During the last couple of weeks, I’ve been involved in quite a number of exchanges on Twitter, as part of my participant observation. There have been a number of occasions when I was moved to consider the ethics of a particular situation, as indeed a researcher should always do. Developing your ethical sensibility doesn’t end the moment a Research Ethics Committee has signed off your submission. Instead it should be an ongoing critical process of reflection and renegotiation (Fileborn, 2015), a fluid dialogue interwoven with the fabric of your research endeavours (Madge, 2007). Whilst that sounds rather grand, for me it means being continually alert to the ways that you conduct exchanges and being sensitive to situations which unfold which you may have originally never have anticipated. Let’s take a look at some of the issues which arose.Read More »
During a recent interview, Joe Dale mentioned a useful new app he’d found which offered some potential in the context of professional sharing – Anchor. It’s a free (as of Jan 2017) smartphone app (Android & Apple) through which you can create a two-minute audio posting (a ‘Wave’) which others can listen to, then respond, again in audio. Joe (with Rachel Smith) had experimented with it by posting a question posed by one of the #mfltwitterati, then crowdsourcing responses from Anchor users. The final combined thread is then presented as a single, stitched audio stream, where the question and responses form a coherent whole.Read More »
If potential research participants gave their permission, what would be the implications of posting interview recordings online? That was essentially the theme of the preceding post. I wasn’t so sure of which way to jump, but the encouraging and supportive comments I received there and on Twitter prompted me to take the trickier route of writing a new ethics submission. In addition to rewriting the University pro forma submission document, I had to rewrite a couple of consent forms and their associated participant information sheets, in order to accommodate the possibility that participants might give their permission to ‘publish’ their recording. I also had to write a consent form and participant information sheet for an new, additional method I want to use. I then had to amend and extend the matrix I composed which summarises the ethical issues for each method. Finally, but perhaps most importantly, I felt it was important to attempt to justify the rather radical notion that interview recordings might be posted as podcasts. Here then is that supplement to my ethics submission:
Why am I proposing a change?
The usual position is that interview participants are afforded confidentiality and anonymity – that the data they provide will only be available to those specified, and that all features which might identify them will be removed before making the findings more public. In the interests of speed and given the small scale of my pilot study, I adopted the aforementioned approach. As I move forward into my main study, I would like to propose a different stance, building on those issues discussed in Appendix 02(?): Anonymity. This also contributes to the University’s and wider Open Access policies.
The arena from which potential participants will be drawn is highly participatory, where members generally adopt a performative approach. The norms of the space include a sense of sharing what you have and what you know; where people acknowledge and give credit to those who have supported or helped them. I’d like to suggest that this participatory space invites a more participatory research approach. As Grinyer (2002) noted, researchers have to balance the need to protect participants from harm by hiding their identity, whilst preventing loss of ownership ‘on an individual basis with each respondent.’ This is manageable, provided the sample size is small, as it will be in this study,
It’s perhaps helpful at this stage to reiterate that the topic of this research is not sensitive, participants are not vulnerable and the data they share will not be ‘sensitive personal data.’
How this differs from the interviews in the pilot study
In the pilot study, the participant was assured confidentiality, anonymity, that the transcript would be deleted at the end of the study and that the findings would not be reported (only used to inform the next stage of research).
For the main study I propose a shift in emphasis from ‘human subject to ‘authored text.’ This would be achieved by allowing interviews to contribute to the participatory agenda, by releasing the interview recordings as podcasts (streamed online audio files), if participants give their permission. Links to the audio files would be embedded in a web page associated with the research project, the interviewees would be named and their contribution credited. This represents an attempt to move beyond the notion that participants are merely sources of data to be mined. In Corden and Sainsbury’s (2006) study, participants responded positively when offered a copy of the audio recording of their interviews and were given the option to amend their responses, though few chose to exercise that control.
This is a very different approach to that found in most studies, but is not without precedent. The ‘edonis’ project, part of an EdD study by David Noble, included a series of interviews with teachers on the theme of leadership in educational technologies. The interviews from those people who gave permission were posted online. It could be argued that this proposed approach is only one step further on from conducting ‘interviews’ in visible online public spaces like blog comments, forums, and some chat rooms.
Risks and benefits
Once participants’ identities are no longer disguised, both potential risks and benefits become more significant. Table xxx summarises possible risks and benefits:
|Loss of privacy which could lead to exposure to ridicule and/or embarrassment.||Direct: Increase in participant agency, moving beyond the notion of participants merely as sources from which researchers abstract data.|
|Change in future circumstances which renders what participants originally said to be viewed in a less-positive light.||Direct: Makes provision for participants to amend or extend what they said in the original interview.|
|Indirect: Increasing the awareness and understanding of the wider community of issues associated with professional learning and social media.|
|Increased attention through increased exposure.
This could be perceived as either a risk or benefit and would depend on the participant’s preferred online behaviours.
As with conventional approaches, in order to make an informed decision, potential participants would need to be made fully aware of:
- Purpose and potential consequences of the research
- Possible benefits and harms
- The right to withdraw
- Anticipated uses of the data
- How the data will be stored and secured and preserved for the longer term.
With items 4 and 5 the circumstances will be different, depending on whether participants accede to their interview recording being released. This distinction needs to be made absolutely clear at the outset so participants are able to decide whether to be involved at all and whether they want to take that additional step.
At the start of an interview, participants who agreed to allowing the interviews to be posted would be reminded of the above once more and their verbal consent captured in the recording. In the debriefing after the interview is complete, participants will be asked whether they wish to change their minds, and reminded that should they do so subsequently, how they can make those views known.
As in the pilot study, potential participants would be provided with a participant information sheet, but one extended to include the additional considerations (see Appendix xxx). The form through which they provide their consent will also be amended to offer options for the different levels of involvement (see Appendix xxx) and whether they are prepared to allow the recording to be released under a Creative Commons license (see next section)
Given the small number of interviewees (<5), coping with different levels of involvement should be a manageable process.
Copyright and Intellectual Property
These issues will also need to be made clear to participants through the participant information sheet.
