult problem, and all of us are charged with continuing to address it.