Use social media evidence ethically

The purpose of research is to make a convincing argument based on the evidence you collected. As your geography research is science (a social science), ideally it is reproducible, meaning given the same inputs, another researcher could use what you have written to repeat your study and come to the same conclusion. For a variety of ethical reasons (confidentiality, security, etc) you sometimes need to obscure some inputs into your research. However, when doing so you also need to ensure you demonstrate the evidence drawn from these observations to build your case and ensure the project is reproducible. The ethical research guidelines around protecting human subjects are often drawn from analog methods like participant observation or interviews but don’t always apply digital to platforms where the individual posting is trying to reach the broadest audience possible.

For more on social media ethics check out: 
Townsend and Wallace (2016) Social Media Research: A Guide to Ethics 
Zook M, Barocas S, boyd d, Crawford K, Keller E, Gangadharan SP, et al. (2017) Ten simple rules for responsible big data research

Ethics is a broad category that goes beyond just the impact to human subjects. When working with social media data other ethical and legal frameworks are necessary to consider, such as copyright, trademark, attribution, and platform conditions.

Step #1 – read the terms and conditions of the platform

Every platform has terms and conditions or community guidelines, and breaking these violates the basics of ethical research.

For example, if you wanted to study Ukrainian soldiers on Instagram, you can not create a profile pretending to be a Ukrainian soldier, because Instagram’s terms and conditions prohibit impersonating others or providing inaccurate information: “you may not impersonate someone or something you aren’t, and you can’t create an account for someone else unless you have their express permission.”

However…. This doesn’t stop you from studying a particular group, this just says “don’t catfish them.” You could still identify the accounts of soldiers (particularly if they are public), observe the photos, observe the comments, etc. You could even use the photos for your research observations if they are public and you cite them appropriately.

Step #2 – Think about the intention of those being observed

This post is not just about ethics–it’s about using evidence effectively. So the next step is to think about the intention of the person posting the material you want to research. With Twitter, Instagram, Reddit and Tiktok, users are often publishing material and trying to reach the broadest audience possible. Influencers intend their post to be public and to be seen by strangers, so they have no intention or assumption of privacy. To double check this, copy the URL for the post and be sure you can access their it from a private/incognito browser window without logging in to your account.

For example, if I wanted to study street art or graffiti on Instagram, or Reddit, any of the posts I can access from a private browser when not logged in are subject to fair use. You should treat them as if they are copywritten materials and be sure to cite them appropriately.

Written consent vs attribution

Some forms of social media–Twitter, Instagram, Reddit, and Tiktok are micro-publishing platforms. Influencers producing content on these platforms are thinking of it as published works (many have monetized their posts and the platform is a source of income). You wouldn’t contact the reporter of a newspaper article to request written consent before you quote it, and it would similarly be unnecessary to contact an influencer asking for written consent for your dissertation work.

Just as you would give attribution to the author of the newspaper quote you are using through citing it, you need to give the author credit for the social media works you are referencing. This means to cite it properly:

Instagram:
Author(s) Last name, Initials. (Year of post) 'Title of post' [Instagram]. Day Month of post. Available at: URL (Accessed Day Month Year).

Reddit:
Author(s) Last name, Initials. (Year of post) 'Thread title' [Reddit]. Day Month of post (time stamp if possible). Available at: URL (Accessed Day Month Year).

With Twitter, individuals with public profiles or those using hashtags expect to engage in a conversation with strangers or influence opinion with their tweets. Similarly you do not need to get consent, but if the text of one tweet is singled and referenced your research or demonstrate a point, then it should be cited/attributed appropriately, and you may want to obscure/retract some details (see below).

Author(s) of the tweet. (Year of post) [Twitter]. Date of posting. Available at: URL (Accessed: Date of access).

Twitter is used in different ways for research. Often Twitter research involves harvesting a very large number of public tweets to indicate sentiment or look at broad patterns in the attribute data. In this case it would be impossible to seek consent. It would also be completely unnecessary as this use is within the terms and conditions of the platform, it is even encouraged by the platform, and expected by users. In this case, you would cite the API you are using to obtain the tweets, rather than the tweets themselves. For example:

Kearney MW (2019). “rtweet: Collecting and analyzing Twitter data.” Journal of Open Source Software, 4(42), 1829. doi: 10.21105/joss.