…for data collected via interviews that are recorded and/or transcribed, the researcher holds the copyright of recordings and transcripts but each speaker is an author of his or her recorded words in the interview.
Rather than seeking formal copyright release from participants, it is proposed that the interview recordings will be released with Creative Commons, Attribution – NonCommercial – ShareAlike 4.0 International licensing. Participants will be asked at the point of providing consent to state whether they agree to that release; if they don’t, then the recording would not be released. Once more, potential participants are likely to be familiar with the principles of CC licensing; many of them release their own materials under these licenses.
Eynden et al (2014) recommend the use of Open Data Commons licenses for data released through research, however this licensing system is more appropriate where data is stored in databases and the database itself need licensing separately from the content. CC licensing was chosen since the content will not be wrapped within a database; at least not one which the public will be able to manipulate (copy, remix, redistribute).
Corden, A., & Sainsbury, R. (2006). Using verbatim quotations in reporting qualitative social research: Researchers’ views University of York York, UK.
Eynden, V. v. d., Corti, L., Woollard, M., Bishop, L., & Horton, L.,. (2014). Managing and sharing research data : A guide to good practice SAGE Publications Ltd.
Grinyer, A. (2002). The anonymity of research participants: Assumptions, ethics and practicalities. Social Research Update, 36(1), 4
Padfield, T. (2010). Copyright for archivists and records managers (4th ed.). London: Facet Publishing.
I’m delighted to be able to report that my revised submission has passed the ethics review process. It’s highly unusual for interviews to be allowed to be published in this way; standard practice is to afford anonymity to interviewees. Perhaps it’s indicative of the need to make our research more open, or the more performative behaviours of potential participants … or perhaps a bit of both. Whatever the case, I’m chuffed to bits, as we’d say up here in the ‘North.’ Now all that remains is to find participants sufficiently confident and generous enough to give it a shot. Know anyone …?
Right from the outset, one of the options I’ve tried to keep in mind is that of ‘publishing’ those data that are amenable. Publishing in this sense refers to sharing interview recordings, as podcasts, back with the community. This feels like the right thing to do; when teachers experiment with new techniques that someone else showed or explained to them, they often share those insights more widely. If that is the norm, why wouldn’t my research study, conducted within this environment, be any different? Well there are a number of reasons, mostly arising as a result of a researcher’s’ ethical sensitivities and obligations towards potential participants.
The default ethical stance is to maintain participants’ anonymity and confidentiality; with an interview transcript, this isn’t too difficult. If on the other hand, the audio file of the interview is shared, the potential for the participant to be identified is so much greater, even if personal identifiers are edited out of the audio. However, it could be argued (as I began to discuss here) that in the online performative space with which participants are comfortable, anonymising what they have created actually does them a disservice. Much better to acknowledge their co-authorship and give credit where it’s due. I wonder how many researchers conducting interviews as one of their methods, discuss the issue of ownership, copyright or intellectual property with their interviewees, beyond explaining where their data will be stored and how it will be used. In fact ‘for data collected via interviews that are recorded and/or transcribed, the researcher holds the copyright of recordings and transcripts but each speaker is an author of his or her recorded words in the interview.’ So I find myself speculating what the implications and potential consequences of that are? As Van den Eynden et al (2011) explain, an author could at some time in the future assert their rights over the words they provided and you would be obliged to comply. It is possible however for the researcher to have sought ‘transfer of copyright or a licence to use the data’ ideally at the outset of the project. There are even templates available through the Data Archive to make things easier. I wonder though whether taking the route towards Creative Commons licensing might provide a route forward? Potential participants are likely to be familiar with it; many will indeed use it with their own material. But that then has me wondering whether that’s permissible under the University regulations for PhD research (which of course I could doubtless find out), but also what the implications might be if you subsequently wish to publish your research through conventional commercial channels.
My work this morning has been with the apparently less sticky technical issues – where would the audio files be stored, how would they be served/streamed etc. In the past I’ve used the free versions of various podcast services like AudioBoom, SoundCloud, Spreakr etc, but they’re of course limited in some way and would not be adequate for several hour-long podcasts. Paying for upgrades is an option, but I don’t fancy picking up the tab of tens of pounds per annum, just for this project. Online storage can be bought for a much more manageable outlay through services like Amazon S3, or perhaps more ethically(?) through Reclaim Hosting, but which of course demand a higher level of technical capability to configure, manage and maintain the site. I probably have enough background to cope with that, especially if supplemented by online tutorials … and I have been considering securing a new domain name anyway. But then what happens in the longer term? How long will I need to maintain the site and content?
I can’t help but be drawn back to ethical principles, specifically those of non-malfeasance and beneficence. Would sharing podcasts of interviews be likely to result in any harm befalling participants and are there ways in which they might benefit? Is is not easy to speculate what harms an interviewee might incur, but not dissimilar perhaps than those from potentially any online activity. In most cases (assuming the material is not inflammatory or illegal) the most harm is likely to be reputational damage from an inappropriate or ill-judged comment. It might be possible that potential future employers might be put off by opinions or ideas expressed – if as a teacher, you expressed particular pedagogical approaches you favoured and they were at odds with the views of a potential employer who heard your interview, then s/he might be less inclined to offer an interview. Again though, if you hold a particular set of values and have an online presence, it’s likely you’ll have already burnt that bridge. This can of course be flipped and work in your favour as it did for Daniel Needlestone – a benefit? For those who share widely, seek exposure and an audience, then being provided with an opportunity for that through an interview, then this might indeed be considered to be in their interests. And of course, as for many research participants, but perhaps particularly for teachers, there is the sense that their participation is contributing the pool of knowledge from which we all sip … or gulp.
I’m obliged to also ask myself why I might want to do this; what do I stand to gain? Am I being selfish and actually seeking kudos from the community? Am I attempting to follow in the spirit of making research more open and more accessible? Am I attempting to be more faithful to my participants in seeking to ensure their voice is not lost through my transcription. Is this one way in which I can be more transparent about my analysis and interpretation? Is this an additional channel through which I can make my research accessible to a wider audience? Perhaps a little of all of the above?
