Not anonymous enough? Research data and issues of anonymity.

by Carol Robinson, doctoral researcher, University of York.


Recently, I settled down to enjoy an article by one of my favourite academic writers. It was everything I’d hoped it would be: well written, thought provoking and interesting. It took a new approach to its subject and had a campaigning edge that I sympathised with.  And then, towards the end of it, I realised that I knew one of the people who had participated in the study being reported. Not that I knew them in terms of recognising a type, but that I actually knew them. My first response was one of disappointment. I want my academic heroes to be flawless. My next thought was along the lines of ‘will anyone else know them?’ followed quickly by the question ‘does it matter?’

A quick search on-line resulted in a Wikipedia page that confirmed other people would be able to identify the participant if they wished. The academic had not revealed their interviewee’s age or location, but from the context it was clear that they were referring to a member of a small group and once more specific information was given, anyone with a curious mind and an internet connection could produce a name. From my knowledge of the individual, further details in the article then confirmed what I had found.  Anyone else would be able to identify them, even if they lacked my certainty.

So, does it matter? The article probably won’t be widely read, even in academia, and it’s therefore doubtful that anyone else will do the searching to put a name to this participant. It’s possible the participant wouldn’t mind if they were named, although the author gave no indication that they’d consented to this. The encounter that was described didn’t include anything particularly controversial or personally revealing. If they read it, the person might not like some of the ways they were portrayed but there was no obvious information that could be used against them. But shouldn’t the participant have been assured of anonymity regardless?

Anonymity is one of the things I have to think about in my own research, which is around deaths in prison, two subjects with particular sensitivities. It is one of the hallmarks of ethical conduct, together with confidentiality and informed consent, necessary not least because twentieth century history has too many examples of exploitation and damage occurring in the name of ‘research’. Anonymization arguably has a value in its own right.  Attempting anonymization, even if we secretly admit we may fail, is a way of preserving the idea of academic integrity, of seeking to avoid the exploitation of other people’s generosity that would taint our work. It is evidence of academic vigour. This links back to my initial disappointment that an experienced academic had made a mistake. If the anonymization was ineffectual, were there other aspects of this article that were in some way dubious?

Demonstrating that we have followed the conventions of academic research, whether by correctly referencing our sources or by using recognised methodologies, is part of staking our claim to be academics. It shows a respect for the traditions of our particular discipline, and in the case of techniques such as anonymization, establishes our research as ethically valid. And if ethical validity is lost, it is arguable that other forms of credibility are lost too.

Research ethics committees usually insist on anonymity and confidentiality for people participating in any research, especially vulnerable participants, as a way of protecting them. It is assumed that some harm or loss may befall an individual if their identity is known, if the stories and experiences they share and which become the researcher’s data are in some way linked backed to them as a person living in the real world, beyond the study report or academic article. Sometimes, as in my own research, this is associated with taboo subjects or criminal activity, where there may be very real consequences if anonymity is not maintained.

In seeking ethical approval for research involving prisoners, deemed to be vulnerable because of their incarcerated status, I am encouraged to think through how I will record and store my data in a way that protects their identity. The specific threat is rarely stated. Although it may be poor practice, is failing to anonymise a person really putting them at risk of harm? In many cases, there is perhaps no direct link between a possible failure to anonymise effectively and a harmful consequence for the participant; the information revealed has to have the potential to be used in a way that would confer harm. However, there is often a simple presumption that all people participating in research should be protected, which ignores the question of whether harm is likely to follow from identification.

In all aspects of our lives, most of us share personal information continually.  We willingly offer up personal information all the time, giving our names, addresses and even bank account details to near strangers, trusting without evidence that they will be used for the purpose we intend. We share our views in conversations that can be overhead by others and via on-line discussions with unknown interlocutors. We post pictures on social media, link them to others without their consent, and live surrounded by cameras. Why do we persist in thinking we can anonymise research participants?

Researchers may use pseudonyms, but often a participant’s gender, age, nationality, race or class are pertinent to the research and so cannot be hidden. We can limit access to some findings, but that poses its own ethical dilemmas. And when the research needs to focus on participants from a small group, as in the case of the article I was reading, anonymization becomes so much harder to achieve.

I have experienced this in my own research. Last year, I interviewed uniformed prison staff with experience of working with terminally ill prisoners, in a prison where there were few female officers. The interviews gave really useful insights into the work prison officers perform with dying prisoners but I was painfully aware that the female interviewees may be identifiable by other staff in the prison, despite my best efforts at anonymization, simply because they belonged to such a small group. Even with a wider pool of participants, in a tight-knit world such as a prison anonymization is hard to maintain. Surely we should not abandon useful research because it involves a small group or close-knit communities?

Indeed, should we even try to anonymise our research participants? Most of the time I would say yes, but there are times when far from protecting our participants, doing so actually risks inflicting a harm.  As researchers, we promise anonymity to ethics committee on behalf of other people, who may not wish for it. Very often, participants may have offered to help the researcher because they too care about the issue that is driving the research and want to have an impact on the situation. They may want to have their voices heard, and by extension, themselves credited. When we anonymise them, we keep their voices, but hide their faces. For vulnerable participants in particular, this is potentially a misuse of power. It is a way for the researcher to exert their positional power and claim control. Nicely anonymised, our participants may not even be able to spot themselves in our final reports and presentations. They can’t see how they are represented, and so they can’t hold us to account. There are ways round this, involving them in the production of the final report, but in my discipline at least, few researchers seem to opt for these approaches.

Lastly, I found myself thinking ‘what does one do if one spots that an academic has not sufficiently anonymised their data?’. It is not easy to be certain what responsibility we have when we spot something problematic with someone else’s work. In the case of the article I read, the peer reviewers had been content with the text, the editorial board satisfied and the article is now published. The damage, if there were any, is done and in an age of on-line journal access, probably un-doable.

I asked colleagues, and was struck by two responses in particular, widely divergent but both from science faculties. One, coming from a discipline where the professional accountability of practitioners is paramount, felt strongly that I should contact either the author directly to alert them to the problem, or the journal anonymously to suggest they review their procedures. From another department, a colleague suggested I keep quiet, and not draw attention to the problem or myself. For them, raising the matter with the author would only make things worse. Each response of course reflected the culture and values of the particular academic disciple. In some academic disciplines, where the use of human participants is rare, the question of the quality of participant anonymization may rarely come up. But for many disciplines, including my own, where the involvement of human participants is so often essential to a research project, this is an issue that can occur at any time. Do we as academics have a collective responsibility to revisit anonymization?



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s