Social media research

• Types of social media research: active versus passive

  We can consider two different types of social media research: the active use of social media where researchers participate, or passive use where data are collected and may not involve interaction with participants.

  Active:

  Social media is increasingly used as a platform for recruiting and interacting with research participants. Activity on social media presents a challenge to protecting participants' confidentiality because it can leave a visible, permanent and traceable online record that it might not be possible to fully delete.

  When planning or seeking ethical approval for this type of research, researchers should:

  • arrange to keep personal and professional social media separate as much as possible, for example adjusting privacy settings to avoid sharing personal information with participants, not accepting ‘friend’ requests from participants, and keeping communication professional and not personal
  • use direct messaging services and/or email to communicate with potential participants rather than engaging in discussion threads in public posts
  • ensure they will never publicly post private phone numbers, emails or addresses, nor require this of their participants
  • ensure that the data collected will only be used for the purposes detailed on the participant information sheet
  • ensure that no data such as images or audio-visual recording relating to participants are collected without the explicit consent of the participants
  • avoid disrupting (including deceiving) online communities, especially those which may be considered by their users as private spaces.

  Passive:

  Passive social media research comprises passively 'scraping' or collecting data from social media sites, or reusing social media data collected by others. There may be no direct interaction with participants. 

  If there is no interaction with participants, then it may not be practical to obtain informed consent. Indeed it may not be possible: for instance Twitter may classify requests for consent as spam (Kotsios et al. 2019). 

  If consent is not being sought from individuals for the collection of primary data (i.e. the researcher is collecting the data themselves) the researchers should:

  • take care to ensure the rights of the social media users in their research.
  • consider what social media users might think about researchers repurposing their data without permission. See for instance (Reuter et al. 2019) (Norval and Henderson 2019) (Hutton and Henderson 2015).
  • note that direct quotations scraped from social media sites (e.g. Twitter posts) cannot be considered as anonymous data and cannot be pseudonymised as they can always be linked back to their author via simple web searches. Special consideration therefore needs to be given to protecting the identity of social media users.

  If the research is using social media data already collected by others, this should be in accordance with any consent originally obtained from participants and the requirements of the data source (whoever collected the data originally). If consent was not obtained, then the same consideration should be given as with primary data.

  Images or audio-visual recordings of users should not be collected without the users explicit consent.

• Legal considerations

  Just because something has been posted “publicly” does not mean that social media users lack privacy protection (Edwards and Urquhart 2016). The GDPR means that data protection principles such as storage limitation, data minimisation and transparency apply (Kotsios et al. 2019).

  A further complication comes from using social media data to infer special category data, such as in the sexual orientation study described earlier. The Information Commissioner’s Office guidance highlights the importance of “whether you are using an inference linked to one of the special categories to influence your activities in any way” (Information Commissioner’s Office 2019). Even inferring special category data brings the requirement to manage and protect this data as you would special category data, in accordance with GDPR.

  There is further information on data protection, confidentiality and definitions of special category data in the Data Protection and Confidentiality guidance web page. If you are not sure if this applies to your research, you should seek guidance from Data Protection: dataprot@st-andrews.ac.uk  

• Methodological considerations

  A broader consideration is whether social media data, whether collected actively or passively, are the most appropriate source of data for a particular research question. The use of social data may introduce various methodological issues around bias, data quality and evaluation (Olteanu et al. 2019). The lack of reproducibility in much social media research may also make it difficult to use such research as a building block for further work (Hutton and Henderson 2018).

• Questions for researchers to consider

  Some questions that may be useful for researchers to consider as they design your research project or complete their application for ethical review:

  • Is the use of social media the most appropriate method? (Olteanu et al. 2019)
  • Can you obtain informed consent?
    • If not are you safeguarding people’s rights? (Kotsios et al. 2019) 
    • Are you inferring or using special category data? (Information Commissioner’s Office 2019)
  • How will you protect the identity of your participants during data collection, analysis and dissemination?
  • Are you making assumptions about whether people would be willing to let you use their data? (Reuter et al. 2019)(Norval and Henderson 2019)
  • Are you following any subject-specific codes or guidance? e.g. (Benton, Coppersmith, and Dredze 2017)(Bowser and Tsai 2015)(Association of Internet Researchers 2019)

  The ethical review process is not intended to be overly bureaucratic or make researchers risk-averse (Molina and Borgatti 2019) - these questions and the guidance are to help researchers improve their research project and protect participants from harm.

• Codes and guidelines

  Researchers have traditionally relied on “personal ethics” when deciding what is acceptable with regard to social media, which risks inconsistent decision-making (Samuel, Derrick, and Leeuwen 2019). Researchers in specific fields are encouraged to consider relevant codes of practice or these papers which highlight potential ethical tensions:

  • The Association of Internet Researchers ethics guidance (Association of Internet Researchers 2019)
  • The British Psychological Society’s guidance on internet-mediated research
  • The UK Research Integrity Office's advice on good practice in research: internet-mediated research
  • Ethical web research guidance [Bowser and Tsai (2015) (note that this is somewhat US-centric)
  • Guidance for researchers using social media for health (Benton, Coppersmith, and Dredze 2017) 
  • A Taxonomy of Ethical Tensions in Inferring Mental Health States from Social Media (Chancellor et al. 2019)- paper on various concerns about mental health research
  • Questions for those using social media for fisheries research (Monkman, Kaiser, and Hyder 2017)