So which way do I go? My easy route is to stick with the ethical issues I’ve already had ratified for my pilot study and go with participant anonymity. The difficult route, for all of the aforementioned reasons, is to seek to ‘publish’ the data and therefore have to write a new ethics submission incorporating all those issues and explaining how I would address them. That might be time consuming (both in the composition and in the approval process), but is not impossible; the edonis project by David Noble has already set a precedent in fact. Which option would you choose if a) you were me, and b) you were a potential participant – what would your preference be?
Van den Eynden, V., Corti, L., & Woollard, M. (2014). Managing and sharing data.
Today I received detailed feedback from my supervisors on my draft submission for ethical approval. It’s reassuring to have some wiser and more experienced clearly isolate and highlight some of the issues that felt ill-defined or less well-articulated. Not to mention point out the inconsistencies or lack of alignment that you never spotted. The latter are relatively easily corrected; the former need a little more thought.
In particular I need to reconsider the toolkit of methods I’m proposing for my pilot study. I’m clearly asking a lot of myself and in my submission haven’t clearly expressed whether what I’m proposing is achievable or desirable. I could reduce the breadth by prioritising and rationalising my choices, or I could aim for less depth and rather than a full thematic analysis, write research and substantive memos reflecting on the methods. Another way I might explore the demands these choices will make on me is to break down the times required for each of the methods and map that out against the time I’ll have available in following months.
Another area where my submission could be strengthened is by providing a supplement in which I justify my ethical decisions by referencing the relevant literature. This could of course be literature specific to ethical issues, or more generic methodology literature in which ethical issues are referenced. Although I included with my submission the appendix I mentioned previously, it didn’t include any references to the literature which had informed my thinking. I made the choice not to include references for simple practical reasons – in an attempt get all the information in a single summarising sheet. Like my supervisors, the ethical review panel won’t have the time to pore over the posts I wrote, so I need to summarise that thinking into a more succinct form.
The final issue I need to address is one I knew might be problematic; that of preserving anonymity for the participants. If verbatim quotes are not required in any published materials, ensuring anonymity for participants should be possible. As I discussed before though, I’m concerned whether those involved would want to remain anonymous; I think I need to make a case for why citing participants might actually be more ethical, by drawing on those studies where more learned people than me have actually done just that.
Over the next few posts I’ll attempt to address these issues.
That’s the question researchers have to continually ask themselves – why is what I’ve said/written/discovered important or why does it matter? The preceding seven posts have covered various topics around the ethics of conducting research on the Internet. So what? What matters in this instance is how that will help to ensure my research study is ethically sound. In this concluding post I’ll try to frame that learning within this context.
It’s worth reiterating the context with which my study is located. A specific group of teachers engaged in a particular activity – those using Twitter for professional learning. These are well educated people engaged in an activity (of their own volition) which would not be considered ‘sensitive.’ Charged with supporting young people in using technology and social media wisely, they are more likely to be aware of the consequences of engaging in the use of social media.
The pilot methods I outlined in a previous post are restated here:
- Immersion in my Twitter stream for 24 hours (possibly over three shifts) – ‘deep hanging out’. This will provide a snapshot of activity from a self-selecting sample of the two thousand plus educators I follow.
- Closely following the twitterstream of a teacher for a limited period, chosen from those who have made claims regarding the efficacy of Twitter. This is to investigate whether focusing on an individual might yield more informative data.
- Attempting informal interviews using the commenting feature on blogs. This will be across a small number of blog posts in which the authors make claims of how useful they found Twitter for professional learning.
- Conducting a single semi-structured interview with one of the more evangelical of those educators making claims for Twitter. This should tease out areas and themes to explore in more depth.
- Seeking permission, then attempting a focus group interview within a Twitter #edchat. This may push the boundaries of what constitutes a focus group, or the depth of discussion possible in a #chat.
- Using an automated routine to collect tweets over one month which reference a particular term e.g. “professional learning.” This will access the general Twitter stream and therefore a wider sample, offering the potential for unanticipated outcomes to emerge.
- Attempting to open dialogue within Twitter (or elsewhere) with anyone who makes claims about Twitter in relation to their professional learning. A ‘naive’ stance will be taken whilst attempting to draw out further information.
- Small-scale social network analysis of a topic or hashtag to explore the interconnections which are forming. The focus here is not on the content of the tweets, nor the people which are connected, but the ways they are connected with each other and the information flows between them.
In the following table, I highlight how the significant themes discussed in this ethics strand of posts apply to my proposed methods:
These are of course only my interpretations, based on being a user of Twitter for professional learning for the last seven years. It would be imprudent however to assume I am at liberty to speak for all teachers on Twitter. Ideally I should seek to verify the alignment of my perceptions with those of the potential participants. I could survey or interview people, as Beninger et al (2014) and Hudson and Bruckman (2010) did. But I suspect the outcome would be far from definitive, and I would find people expressing the same range of views from ‘this is an open platform and people should know what they’re doing’ to ‘I’m OK with people using my information so long as they ask first.’ The difficulty then is in being sensitive to and addressing the wishes of all participants. Is this even possible? It’s important therefore to behave in a way that responds to context; if a tweet or the content of a blog leans more towards personal reflection, than open debate, I would be less inclined to intrude. As Roberts et al (2004) advise – “expectation of privacy overrides the distinction between public and private spaces.” Better then to consider the micro-context very carefully.
Some factors in the summary (like ‘degree of interaction’) are less subjective, whilst other issues are open to interpretation. Views about what is public and what private often differ in degree, as indeed does the need to seek consent. Where possible, it would seem sensible to be guided by precedent and what has been deemed acceptable practice by researchers who have gone before.
My greatest dilemma is regarding anonymity. Based on experience and personal preference, I tend to concur with Roberts et al (2004) & Sixsmith & Murray (2001), and would prefer to seek to empower participants by offering them the choice of having their authorship recognised and being credited in published works. What I feel I’ve done, as can be seen in the table, is opt for the safer and simpler option of anonymising the data, rather than acknowledging the contributions of participants. Wrong decision? What would you prefer?
