Wednesday, April 24, 2013

The Potential for Digital Data from Social Media

One of the challenges of research support is to keep up to date with the impact of digital technology - and in particular 'social media' on research methods.Whilst there is great deal of research now being conducted into the influence of social media on Society, there is also a growing interest in how they may be used as a tool in harvesting good quality research data.

I recently attended a workshop organised by the National Centre for Research Methods at the Royal Statistical Society in London. 'New Social Media, New Social Science - Blurring the Boundaries' was an event that offered delegates the opportunity to look at some of the issues involved. This was a culmination of a year long series of events, workshops and podcasts as part of  Networks for Methodological Innovation. The project is now winding down but there are plans to keep collaborations going and maintain the network that has been developed as much as possible.

The presentations I attended fell into a number of key areas;

The ethical foundations of using Social Media;  is it encouraging participation or exploitation; the need for strong advocacy of ethics committees; the commercial sector as a driver of social media data and the backlash of public scandals,

The impact of 'Big Data'; working with 'byproduct data' ; mistaking volume for quality; Stapleton's Volume/Velocity/Variety definition; problematic sidelining of context and metadata; hybrid approaches (cross-referencing it with traditional sources),

Research opportunities;  framing research around browsing habits; the goldmine of location data; mismatch of needs of academia and commercial social media owners; examples of successful tools like Mappiness  (an app created by LSE researchers) and Tweetometer being used by traditional media outlets,

Visualisation; powerful tool for analysing social data; great potential for adding impact to findings; opportunities for identifying new connections; dangers of mistaking correlation for causation,

Interesting comparisons were drawn with past debates about the use qualitative data as a source of robust social research data. The need for attention to context and creation of metadata to ground the data was the same. There are also similar concerns about quality and verification and problems of re-purposing. However there were also unique issues highlighted. Use of this material is very much driven by the commercial interests developing the technology who have a great concern with novelty and innovation in how the platforms work. It was noted a number of times that research projects were developed around key elements of a particular form of social data, only to have those elements dropped from newer versions.

Most of these debates are ongoing but if nothing else they should make researchers aware of some of the issues involved in working with this kind of digital data. As investigative projects at the LSE continue to take an interest in this area - as a subject for study and as a tool - collections of data will naturally accumulate. It is therefore important that this data is properly preserved and wherever possible promoted and made accessible as research resources.


  1. Having completed my dissertation where I examined Facebook groups as an additional component to a face to face learning community, thus creating a virtual learning community. I found that the potential for the use of, and study of how social media is used, is immense.

    The research opportunities, especially with software like Nvivo, of following dialogue and online interactions for qualitative research is exciting! In the section you label Research opportunities I feel you miss the authentic online interactions that can be captured (especially in private groups) which hold potential for further research.

    I believe that social media, and the interactions therein, are a rich and burgeoning field for qualitative research!

    ~ JH

  2. A very good point Jerome. Like a lot of qualitative research the challenge can be to capture that feeling of involvement. In this instance though - as you note - the ebb and flow of the interaction itself can become a key area for study.

    Your mention of private groups is a useful one too as it goes right to the heart of one of the debates about using such networks for research. What are the methodological and ethical issues a researcher needs to consider about using such groups in research? Are there conflicts that need working through in terms of informed consent, participation or simply issues of representativeness.