Social media as data source
When I talk here about social media, I am referring to online content that is publicly-generated, typically by unpaid authors and creators who do not have specific skills or training in writing or other creative arts – although of course social media is also used by professional creatives and ‘influencers’ with considerable followings.
Platforms such as Twitter, Facebook, Mumsnet, YouTube, Instagram, TikTok and Reddit provide obvious examples, but social media can also be considered to include blogs and perhaps comments sections of websites such as online newspaper comments, or Amazon reviews.
These are all spaces that allow amateurs to publish online without cost and without having to achieve prior approval from any gatekeeper (although they may of course be removed by site administrators after publication). The internet is replete with opportunities for members of the public to ‘publish’ their commentary or creations, from home, at the push of a button.
The ease of online publishing, and the variety of different online platforms, and online communities, offering space for variously-moderated public online communication, means that the internet is laden with ‘public opinion’ shared by diverse demographics and addressing an almost infinite array of topics. If you can’t find online discussions of the topic you’re interested in, it’s likely that you’re just looking in the wrong place!
Social media is unique in providing a space in which people can communicate with one another across vast distances in an asynchronous manner. The fact that social media platforms are open for contributions 24/7 means that exchanges between different people are not limited by differing time zones or availability. Social media offers an unparalleled opportunity for exploring public opinion amongst different sorts of people and between people from different places and different walks of life.
Social media platforms have significant potential as sources of data, and as an archive, social media platforms have advantages over other data sources due to:
- The ease with which the data can be accessed
- The broad range of topics addressed by a wide variety of different people
- The absence of researcher-influence in the production of the initial data
Understanding social media data in its context
Whilst social media can be understood to hold significant potential, it should not be considered an easy option. There are significant ethical and methodological questions over the way in which social media data should be used, interpreted or related to by researchers.
Social media contributions are well-recognized to be a ‘performance’ in which people post content that they believe reflects an image of self that they want to portray to others. As such, it’s important that researchers don’t use social media data uncritically, presuming it to be a ‘true’ reflection of opinion or interest.
Another key complication is that different social media spaces will have different cultural norms, and different agendas. These may support, and be supported by, different types of contributors and may encourage the development of different sorts of content. Moreover, one social media platform may encompass more than one set of cultural understandings, dependent on the space within the platform that is being used, or the sort of person that is interacting with the platform.
For example, Mumsnet will have a very different demographic to TikTok. However, within Mumsnet – typically thought of as a fairly homogeneous demographic - different discussion boards will have different frequenters and different cultures depending whether they are intended for ‘spilling the tea’, such as ‘AIBU’ (Am I Being Unreasonable), or for discussion of personal issues such as infertility or raising children with specific needs.
These different cultural norms are often not made explicit, but their potential influence on the data makes understanding them important for meaningful analysis and interpretation. As such, researchers should be encouraged to gain familiarity with a platform before using it as a source of data.
How to select and analyse social media content as data
Your choice of data source – which social media platform(s) you choose to draw data from – and your sampling method need to be informed by your project design: your research question or topic, the resources (especially time) you have to address it, the kind of data you want to collect or create, and the sorts of people and you are interested in.
For example, some platforms – such as Instagram, but also Facebook and Twitter – are very useful for access to user uploaded images; YouTube, Instagram and Facebook, for videos. If you’re more interested in analysing text exchanges, then Facebook, Twitter, Mumsnet and Reddit, alongside Newspaper comments sections provide great opportunities. For young people’s social media interactions, TikTok and Instagram provide useful repositories, whilst middle aged people may be better accessed through Facebook or LinkedIn and older respondents might be found in spaces like Gransnet.
Of course, these suggestions shouldn’t be considered to constitute ‘rules’ as many people enjoy spending time in online space that isn’t marketed directly at them or where they are not the majority demographic. Some social media platforms even have spaces dedicated to minority platform users, such as Dadsnet, a message board within Mumsnet.
Once you have chosen your platform, or platforms, there are a wide range of possible approaches to sampling and analysing this data. For example, you might choose to sample content from one or a small number of contributors between specific dates, you might choose to sample according to a specific topic, or to analyse data from a particular discussion thread or set of threads.
Alongside sampling decisions, you’ll also need to make a decision about how you will approach the analysis of the data. Will you choose an approach such as
Discourse Analysis that allows you to take a really deep look at the language and expression that social media users have employed in their discussions, or an approach such as
Reflexive Thematic Analysis that allows you to identify the broader themes that emerge from an online exchange. It would not be possible to cover the entire scope of potential approaches here. However, a useful introduction is provided in von Benzon and Wilkinson’s chapter on Selecting and Analysing Publicly Generated Online Content in
Creative Methods for Human Geographers.
Getting permission
It is important to bear in mind that whilst it might be technically possible – and legal - to access social media data for use in research, it might not always be considered ethical. A key concern is whether it is ethical to use material published to the internet by members of the public, without their consent. This issue is magnified when the material is either of a sensitive nature or appears to have been published without forethought or consideration of the extent of the potential audience – such as comments in online discussion threads. Of course, obtaining consent for using this sort of data is often very difficult, if not impossible.
Additionally, using social media content for research isn’t always in keeping with the guidelines for use provided by the individual platforms. For example, Mumsnet does allow researcher use, but requires researchers to write to request specific permission from the site administrators. Please look carefully at the rules of use associated with the platform that you choose.
Social media research: Some interesting examples
Some varied and interesting examples, that might provide useful ideas for developing your own research using social media as archive, might be found in the following papers:
- Sarah Pedersen has been a forerunner in this sort of research, particularly focusing her attention on the parenting website and blog, Mumsnet, such as in this example of her work.
- This Open Access paper in PLOS ONE is rather an oldie but a goodie, looking at how you might measure happiness in particular places through exploring Tweets made in those locations.
- Finally, another open access paper, explores the way in which urban inequalities can be reproduced through Instagram.
Final thought
Whilst social media platforms are replete with data that can be easily harvested, researchers need to be alive to the context of that data and the potential ethical implications of using it.
AUTHOR BIO: Nadia von Benzon is a lecturer in Human Geography a Lancaster University. Her current research uses Facebook as a platform for sharing birth stories and she has also published from research using Mumsnet and . She is editor of the textbook Creative Methods for Human Geographers and is a member of the editorial board of the SRA blog.