Despite all the dilemmas, dichotomies and disagreements, the touchstone to which I’ll always return is to ensure minimal impact/harm for participants and to maximise beneficence.
BENINGER, Kelsey, et al. (2014). Research using social media; users’ views. [online]. NatCen social research, .
HUDSON, James M. and BRUCKMAN, Amy (2004). “Go away”: participant objections to being studied and the ethics of chatroom research. [online]. The information society, 20 (2), 127-139.
ROBERTS, Lynne, SMITH, Leigh and POLLOCK, Clare (2004). Conducting ethical research online: Respect for individuals, identities and the ownership of words. In: BUCHANAN, Elizabeth A. (ed.). Readings in virtual research ethics: Issues and controversies. IGI Global, 156-173.
SIXSMITH, J. and MURRAY, C. D. (2001). Ethical issues in the documentary data analysis of Internet posts and archives. [online]. Qualitative health research, 11 (3), 423.
With one of the fundamental principles of research being to minimise harm, perhaps we should be aiming to tread as lightly as possible in the field? Minimise disturbance to participants? Could it be argued therefore that we should aim to be unobtrusive as possible?
If it is important to gather data based on naturalistic behaviour, as untainted as possible by researcher presence, then being unobtrusive becomes a primary goal. The intention is to avoid observer effects by acting as an ‘overhearer’ (D’Arcy, 2012), rather than a participant who may influence outcomes. However it is a fine line between being unobtrusive and covert; between minimising influence and hiding from view. As Hine (2011) observed:
although we might be able to easily access data using unobtrusive methods, this does not make this ‘ethically available’
The ethical issues do not prevent us from using unobtrusive methods, so much as to remind us of our obligations to participants …and authors.
Unobtrusive research methods predate the Internet, but they have certainly become easier through online channels. Some techniques like content or sentiment analysis of large data sets which provide summaries, rather than specific, identifiable details come with less ethical baggage. Entering a chatroom or monitoring the twitterstream without announcing your presence is also unobtrusive and some might say, crosses an ethical line. This is now covert behaviour, albeit in a ‘public’ place. Lurking, as it is known, is a legitimate online activity however and is common when people enter a new environment for the first time; it allows them to become familiar with norms and conventions.
An ethical case can be made for unobtrusive, even covert research on the grounds of the reduced impact it has on participants. They’re not required to give up any time, to fit appointments into their schedule or to worry whether what they’ve said is helpful/useful to the researcher. Once more we’re confronted with shades of grey rather than definitive answers, however Whitty (2004) draws a line in the sand for us:
“While it might be unclear as to how ethical it is for lurkers to collect data on the Internet, there is less doubt as to whether it is acceptable to deceive others online in order to conduct social research,…”
Being unobtrusive tips over into deception if researchers deliberately conceal their purpose, do not fully disclose relevant information to participants, or provide false information. (Madge, 2007; Frankel and Siang, 1999). Whether lurking constitutes deception is open for debate. Perhaps we need to return to some of the issues discussed in earlier posts, like privacy.
There are circumstances however where deception might pass scrutiny from an ethical review panel. Those situations where the research could not otherwise be undertaken for example, but only if participants (and researcher) are protected from harm and they are debriefed after the research.
In some online arenas, deception (withholding information, pretending to be someone other) might be the norm; MMORPGs and virtual worlds for example, where a player might take on the role of a character or choose an alternate identity. Without good reason, a researcher should avoid such behaviour, instead opting to find the means to disclose to fellow participants that you are conducting research as (Eynon, 2009; Krotowski, 2010).
How you disclose your status as a researcher to the group with whom you are participating online will depend on the conventions in that space. Providing details in your profile together with a link to an institutional website ‘can increase the credibility of the researcher’s claimed identity and shows respect and courtesy to members of the newsgroup.’ (Madge, 2007). Although some groups are openly hostile to the presence of researchers (Hudson and Bruckman, 2004), a respectful approach and involvement might not only grant access, but also pay dividends:
“Such efforts to establish cultural membership and disclose research aims were foundational to creating relations of caring and trust with group participants.” (Walstrom, 2004)
Porr & Ployhart (2004) consider this rendered even more powerfully through the disclosure-reciprocity effect – “we reveal more to those who have been open to us.” By being completely open and transparent with our participants, it is likely that they will reciprocate.
Since they were conducting observation-only research in a public space and did not need to interact with the participants (using interviews, surveys or experiment), Coughlan and Perryman (2015) felt justified in not disclosing their status. However they also took great care in anonymising the data they gathered, even going so far as to break Facebook’s terms of service by altering the screenshots they captured; a practice regularly undertaken by many reputable institutions and other academics.
Perhaps it is our ethical sensitivity that makes us feel uneasy with unobtrusive, covert or even deceptive research. That is right and proper, but we should also take care that we do not sacrifice a potential contribution to knowledge by playing it too safe. Under the right circumstances, we are granted ethical latitude:
“If research requires any kind of deception, then only by the clear demonstration of the benefits of the research can it be justified.” SRA, 2003
“Education researchers do not use deceptive techniques unless they have determined that their use poses no more than minimal risk to research participants; that their use is justified by the study’s prospective scientific, scholarly, educational, or applied value; and that equally effective alternative procedures that do not use deception are not feasible.” AERA, 2011
“Researchers must therefore avoid deception or subterfuge unless their research design specifically requires it to ensure that the appropriate data is collected or that the welfare of the researchers is not put in jeopardy.” BERA, 2011
AERA Code of Ethics: American Educational Research Association Approved by the AERA Council (2011). [online]. Educational researcher, 40 (3), 145-156.
BERA Ethical guidelines for educational research. (2011). London, BERA.
COUGHLAN, Tony and PERRYMAN, Leigh-Anne (2015). A Murky Business: Navigating the Ethics of Educational Research in Facebook Groups. [online]. European journal of open, distance and e-learning, , 146-169.
D’ARCY, Alexandra and YOUNG, Taylor Marie (2012). Ethics and social media: Implications for sociolinguistics in the networked public1. [online]. Journal of sociolinguistics, 16 (4), 532-546.
