The Battle of Stokes Croft on Youtube: The ethical challenges associated with the study of online commentson 30 April 2014
By Paul Reilly
Should online comments be considered published artifacts or the property of human participants? Clearly if such content is treated as the former then there should be no need to seek informed consent from individual commentators for the use of their contributions in academic research. Early work in the field of Internet research ethics suggested that such data could be considered ‘public’ by virtue of its publication on a website that could be accessed without the need to use a password. However, such an approach might be inappropriate for the study of ‘semi-public’ social media sites such as Facebook. Researchers should consider the different expectations participants might have in relation to the publicness (or not) of the online environments in which they operate.
A ‘human participant’ perspective, which obliges researchers to take appropriate measures to minimize the potential harm that might be caused to participants through the use of their content, has been evident in the most recent iterations of the ethical framework for the study of online communities developed by the Association of Internet Researchers (AoIR). These guidelines have encouraged researchers to obtain informed consent from participants (if possible) or anonymise datasets in order to protect their privacy. Nevertheless, research into the role of social media during the disturbances seen in several English cities including London and Manchester in August 2011 arguably treated user-generated content as an artifact through the identification of users in data visualisations without seeking their consent to do so prior to publication. Such breaches of anonymity appeared to be justified on the basis that many of these Twitter users were public figures, such as politicians and journalists, who usually had the ability to influence or control the spread of information about them. However, concerns continue to be raised in relation to the potential reputational harm that might be inflicted upon those individuals who are identified in studies such as Reading the Riots but do not possess such resources.
Natasha Whiteman argues that such issues are best addressed through the development of localized ethical stances, which take into account not only the guidance provided by organisations such as the AoIR but also the specific circumstances in which the data is collected. This can be illustrated by the strict ethical stance that I adopted in my study of YouTube footage depicting the so-called ‘Battle of Stokes Croft’ in April 2011. My analysis of 1018 YouTube comments was designed to examine whether the footage of the ‘anti-tesco’ riot, presumably shared by eyewitnesses online in order to highlight the allegedly ‘heavy-handed’ tactics of the Avon and Somerset Constabulary, had generated sympathy amongst commentators for the local residents. The study also examined the extent to which the views of these commentators appeared to have been influenced by the news media framing of the riot as a manifestation of the ongoing campaign against a newly opened Tesco supermarket in the area.
A number of ethical issues were raised by the use of YouTube comments in the study. Did YouTube’s disclaimer that users could not expect confidentiality in relation to the use of their content mean that I was free to use these comments without seeking permission to do so from their authors? What, if any, measures would be taken to protect the privacy of these ‘unaware’ participants? Such questions were addressed in my project through the adoption of an ethical stance that was influenced by the ‘human subject’ perspective. Obtaining informed consent from those responsible for these comments was considered neither feasible nor appropriate for the study. Participants were not identified either via username or a pseudonym and their comments were paraphrased in order to illustrate key themes that emerged from the dataset. They would only be identifiable in relation to the video under which they had posted. Such a decision was justified on the basis that there would be no public benefit served in exposing these participants to the risk of prosecution should their ‘semi-published’ comments be deemed offensive by the authorities. In this way, the study employed a strategy that attempted to provide the maximum level of disguise possible for these unaware participants.
This strict ethical stance not only conformed to the ‘do no harm’ principle implicit in many of the ethical frameworks for online research but also met the requirements of the study. The results were presented in such a way as to allow the reader to view the key themes that emerged in these comments in relation to the policing of the Stokes Croft riot. Direct quotes were not considered necessary as the paraphrasing of comments allowed the voice of participants to be heard without exposing them to any potential harm. While no such strategy can ever fully guarantee anonymity for participants, this at least made it harder for these comments, many of which used offensive language to describe the ways in which the police responded to the Stokes Croft riot, to be located using search engines.
But this is not the only ethical stance that could have been adopted for the study. Some researchers would have used direct quotes from this dataset in academic publications on the basis that YouTube is widely perceived to be a ‘public’ online space. Others might dispute the notion that scholarly research should develop strategies in order to protect online participants from the negative consequences that might occur as a result of their online behaviours. This reflects the subjective nature of qualitative research and the difficulty in prescribing universal rules for researchers that define what is (and is not) ethical practice. Researchers must explicitly justify the decisions they take with reference to the ethical frameworks of not only their institutions but also the online communities that are the focus of their studies. This doesn’t mean that the default position should be to please participants through the redaction of potentially harmful content from datasets. Rather, researchers need to assess each specific research context and tailor their ethical stance accordingly. They should weigh up the public benefit of verbatim reproduction of social media texts against the potential risks associated with the use of this data without the informed consent of its authors. Factors such as the contextual notions of privacy held by participants, the perceived openness of online platforms, and the requirements of the study should all be explored as part of this process. Researchers should be empowered to make informed ethical decisions that protect the right to privacy for unaware participants when it is appropriate to do so.
Paul Reilly is Lecturer in Media and Communication at the University of Leicester. He tweets as @PaulJReilly. His article ‘The ‘Battle of Stokes Croft’ on YouTube: The development of an ethical stance for the study of online comments,’ will be published in SAGE Cases in Methodology in May 2014.
Image: Spencer E Holtaway CC BY