What about privacy?

If a profile and post is public, and has remained active and online in a public form for weeks/months, it’s unlikely the original author will remove it. But, the person posting it should have the right to withdraw their post from your data collection. Keep in mind that Article 89 of the GDPR protects researchers using social media for research purposes and the UK follows these guidelines.

If you are using a platform like Facebook–where users join closed groups with gatekeepers/moderators accepting/rejecting new group members, there is an expectation of privacy. Profiles on Facebook are often private, users expect privacy, and the platform has a large number of restrictions preventing research. In this case you need to seek permission from the group admin/moderator. Once granted you need to make your intentions as a researcher clear to the group members and give group members the opportunity to opt out of being studied. If this group surrounds a sensitive/private topic, or involves children/vulnerable people then there are further concerns, and research on these groups is discouraged. However if it’s a more general purpose group (e.g. memes, community groups, recipe exchanges, etc) then it is still possible to study (with precautions in place) as long as individuals can opt out of being included in your research. With Facebook, especially in a private group, it is important to protect the identity of your subjects by obscuring their names. Please noe that often these large groups have parallel pages on other platforms or the content spills over and is replicated elsewhere. In this case, it’s often easier to study, use examples or cite the posts that are on Reddit or Twitter.

Online dating apps–Tinder, Hinge, Bumble, are not social media, and it is likely your research use of the platform is breaking some term of their user agreement. Users engaging with those apps do expect privacy and content and research requires explicit written agreement and a substantial ethics review. So, instead of studying Tinder by catfishing or examining details from unsuspecting users, try studying the Tinder community on Reddit, which is a public form where users do not expect privacy. You can cite the posts and redact any identifiable details.

Reproducing and quoting research materials in your text

Many students are overly cautious about respecting the privacy of people posting to social media.  Even even respecting privacy, it’s important to show evidence for the argument you are making. This makes the research reproducible. This may mean including some posts (with appropriate citation/attribution) in the dissertation document, and it also may mean obscuring some details (like a name or a face). If you are using (with attribution) a screenshot from a public Instagram, or the commentary attached to it, it may be necessary to black out the screen names of the people commenting on the post.

For example, Michael J. Fox frequently posts on Instagram about his struggle with Parkinsons disease. His account is public (and verified), and he has 1.2 million followers, so he has no expectation of privacy in his posts, and it would be inappropriate to request consent from him. He’s using Instagram as a publishing platform, so including his posts in the dissertation is more a question of reproducing copyrighted material than ethics. However, when people comment on his posts, at times they are commenting on their own experiences with Parkinson’s disease or the experience of their loved ones. Those comments and the usernames should be treated with sensitivity, as the person posting may not intend it to be public. Using their comments, and publishing it in your dissertation may violate their privacy. In this case, if you need to use the comment, you could black out the username or any identifiable details (places, avatars, photos etc..) from the comment.

If my research were on online communities of Parkinson’s and the image below (a screen shot of this post) was really important evidence of that community. I would need to show the language of the comments in my analysis. I would cite the post by @realmikejfox and redact (with black boxes) details that I’ve circled in red below as they can easily be tied back to the identity of the poster, and they are unnecessary. Alternatively, I would use the quote “Hi Mike, my dad had Parkinson’s for the last ten years and just passed away this summer at XX…. I feel protective of you now more than ever. You would have loved my dad…” in the text of the dissertation but redact identifiable details (circled in red below).

Fair use of materials

When you reproduce an image, a screenshot, or take a quote from social media, do be careful that you are not publishing (reproducing) more of that material than necessary in the dissertation. Most copyrighted material is available for reproduction as long as it’s “fair use” which includes educational/research purposes as well as criticism and comment. Really, what this means when applied to your dissertation, is don’t put the collection of someone else’s social media content in your appendix, even if it is the data you observed and analysed for the dissertation. Aggregates, summaries, graphs are all fine, but a excessive number of images could be copyright infringement.