EYNON, Rebecca, SCHROEDER, Ralph and FRY, Jenny (2009). New techniques in online research: Challenges for research ethics. [online]. Twenty-first century society, 4 (2), 187-199.
FRANKEL, Mark S. and SIANG, Sanyin (1999). Ethical and legal aspects of human subjects research on the Internet. Published by AAAS online, .
HINE, Christine (2011). Internet research and unobtrusive methods. [online]. Social research update, (61), 1.
HUDSON, James M. and BRUCKMAN, Amy (2004). “Go away”: participant objections to being studied and the ethics of chatroom research. [online]. The information society, 20 (2), 127-139.
Social Research Association (2003). Ethical guidelines. [online]
KROTOSKI, Aleks (2010). Introduction to the special issue: Research ethics in online communities. [online]. International journal of internet research ethics, 3 (1), 1-5.
MADGE, Clare (2007). Developing a geographers’ agenda for online research ethics. Progress in human geography, 31 (5), 654-674.
Social Research Association (2003). Ethical guidelines. [online].
PORR, W. Benjamin and PLOYHART, Robert E. (2004). Organizational Research Over the Internet: Ethical Challenges and Opportunities. In: BUCHANAN, Elizabeth A. (ed.). Readings in Virtual Research Ethics: Issues and Controversies. Hershey, USA: Idea Group Inc, 130-155.
WALSTROM, Mary K. (2004). Ethics and Engagement in Communication Scholarship: Analyzing Public, Online Support Groups as Researcher/Participant-Experiencer. In: BUCHANAN, Elizabeth (ed.). Readings in Virtual Research Ethics: Issues and Controversies. Hershey, USA: Idea Group Inc, 174-202.
WHITTY, Monica (2004). Peering into Online Bedroom Windows: Considering the Ethical Implications of Investigating Internet Relationships and Sexuality. In: BUCHANAN, Elizabeth (ed.). Readings in Virtual Research Ethics: Issues and Controversies. Hershey, USA: Idea Group Inc, 203-218.
Research participants have a right to anonymity, confidentiality and being able to provide informed consent. This would appear to be ethically unproblematic, setting aside the issues discussed in the previous post for a moment. To incorporate data drawn from a blog into a report or article would therefore require anonymising any details which might identify the author. But here’s the thing; what if they don’t want to be anonymous? What if their livelihood or reputation is based on the success of their posts, as distinguished by how widely they’re shared and reposted? To unpick this tension, we need consider two different perspectives: the online world can be viewed as a cultural realm wherein people perform and interact, or that what we see online consists instead or authored texts.
A social world
The default (and safe) option is to assume that the activity we view online is intrinsically linked with the people who produced it and consequently we should adopt a ‘human subject’ ethical stance. Originating in medical research, the foundations can be traced back to the Belmont Report and further. Walther (2002) provides the specifics:
Human subjects research is that in which there is any intervention or interaction with another person for the purpose of gathering information, or in which information is recorded by the researcher in such a way that a person can be identified directly or indirectly with it.
The question then is whether is a human subject approach is still valid where the research process neither involves biomedical procedures, nor interacting with people? Markham and Buchanan (2012) in the guidance from the Association of Internet Researchers caution that:
Because all digital information at some point involves individual persons, consideration of principles related to research on human subjects may be necessary even if it is not immediately apparent how and where persons are involved in the research data.
but acknowledge that the Internet may call into question the notion of personhood, asking whether the digital traces we produce online can be considered an ‘extension of self.’ Although the data a researcher might be linked with a person, if we interact with that text, are we still interacting with a human subject?
A textual domain
Some argue for a different sensibility where the data we access on the Internet ought to be considered as ‘authored texts’ (Basset and O’Riordan, 2002; Walther, 2002). Whether a web page, blog post, forum thread or tweet, these textual artefacts have been inscribed onto the world wide web and persist over time. More akin to books, newspaper articles and company reports, they are divorced from the originator and require us to draw from a different set of ethical principles. We think instead of issues around ownership, copyright and attribution, where the text may have been authored by someone, but does not constitute a part of them; the two need not be conflated.
This perspective does not mean we are at liberty to drop the fundamental ethical principles of autonomy and beneficence; they simply shift to ensure them in a different way. Rather than provide anonymity and confidentiality, we acknowledge authorship and attribute it in the same way we would when quoting more traditional works. This assumes of course that we remain sensitive to issues of privacy and sensitivity of subject, though Wilkinson & Thelwall (2011) point out we need not feel obliged to be drawn too far back towards the human subject stance:
Although web texts can be treated as documentary research sources or cultural artefacts, they deserve special consideration because they are less obviously public than a published book and often contain personal information. This is an issue of privacy rather than consent, however.
I’m not sure I’d agree that online texts are less ‘public’ than a published book, when (assuming the text was not written to a private online space) visibility of that text may be no more than a click or two away. A book has to be first bought or borrowed through a conscious act; online texts can be encountered by chance, at any time, by anyone.
We must remember that if someone is publishing in a performative space, where the notion of public audience (and possibly recognition) is the norm, then by anonymising rather than attributing their contribution, we may be doing them a disservice. A harm, rather than a benefit, or as Basset and O’Riordan (2002) put it, a diminution of ‘the cultural capital of those engaging in cultural production through Internet technologies.’
Shades of grey
Simple, accessible tools have become available which enable researchers to hoover up large corpora of data from social media. Big data research brings with it a set of ethical concerns of its own, but within the context of this post, we should ask to what extent the people who generated the texts within these corpora are still present. This becomes even more pertinent if the findings of the research are published in aggregate form, rather than being attributed to (anonymised) individuals. This introduces the ‘distance principle’ (Lomborg, 2013); the degree of conceptual and experiential separation between the researcher and the participant/author, or between the research object and the person who produced it. The closer the distance, the more vividly the ‘human’ comes into view and therefore the more likely it ought to be classified as human subject research. Greater distance can be achieved when the participant identity is less distinct, or when the interaction between researcher and participant is lower.
