Editors’ note: Though it plays with “advice column” in form, the goal of Yours Sincerely, An Uncertain Anthropologist is not to elicit prescriptive answers nor to develop universal disciplinary principles. The views and opinions expressed in the letters do not represent the breadth of perspectives within the field nor do they necessarily reflect those of the editors. By facilitating these exchanges, we do not expect to reach consensus or offer certainty. Instead, the aim is to create a space for grappling with ethical uncertainty and to invite ongoing dialogue within our professional and intellectual communities. Read more about this series and the call for letters here.
I recently submitted a proposal for a fieldwork grant funded by the U.S. government that requires applicants to provide a “data management plan” outlining how they will give other researchers access to the data they produce. In researching how to write a data management plan, most of the guidance I encountered focused on how to de-identify and safeguard data so that researchers can fulfill the social value (and grant requirement) of “sharing data” with the research community without compromising the ethical obligation to research subject privacy. For example, the American Anthropological Association’s data management modules provide practical information on de-identifying and preserving data alongside cautionary tales of anthropological data consigned to dumpsters because their writers had not prepared a “fieldnotes will.”
In the context of my ethnographic research, where the stakes of data security are relatively low, my biggest worry about “sharing” data is not the risk of re-identification (though I care about that too). Instead, it is the unknowability of how my field notes and interview transcripts might be used by others once they are detached from the conditions of their production—that is, from my relationships with people I know and care about, and all the commitments those relationships entail. What advice do you have for deciding whether (or not) to preserve and share data and how to ethically do so? Are (digital) dumpsters always a bad place for data?
Unsure about Sharing
Dear Unsure about Sharing,
This is a tricky question, and it’s one that the two of us have spent a long time thinking through. It’s extra tricky because there’s already so much out there about data sharing, but only a bit of it really helps us as anthropologists think about what we should be doing with our data. It’s one thing to upload an anonymized table of survey responses to a repository; it’s quite another to make the life stories, everyday impressions, and personal experiences public. Anthropologists know how to care for our interlocutors and our relationships when we translate our field notes into an ethnographic article or book. But how do we exercise the same care when it’s the field notes themselves that are out in the open?
You’ve already tuned into something crucial. Data embodies relationships, the most important of which might be the ones we have with our interlocutors. So the real question might be “How do we care for those relationships in the way we share (or don’t share) data?” What’s tricky is that data embodies far more than the relationships we have with interlocutors: mandates for data sharing often presume that the most important relationship embodied in data is the one between a researcher and the broader scholarly community, or between the researcher and the public that has funded their work. This might be fine for a medical researcher working on a major public health issue: their data tends to be derived from the very people that might benefit from broader attention to that data. But anthropologists can’t really take this for granted. We need to be mindful of different ethical concerns, such as the risks that data sharing might pose to vulnerable groups, or the uneven distribution of the benefits of the use of data. In addition, what works for one group of interlocutors will not necessarily work for another, so there’s no one-size-fits-all principle that can decide when and how we share what kinds of data. So the first step needs to be for us to try and make clear what relationships, what responsibilities, and what obligations our data embody, and how best we can care for the relationships that are most important.
This means that we should be asking questions about our data. Who has a stake in this data? Who will have access? Where will it be stored and how will it be tagged? What skills do you need to care for this data? One of us (Angela) has developed questions that have helped us think about data relations and what acts of care-through-data are most necessary. It might mean that we do not share data at all, or that we share some things, but only in certain places or to certain groups.
It could also mean that what you end up sharing is not what we usually think of as “data” at all. Things like interview protocols and IRB applications capture a slice of your relationships with your institution, your funders, your disciplinary peers, and your interlocutors. They may help others grappling with similar questions work their way towards their own answers with their own interlocutors.
It could also mean that the data we’ve gathered is not ours to share at all or in the ways that are too often taken for granted. In Aotearoa New Zealand, where the other of us (Grant) is located, it’s important to recognize that sovereignty over data is inseparable from Māori sovereignty. In some cases, the decision to share your data may not be yours to make.