• Bibliography

  You can find below the references and resources used to develop this guidance and which may provide useful further reading:

  • Andrews, Lori. 2012. I Know Who You Are and I Saw What You Did: Social Networks and the Death of Privacy. New York: Free Press. (Book)
  • Association of Internet Researchers. 2019. “Internet Research: Ethical Guidelines 3.0.”
  • Benton, Adrian, Glen Coppersmith, and Mark Dredze. 2017. “Ethical Research Protocols for Social Media Health Research.” In Proceedings of the First ACL Workshop on Ethics in Natural Language Processing. Association for Computational Linguistics.
  • Bowser, Anne, and Janice Y. Tsai. 2015. “Supporting Ethical Web Research.” In Proceedings of the 24th International Conference on World Wide Web. ACM Press.
  • Boyd, danah m., and Nicole B. Ellison. 2007. “Social Network Sites: Definition, History, and Scholarship.” Journal of Computer-Mediated Communication 13 (1): 210–30.
  • British Psychological Society. 2018. “Supplementary Guidance on the Use of Social Media.”
  • Chancellor, Stevie, Michael L. Birnbaum, Eric D. Caine, Vincent M. B. Silenzio, and Munmun De Choudhury. 2019. “A Taxonomy of Ethical Tensions in Inferring Mental Health States from Social Media.” In Proceedings of the Conference on Fairness, Accountability, and Transparency (FAT* ‘19).
  • Edwards, Lilian, and Lachlan Urquhart. 2016. “Privacy in Public Spaces: What Expectations of Privacy Do We Have in Social Media Intelligence?” International Journal of Law and Information Technology 24 (3): 279–310.
  • Flick, Catherine. 2016. “Informed Consent and the Facebook Emotional Manipulation Study.” Research Ethics 12 (1): 14–28.
  • Hutton, Luke, and Tristan Henderson. 2018. “Towards Reproducibility in Online Social Network Research.” IEEE Transactions on Emerging Topics in Computing 6 (1): 156–67.
  • Hutton, Luke, and Tristan Henderson. 2015. “‘I Didn’t Sign up for This!’: Informed Consent in Social Network Research.” In Proceedings of the 9th International AAAI Conference on Web and Social Media (ICWSM), 178–87. Oxford, UK.
  • Information Commissioner’s Office. 2019. “What Is Special Category Data?”.
  • Kotsios, Andreas, Matteo Magnani, Luca Rossi, Irina Shklovski, and Davide Vega. 2019. “An Analysis of the Consequences of the General Data Protection Regulation (GDPR) on Social Network Research."
  • Mayer-Schönberger, Viktor. 2009. Delete: The Virtue of Forgetting in the Digital Age. Princeton: Princeton University Press.
  • Molina, José Luis, and Stephen P. Borgatti. 2019. “Moral Bureaucracies and Social Network Research.” Social Networks, November.
  • Monkman, Graham George, Michel Kaiser, and Kieran Hyder. 2017. “The Ethics of Using Social Media in Fisheries Research.” Reviews in Fisheries Science & Aquaculture 26 (2): 235–42.
  • Norval, Chris, and Tristan Henderson. 2019. “Automating Dynamic Consent Decisions for the Processing of Social Media Data in Health Research.” Journal of Empirical Research on Human Research Ethics.
  • Olteanu, Alexandra, Carlos Castillo, Fernando Diaz, and Emre Kıcıman. 2019. “Social Data: Biases, Methodological Pitfalls, and Ethical Boundaries.” Frontiers in Big Data 2 (July).
  • Reuter, Katja, Yifan Zhu, Praveen Angyan, NamQuyen Le, Akil A Merchant, and Michael Zimmer. 2019. “Public Concern About Monitoring Twitter Users and Their Conversations to Recruit for Clinical Trials: Survey Study.” Journal of Medical Internet Research 21 (10): e15455.
  • Samuel, Gabrielle, Gemma E. Derrick, and Thed van Leeuwen. 2019. “The Ethics Ecosystem: Personal Ethics, Network Governance and Regulating Actors Governing the Use of Social Media Research Data.” Minerva 57 (3): 317–43.
  • Tromble, Rebekah, and Daniela Stockmann. 2017. “Lost Umbrellas: Bias and the Right to Be Forgotten in Social Media Research.” In Internet Research Ethics for the Social Age: New Cases and Challenges, edited by Michael Zimmer and Katharina Kinder-Kurlanda, 75–91. New York: Peter Lang Publishing.
  • Wang, Yilun, and Michal Kosinski. 2018. “Deep Neural Networks Are More Accurate Than Humans at Detecting Sexual Orientation from Facial Images.” Journal of Personality and Social Psychology 114 (2): 246–57.
  • Weaver, Matthew. 2018. “Cambridge University Rejected Facebook Study over ‘Deceptive’ Privacy Standards.” The Guardian.
  • Zimmer, Michael, and Katharina Kinder-Kurlanda. 2017. “Introduction.” In Internet Research Ethics for the Social Age: New Cases and Challenges, edited by Michael Zimmer and Katharina Kinder-Kurlanda, XIX–XXVII. New York: Peter Lang Publishing.

• Acknowledgements

  This guidance was developed by:

  Dr Tristan Henderson, School of Computer Science 
  Dr Juliette Summers, School of Management
  Dr Vicky Ward, School of Management