In the following graphic, on a continuum between human subject and authored text, I’ve added a distance dimension. The examples are notional online research situations, but their precise position will depend on context.
The examples coloured red involved close interaction between researcher and participant and high proximity between the object being analysed (interview ‘transcript’) and the producer of the object. The green examples show much greater separation or distance. Red examples will undoubtedly require informed consent; green perhaps not. Would you argue for any of the examples being moved? Can you think of other examples which could be included?
BASSETT, Elizabeth H. and O’RIORDAN, Kate (2002). Ethics of Internet research: Contesting the human subjects research model. Ethics and information technology, 4 (3), 233-247.
LOMBORG, Stine (2013). Personal internet archives and ethics. Research ethics, 9 (1), 20-31.
MARKHAM, Annette and BUCHANAN, Elizabeth (2012). Ethical Decision-Making and Internet Research: Version 2.0. [online]. Association of Internet Researchers.
WALTHER, Joseph B. (2002). Research ethics in Internet-enabled research: Human subjects issues and methodological myopia. [online]. Ethics and information technology, 4 (3), 205-216.
WILKINSON, David and THELWALL, Mike (2011). Researching personal information on the public web methods and ethics. Social science computer review, 29 (4), 387-401.
When conducting research which involves people, it is important that they are able participate of their own free will. To be able to decide whether to participate, they will need to know in advance what the study involves, what will be expected of them and what any potential risks or benefits might be. They should be able to make their decision freely and without coercion and be able to change their mind should they feel the need. These are the foundations upon which the principle of informed consent is based:
- autonomy – participants should have sufficient time and (appropriate) information to make a choice.
- beneficence – participants should be able to experience any benefits and be kept free from harm.
- justice – ensuring any opportunities, benefits and risks are distributed fairly and equitably.
As researchers, we are guided by the principles produced by our institutions and professional associations mentioned in an earlier post. We also have certain legal obligations enshrined in the Human Rights Act 1998 and the Data Protection Act 1998.
Gaining consent should not be considered a single, one-off event conducted at the start of (or possibly later in) a study. Rather it should be an ongoing process of mutual exchange, where the researcher regularly reminds the participant that they are involved in a study and the participant is free to ask the researcher questions and indeed is free to withdraw at any point. Practically, this often involves the researcher producing a Participant Information Sheet (PIS) which provides the details of the research in clear, unambiguous language. (An exemplar from my institution can be found here) The prospective participant is given this and provided with adequate time to make a choice about whether to proceed. They are then asked to sign a consent form, which asks them to confirm they understood the information they’ve read, reminds them of what they are providing their consent for and reminds them they can withdraw at any point. Would that it were only that simple!
There are numerous circumstances in which obtaining consent in the way described above are problematic. In some cases, participants may be poorly placed to provide their consent because they have reduced faculties (those with Alzheimer’s for example), lack maturity (the young) or may be indisposed (in a coma). Some studies may not be on specific individuals, but be on communities or organisations; would this require consent from all members, or just a ‘gatekeeper’ or advocate? Research might be conducted at particular events or in locations where people are involved en masse and seeking consent from all would be impractical. Some situations involve a participant who has provided consent, but their usual activities continually bring them into contact with other people on a transient basis; a market stall holder, a librarian or police officer for example. How can the consent of those additional people be obtained without having an impact on the natural activity of the participant, or perhaps compromising the research data? Which all raises the question whether informed consent is always required, or are there circumstances under which obtaining it might not be necessary and if so, what are the factors which need to be taken into account?
It is not always possible or practical to obtain consent, especially in observational studies in unbounded areas like a shopping mall. Using posters or notices, it may be possible to forewarn people that research is taking place, but obtaining consent would be challenging. There is also the issue that by providing notice that research is being undertaken, naturalistic observation becomes impossible because people change their behaviour. Research conducted in this way is said to be ‘covert;’ the participants are unaware that they are being observed. In some circumstances, it may actually be harmful to obtain the consent of individuals, such as in obtaining a signature from someone involved in illicit or illegal activities. This is when ethical issues and the needs of scientific discovery need to be very carefully considered and balanced. Hudson and Bruckman (2004) looked to the U.S.6 regulations governing academic research, which advise that research can only be conducted without informed consent if:
- the research is not human subjects research
- the research is exempt from IRB (insitutional review board) oversight, or
- when an IRB issues a formal waiver of consent.
Like Hudson and Bruckman, I too have concerns about the first two. I’ll discuss human subject research in a future post, but side-stepping ethical review is neither an option, nor something I’d even consider. Here in the UK, we don’t, to my knowledge, have the notion of a ‘formal waiver of consent,’ but our ethics panels are at liberty to approve research where consent will not be sought, provided an adequate case can be made. What indicators can we use then to suggest when informed consent may not need to be sought. The American Sociological Association Code of Ethics (1999) provides guidance on when consent should be sought:
when behavior of research participants occurs in a private context where an individual can reasonably expect that no observation or reporting is taking place.
but qualifies this by suggesting that consent may not be needed where ‘naturalistic observation in public places’ is conducted, provided participants are likely to suffer minimal risk and where the research could not otherwise be conducted practicably. This of course means we need to refer back to the previous discussion on how we differentiate between public and private. McKee and Porter (2009, 11) proposed that for a given set of circumstances, the degree of privacy and level of sensitivity should be established, then a point representing that located on the Sveningsson diagram. By adding a diagonal, it is proposed that anything above and to the right would need informed consent.
Of course different researchers or participants might locate the same situation at different points on the chart, but the intention is not for the chart to provide definitive answers, but act as a heuristic to facilitate discussions and negotiations. Ethical issues often manifest themselves as gray, rather than black or white. By also considering the degree to which the researcher would need to interact with participants, McKee and Porter offer a more sophisticated graphic which can further assist decision making:
Of course these determinations make the assumption that it is the researcher(s) making the decision, but perhaps there is space too for including members of the group under study in discussions, thereby shifting the power relationship back towards participants somewhat?