To answer your last question: Neither of us think the digital dumpster is a great place for data to end up. Grant has done archival research, sifting through hundred-year-old letters between early information theorists in Japan and the US that dealt with things that were trivial at a time, but today have become invaluable for seeing connections that left few other traces. In Kenya, Angela has engaged with those who have repeatedly been subjects of research without anything to show from the research exchange, noting their feelings of being “over-researched.” Data as traces of relationships can allow us to see, for instance, that in Grant’s work, there were personal and intellectual exchanges happening between Japan and the US that contributed to the establishment of information theory. Or, as in Angela’s work, data traces can enable us to know the kinds of research investments that have already transpired in a community, and where her own efforts should be directed. Today, e-mails or text messages capturing these traces might vanish moments after they’re received, except for the tenuous marks that they leave in corporate repositories waiting to be mined. We should be thinking about what we can do as researchers to preserve and share data differently, which might help to keep in the foreground what the commoditization of data will erase.
Which is not to say that we’ll be making all our e-mails public. It’s natural to worry that our data might be taken out of context and misused. But it’s also worth remembering that it might have value beyond our own lifetimes that we can’t yet imagine. When so much data is becoming aggregated into “big” data, the stories, impressions, and experiences that our ethnographic data capture might turn out to be just what a 22nd century historian needs to look at our time otherwise.
Director of Programs, Code for Science & Society
Associate Editor, Engaging Science, Technology, and Society
Grant Jun Otsuki
Programme Director and Senior Lecturer, Cultural Anthropology
Te Herenga Waka—Victoria University of Wellington
Conference Co-Chair, 4S 2023 Honolulu
Associate Editor, Engaging Science, Technology, and Society
Dear Unsure about Sharing,
One of the things I appreciate about your letter is your willingness to take the completion of a Data Management Plan (DMP) as an occasion for ethical reflection, rather than just an administrative burden to be fulfilled. Early advocates of DMPs had high hopes for their potential to get researchers thinking about the data their work would generate, but in practice DMPs often suffer from what has been called “the boilerplate problem” (Keralis, Grumbach, and Potvin 2019). They rely heavily on templates or previous examples instead of grappling with the particulars of a project, as you are clearly doing.
I also sense in your letter a receptiveness to the idea that data sharing can have real value, beyond the practical need to comply with grant requirements. Travel restrictions during the COVID-19 pandemic, the environmental impacts of long-distance travel, and a growing awareness of research fatigue in overstudied communities have gotten many qualitative researchers thinking about creative opportunities for reusing data generated by others (e.g., Hamill 2022). And while, in my own work, I’ve argued that institutional efforts to promote data sharing and reuse can be understood as a bid to reshape researcher subjectivities (LaFlamme, Poetz, and Spichtinger 2022), I am nonetheless sympathetic to these efforts—especially when they take the distinctive needs of fields like ours into account.
As you decide whether and how to share the data resulting from the research you’ve proposed, the key question I would invite you to tarry with is this: what do you (in conversation with your interlocutors) perceive that you could ethically share, even if you cannot share everything? The Open Science movement has thankfully moved on from the maximalist and masculinist position that everything must be open, always. The influential UNESCO Recommendation on Open Science spells out justifiable restrictions on sharing that include the right to privacy and respect for human subjects of study, as well as the protection of human rights and of Indigenous knowledge. But rather than using those restrictions as an alibi to avoid discussions of data sharing altogether, I would like to see anthropologists move beyond an ideology of “all or nothingism” (Poirier et al. 2020: 220) to embrace partial, pragmatic solutions.
At the scientific publisher where I work, we’re talking more and more about controlled sharing as an option for researchers who work with sensitive data. There are two basic approaches to controlled sharing: transforming data and regulating access to data. Each of these approaches can be operationalized in different ways, and they can also be combined. For instance, textual materials like interview transcripts or field notes can be transformed by masking certain details or redacting longer passages altogether. Such materials can also be deposited in a repository that requires prospective users to complete a data use agreement, making access contingent on certain conditions being met. The dataset underlying my paper cited above employs both approaches, bringing together manual deidentification with a license that permits reuse for a specific scientific purpose only.
