Last Friday, I attended an excellent workshop on ‘Anonymity, Identity and Credibility: Challenges of Using Social Media Data for Research’, run by the Social Media Knowledge Exchange (SMKE), based at Centre for Research in the Arts, Social Sciences and Humanities (CRASSH) at Cambridge University. It was organised by Dr. Ella McPherson, an LSE fellow in Media and Communications and an SMKE scholar at Cambridge University, and the workshop was attended by PhD students and academics from institutions around the UK from a variety of disciplines, including History, Sociology and Media.

The workshop featured four presentations from speakers from NGOs at the forefront of using data from social media data for campaign activities, policy making and advocacy; especially from areas where conventional sources of information may be difficult or impossible to get (such as conflict areas or disaster zones). The workshop also highlighted some of the techniques these organisations use to verify the credibility of social media data, many of which could cross over to how academic researchers use social media sources. These included:

  • Verification: what is the actual source of their information? Can this information be validated by offline sources?
  • Triangulation: how many other sources are reporting the same information? Do their views align?
  • Authentication: who is re-posting this information? Has this information been picked up by traditional media sources?

Sarah Jackson, Communications manager at Womankind Worldwide, talked about how she evaluated a source from Twitter, which she had storified to share with us as a case study.  She was able to give some handy hints and tips on making sure information from social media is credible, including:

  • Evaluate the profile of the source. How many subscribers/followers do they have vs the number of people they follow/subscribe to? What sort of information have they been sharing? Does their profile look professional?
  • Evaluate the content of the information. Are unverified or spurious numbers and figures being used? Are exclamation marks used excessively? Are there any grammatical errors? How many hashtags are being used?
  • Evaluate the tone of the information. Is the source trying to be sensationalist? Neutral tones would suggest that the source is only interested in sharing the information, and is not biased in the situation it pertains to.

Maha Abu Shama from Amnesty International gave a fascinating presentation via Skype on triangulating sources of YouTube videos charting the conflict in Syria, and the toolkits Amnesty provide which advise people on how to effectively campaign on social media. Maha highlighted some of the benefits of using social media in conflict zones, such as being able to get news of human rights violations quickly and getting on the ground photo and video evidence of violations. However, she also mentioned some of the challenges of working with social media from conflict zones, such as ethical concerns about sharing information containing graphic violence, misinformation and lack of details for sources, such as times, dates and locations, which could be used to verify claims of human rights abuse.

Emma Marshall from Peace Brigade International UK (PBI) spoke about the importance of offline contacts in verifying online information. She spoke about how PBI use social media to disseminate information and raise awareness. She also echoed Sarah Jackson’s point that NGOs increasingly act as curators of information on social media, which in turn provides credibility to that information for downstream users, such as traditional media.

The final presentation was given by Richard Dent, founder of Crowd Aid Exchange; a social media platform which aims to map impending disasters, inform people about them and connect victims of disasters to people who would like to help in the area affected. Richard spoke of innovative ways that the platform could help raise awareness of disasters through geo-mapping, and for rating the reputation of individuals by linking their existing online presence to the platform to give users an idea of the reliability of their information.

Some of the common pros and cons of using social media data were:

Pros

  • Access to speedy, real time information on developing situations
  • Access to unique, unfiltered information
  • Using data which gives agency to the people involved in that situation

Cons

  • Social media may be used to spread propaganda and misinformation
  • Social media provides a large quantity of data, but variable quality of data – sources often do not contain dates, times, locations or the identities of the sources
  • Ethical concerns – could vulnerable parties, such as children, be affected by the information sourced or shared from social media, such as videos of violence in war zones?
  • Privacy issues – sources may not wish to be cited, particularly in cases where they may be at risk of harm

In the afternoon group-work session, we discussed the common challenges participants faced when using social media sources in their own research, and came up with three overarching questions researchers should bear in mind when using social media data:

  1. To what extent do issues with using social media data apply to research in general? E.g, are there similar issues which researchers have to contend with using traditional sources, such as peer-reviewed journals? Could a researchers own bias be an issue?
  2. How much does the veracity of the source actually matter? Are they just trying to make a point, or get attention to their cause, and does that affect your research? What value does the veracity of the source bring to your research? E.g, the veracity of a source may not be all that relevant if the study you’re doing is attitudinal or quantitative research.
  3. To what extent does using these sources reinforce existing power structures regarding information dissemination and management? Does using social media data counter, mitigate or reinforce the views of certain parties, such as governments, military forces, academia and/or industry? E.g Does putting more trust in a tweet from a traditional media source rather than a lay-person during a developing situation give an accurate account of the situation, or reinforce bias towards the view of that media source and its affiliates?

This was a very enjoyable and interesting workshop, and my thanks to Ella McPherson for inviting me to attend.