Having considered at length whether consent is necessary, the actual act of obtaining it is not without its challenges. In an online setting, researcher and participant are likely to be geographically separate. Rather than require a physical signature on a printed document, the Internet offers a more direct, immediate option. On many websites, it is standard practice to require new subscribers to ‘sign’ their acknowledgement of terms and conditions by ticking a box; the same technique could be applied for giving consent. Of course it is also true that the majority of people don’t read terms and conditions before ticking the box; this is not something which would be acceptable in a research context, since participants could hardly be considered ‘informed’ if they are not aware of what they are agreeing to. This can be ameliorated to some extent by ensuring the PIS is concise and doesn’t require too much reading, or by breaking down the terms on the consent form into separate steps, each requiring a tick to acknowledge acceptance. Even then, as Flick (2015) observed, it is difficult for a ‘remote’ researcher to be able to tell whether a participant gave considered consent in the same way that a researcher in a face-to-face setting might be able to. The option to quickly seek clarification about an issue they’re unsure about, is somewhat more difficult for an online participant. Nor does the online researcher receive any visual clues that a participant might be hesitant about something, in the same way that a face-to-face researcher might. Flick also argues that authenticity is compromised online; how can a researcher know that the person giving consent is who they say they are; though of course the same could be equally true in face-to-face setting, or indeed in traditional posted survey research. Once more we are faced with balancing the need to undertake the research, against the shortcomings of the medium, as far as informed consent is concerned.
Having negotiated these troubled waters and finally obtained consent from participants, the matter is not yet closed. Particularly in online settings, the composition of those ‘present’ in a social space tends to fluctuate; people tend to come and go as time passes. If consent had been sought and obtained at the outset, that might no longer apply to all present. Once more we find ourselves balancing the need to obtain consent against becoming intrusive with constant reminders and perhaps damaging the nature of the group, or the interactions therein. There is another element we must consider when the research approaches conclusion; whether someone who gave consent to participate also meant consent for their contributions to be shared when the research is published or reported. This need not be too troublesome provided this was included in the PIS at the outset and that participants understood that information they provided might be shared, and in what way it would be shared. Providing anonymity tends to be the default means to smooth this path, but that too can be problematic, as I’ll discuss in a future post. It is also becoming increasingly the norm for research data to be shared more openly, together with the report, thesis or article. What guarantees might a participant have that the terms under which they gave their consent will be adhered to by some future researcher, conducting a different study and using the data in a completely different way?
Wherever possible, involving participants in discussions about what they feel constitutes appropriate and acceptable behaviour for a researcher, is a desirable aspiration. Whether consent is required can form part of that discussion. Rosenberg (2010) identified three distinct sets of views amongst participants. Some would not wish to participate in any form of research without having first provided their consent. Others feel that consent is not required, provided the researcher does not become involved or participate within the activities. The final group would be comfortable with the researcher interacting and gathering data without consent, provided they don’t deceive participants to gain information. That there are different views will come as no surprise. If all those views were to be accommodated, there would be some research or some methods which could not be undertaken. The debate then becomes one of the worth and value of the research for the greater good. Do the benefits outweigh the harms?
BUCHANAN, Elizabeth A. and ZIMMER, Michael (2015). Internet Research Ethics. [online]. In: Edward N. Zalta (ed.). The Stanford Encyclopedia of Philosophy. Spring 2015 ed., .
FLICK, Catherine (2016). Informed consent and the Facebook emotional manipulation study. [online]. Research ethics, 12 (1), 14-28.
HUDSON, James M. and BRUCKMAN, Amy (2004). “Go away”: participant objections to being studied and the ethics of chatroom research. [online]. The information society, 20 (2), 127-139.
MCKEE, Heidi A. and PORTER, James E. (2009). Playing a good game: Ethical issues in researching MMOGs and virtual worlds. [online]. International journal of internet research ethics, 2 (1), , 5-37.
ROSENBERG, A. (2010). Virtual world research ethics and the private/public distinction. [online]. International journal of internet research ethics, 3 (1), 23-37.
When thinking about the ethics of a situation, it won’t be not long before the online researcher will be troubled by a ‘private versus public’ debate. The reason why this becomes such a pressing issue is that it will serve as a guide to the way the researcher should behave in particular circumstances. This is particularly significant when determining whether consent should be sought and/or whether it is necessary to maintain anonymity for the participants.
What do we mean by ‘private’ … or ‘public?’
Privacy may be defined in law, but for most of us, it is determined by social norms. Different people have very different conceptions of what it means to them and those conceptions may vary depending on circumstances. For many years private and public have been viewed as a dichotomy, with domestic life and the home representing the former. Although personal too can sometimes mean private, this is not always the case, as I shall discuss later.
Let us first consider to what we’re referring when we wish to distinguish between private and public. There are two main perspectives here: firstly a spatial one in which some regions or locations are viewed as more or less, private or public. A town square might be considered a public space and one’s car a private one. Secondly, information can be viewed in the same way; some is public, like a train timetable, and some private – a message to a lover. But in this last example, we begin to see a further possibility – conversation, or the exchange of information. Here then we see that the intended audience for the information we wish to share matters too. As Rosenberg (2010, 34) suggests
It is not enough to consider whether a space is public, who the intended audience is or whether some information is personal. All three must be considered.
What we’ve seen in recent years however, is a shift in emphasis from the times when privacy could be taken for granted, because sharing information was more difficult. Social media has changed that to ‘public by default; private through effort’ boyd & Marwick (2011, 12).
