One area of the research process which has struck me as being in need of refinement for studying the Web is the procedure for obtaining ethical approval. Specifically, the ethics of using a social network such as Twitter as a data source for a study.
A general ethics application procedure for a project may question whether the study involves human participants, and if so, whether it will collect “personal data” (data which could potentially identify a participant). Often, it is assumed that if there are no human participants, then personal data will not be collected. If there are humans participating, then issues revolving around sensitive data, risk of harm, deception, consent, and right to withdraw become significant. This makes sense for a psychological lab experiment, for example. For a study on the Web, this causes great confusion.
A research study on the interactions happening on Twitter is a prime example of this. Data on Twitter is publically available to anyone, meaning that it lends itself to Web Science research (unless the user has set their account to private – in which case, unless a study of someone’s own followees is occurring, their data will not be accessible to the researcher). But taking data from Twitter, rather than studying people themselves, means there are no human participants – and in some cases the ethical procedures may appear to suggest that because of this, no ethical approval is required, or only a basic assessment is necessary.
This would, however, fail to take into account the many issues that may still be present in the data available to the researcher. Even without participants as such, there is still the very big possibility that personal data could be collected, for example usernames, real names, and profile information including the user’s location and bio. Additionally there is the data included in the individual tweets themselves, which could comprise of pretty much anything imaginable, and therefore cannot be ruled out from containing personal data. The possibility that this type of research may collect data from users who may not be adults underlines how wrong it would be to reduce ethical assessments simply because there are no physical participants.
So returning to the idea of a standard ethics procedure, and assuming now that because of this personal data a more detailed assessment is required, the issues of consent and right to withdraw become apparent. But again, without participants, nobody is ever asked to participate and can therefore not give consent to be studied. Can it be assumed that by posting their data on Twitter and into the public domain, the user has consented to any research study using that data? Personally I would think that a large proportion of users wouldn’t realise that this could happen, and so while technically it may be acceptable, it doesn’t quite seem morally right. If it is the case, however, then these same users need the right and ability to withdraw from the studies. But if they don’t know exactly what studies they’re involved with (because they have never been asked for consent in the first place), then how can they withdraw? These ethical procedures just don’t seem to fit with studies on the Web, and social media in particular.
I am unsure of how these problems with the current procedures should be resolved. I’m certain, however, that there is an entire Web Science project that could be filled with researching the problem and figuring out the solution because it is such an important part of ensuring the integrity of any research study, and the Web is changing it drastically.
I'm inclined to think that, since there are (despite what the police will tell you) no laws about going out into the street and taking a photo of whatever is going on, there shouldn't be any problems with going out onto the web and saving a snapshot of what is, after all, in plain view.
ReplyDeleteClearly we need ethics to protect people where research may have an impact on them, even if that's just taking up an inappropriate amount of their time. The "danger" to the participants in a legitimate study of their tweets is negligible at most.
In many jurisdictions, there is the concept of a "reasonable expectation of privacy" which is used to determine whether someone's rights have been infringed by eg a photographer. I think we need the same concept on the web. Clearly siphoning data from someone's Facebook account using a rogue application would be unacceptable, but if something is out in public, it is fair game.
I wouldn't have a leg to stand on if I walked down the street and demanded that nobody took my photo - Analogous declarations on public websites should be similarly ignored. By all means honour a robots.txt file out of politeness, but lets not pretend that people have the right to run around screaming "hey everybody don't look at me!"