So, to what extent do these approaches address the problem of unknowability that you describe as your biggest worry? Transforming your data likely won’t help much; indeed, data transparency advocates have argued that masking selects attributes of a person or situation to be generalized away that may be precisely those of interest to future researchers (Jerolmack and Murphy 2019). Regulating access, on the other hand, allows you to be more selective about what constitutes acceptable use. It’s best if you aren’t the only person reviewing requests for data, both in the event that you can’t be reached and to avoid conflicts of interest. But many repositories are happy to have depositors and source communities stay involved in such decisions, especially when changing conditions on the ground justify reevaluating existing arrangements.
A closing thought about unknowability: being unable to fully anticipate how our research will circulate or to what uses it will be put is, I submit, part and parcel of the ethnographic encounter. True, the works we produce for publication can be carefully constructed to leave out, where necessary, what may have been captured in the moment. But even these works have lives of their own (e.g., Fassin 2017). Grounding the decisions we make about sharing research outputs in the relationships that gave rise to them is important, of course. Yet this need not extend to claiming total knowledge of or control over what will happen next, not least because our interlocutors are unlikely to believe us. Rather than setting ourselves apart from the fallibility of intentional action in their social worlds, the more honest move may be to take our places there in the mess, doing the best we can by each other without promising what we in the end cannot—that is, protection from all possible harm.
Open Research Manager
Fassin, Didier. 2017. “Introduction: When Ethnography Goes Public.” In: If Truth Be Told: The Politics of Public Ethnography, edited by Didier Fassin, pp. 1–16. Durham, NC: Duke University Press.
Hamill, Julia. 2022. “Reduce Reuse Recycle: Secondary Analysis and Secondhand Data as a Lower Carbon Method.” In: DIY Methods 2022 Conference Proceedings, edited by Anne Pasek, pp. 328–359. https://doi.org/10.17613/dq0x-gs49
Jerolmack, Colin, and Alexandra K. Murphy. 2019. “The Ethical Dilemmas and Social Scientific Tradeoffs of Masking in Ethnography.” Sociological Methods & Research 48(4): 801–827. https://doi.org/10.1177/0049124117701483
Keralis, Spencer D.C., Elizabeth Grumbach, and Sarah Potvin. 2019. “The Boilerplate Problem in Data Management Plans.” ResearchDataQ, October 21. https://researchdataq.org/editorials/the-boilerplate-problem-in-data-management-plans
LaFlamme, Marcel, Marion Poetz, and Daniel Spichtinger. 2022. “Seeing Oneself as a Data Reuser: How Subjectification Activates the Drivers of Data Reuse in Science.” PLOS ONE 17(8): e0272153. https://doi.org/10.1371/journal.pone.0272153
Poirier, Lindsay, Kim Fortun, Brandon Costelloe-Kuehn, and Mike Fortun. 2020. “Metadata, Digital Infrastructure, and the Data Ideologies of Cultural Anthropology.” In: Anthropological Data in the Digital Age: New Possibilities, New Challenges, edited by Jerome W. Crowder, Mike Fortun, Rachel Besara, and Lindsay Poirier, pp. 209–237. Cham: Springer.
Anonymous. 2023. “Yours Sincerely, Unsure About Sharing.” In “Yours Sincerely an Uncertain Anthropologist,” edited by Paige Edmiston and Alexandra Dantzer, American Ethnologist website, August 25 2023, https://americanethnologist.org/online-content/collections/yours-sincerely-an-uncertain-anthropologist/yours-sincerely-unsure-about-sharing/
Okune, Angela,. Otsuki, Grant J. 2023. “Dear Unsure about Sharing” In “Yours Sincerely an Uncertain Anthropologist,” edited by Paige Edmiston and Alexandra Dantzer, American Ethnologist website, August 25 2023, https://americanethnologist.org/online-content/collections/yours-sincerely-an-uncertain-anthropologist/yours-sincerely-unsure-about-sharing/
LaFlamme, Marcel. 2023. “Dear Unsure about Sharing” In “Yours Sincerely an Uncertain Anthropologist,” edited by Paige Edmiston and Alexandra Dantzer, American Ethnologist website, August 25 2023, https://americanethnologist.org/online-content/collections/yours-sincerely-an-uncertain-anthropologist/yours-sincerely-unsure-about-sharing/