The suggestion so far is of a public – private dichotomy, but social norms and the way we view these two poles are shifting. Perhaps with the advent of technologies like email, the mobile phone and text messaging, work-life (public) is beginning to intrude in home-life (private) and similarly, more people access their home or personal lives whilst at work. The private – public boundary is increasingly blurring. Many authors caution against viewing public-private as a dichotomy. A particular space can be public and private at different times, in different contexts and at different scales. The public-private status of a school for example can be quite fluid in a way that, say a library is not. A maintained (state-funded) school is considered a public resource1 available for the benefit of its local population. However the entrance gate, often secured during the school day, suggests a more private place. Access is only allowed to specific people: those who work there, the student population, parents and other legitimate visitors. That of course changes on ‘open’ days, for sports events or theatrical productions; or at different times of the day, week or school year. Context and temporality become important. At a smaller scale, the school can be broken down into smaller spaces, some of which are more or less public. The school hall may be accessible to all, but more or less so at different times of the day. Senior staff may have offices which are mainly private, at least for certain sectors of the school population. Perhaps there are some parallels here with social media where some areas are more public and open than others, but individuals can choose precisely how public their space or the information they wish to share is by using privacy settings? As Ford (2011) observed:
Between the purely private and the purely public there exist an infinite and an infinitely variable number of configurations that fall somewhere between the traditional categories of ‘private’, things that happen or are said behind physical or virtual closed doors, and ‘public’, those interactions and events that take place within full view of an unknown audience.
Sveningsson Elm (2009, 136) offers a system in which the private – public continuum can be subdivided into four categories:
- public – open and accessible to everyone
- semi-public – open and accessible, but only to those who have obtained membership
- semi-private – only available to those who have secured membership, the criteria for which are conditional
- private – hidden or unavailable to anyone but the owner and invited guests
What categorisations help researchers do is to establish when and where it might not be necessary to obtain informed consent. Sveningsson Elm for example, suggests that public and semi-public spaces might not require consent to be obtained. Perhaps, but there are still further factors we should take into account before making that decision. What might be considered private to someone in the UK, might be perceived as more public to someone in Trinidad. A teen’s view of their bedroom as a private space may be viewed differently by their younger sister or father.
The ways in which individuals perceive privacy is rather fluid, varying from person to person, across different contexts and changing at different times. As a citizenry, many people trade-off increasing ubiquity of surveillance cameras for better protection and detection of crime; they accept cookies which track our web browsing for a more personalised experience; and allow in-app monitoring to benefit from location-based services like Tinder. Yet if a fellow passenger on the bus was to overtly read our activity on our smartphone, we would most likely object, seeing that as an invasion of personal space or privacy. However, strangely, as Ford (2011, 556) noted:
protecting all tidbits of personal information is simply not as important to some as it used to be. In fact, some individuals turned the idea of surveillance on its head, broadcasting their daily lives to the entire World Wide Web.
Personal is private?
Although our initial reaction is to conflate private with personal, this is not always valid. People often publicly share very personal, sometimes intimate things through social media, in a way they wouldn’t do offline. This is often completely intentional and with an awareness of potential outcomes, but as boyd and Marwick (2011) noted of people’s behaviour:
practices in networked publics are shaped by their interpretation of the social situation, their attitudes towards privacy and publicity, and their ability to navigate the technological and social environment.
However, the alternate argument can also be posed, where people, either through lack of experience or maturity, may not conceive of the full extent of their audience. They may be writing for a specific subset, or be writing in a more personal way, not for public consumption. Young (2013, 168) found some researchers and some participants who expressed the view that people performed in particular ways with intent and understanding, but Young also found others who held the opposite view. There is often a far greater degree of nuance than might at first appear. Lange (2007) describes two different behaviours – people can be ‘publicly private,’ where they share personal identity information, but restrict access using a range of mechanisms. Alternately, they might be ‘privately public,’ by sharing and connecting widely, but restricting the amount of personal information they share. Tied up in this are three factors: identity information (identifying personal details), content relevance (how closely the shared information matches the interests of the intended audience) and technical access (privacy and password settings). So if someone is openly sharing their identity, not limiting access in any way and making the content relevant to the audience, then perhaps the researcher might feel less inclined to seek informed consent? Some would say that a researcher is clearly not a member of the intended audience; yet as an ethnographer and participant observer, haven’t they gained the status of being considered a member of the ‘community?’
In a further extension of the public-private continuum, Sveniningsson (2004, 56) advises adding the extra dimension of sensitivity of the information. On these axes she has medium on the horizontal, and information on the vertical. I’ve added indicative examples.
If the medium is considered private, or the information sensitive, then the researcher should handle the situation with discretion. If the medium is public and the information non-sensitive, the researcher might feel less obliged to seek consent.
So we need to think about the nature of the information shared and the space involved. The context, scale and temporality of the circumstances should inform any decisions. We need to remember also that different people will have different perceptions of public and private and display fluid behaviour that shifts along a continuum between the two. My feeling is that this brief observation from boyd & Crawford (2012) might prove particularly helpful in guiding researchers’ actions:
there is a psychological difference between being in public (sitting in a park) and being public (actively courting attention)”
In my next post, I’ll take a closer look at the issue of ‘consent.’
BOYD, danah and CRAWFORD, Kate (2012). Critical questions for big data: Provocations for a cultural, technological, and scholarly phenomenon. Information, communication & society, 15 (5), 662-679.
BOYD, danah and MARWICK, Alice (2011). Social Privacy in Networked Publics: Teens Attitudes, Practices, and Strategies. [online].
FORD, Sarah Michele (2011). Reconceptualizing the public/private distinction in the age of information technology. [online]. Information, communication & society, 14 (4), 550-567.
ROSENBERG, A. (2010). Virtual world research ethics and the private/public distinction. [online]. International journal of internet research ethics, 3 (1), 23-37.
SVENINGSSON ELM, Malin (2009). How Do Various Notions of Privacy Influence Decisions in Qualitative Internet Research?. Internet Inquiry: Conversations About Method. SAGE Publications, Inc. In: MARKHAM, Annette M. and BAYM, Nancy K. (eds.). Internet inquiry: Conversations about method. Thousand Oaks, CA, SAGE Publications, Inc, 69-88.
SVENINGSSON, Malin (2004). Ethics in Internet ethnography. In: BUCHANAN, Elizabeth A. (ed.). Readings in Virtual Research Ethics: Issues and Controversies. 45-61.
YOUNG, Kirsty (2013). Researching Young People’s Online Spaces. [online]. Negotiating ethical challenges in youth research, , 163-176.
1In the UK, we have an unfortunate legacy naming convention where public schools are actually private(!) and independent of the state.