In fall 2014, I began building Tinn, a health data tracking application for people living with tinnitus. I approached building Tinn as an opportunity to explore what a socially conscious, feminist, and anti-colonial process for developing mobile applications might look like. The idea of pairing design, building, and anthropology is hardly all that innovative; “design anthropology,” a subfield of cultural anthropology focused on empirical studies of use cultures, has been around since the 1980’s (at least). What sets Tinn apart, however, is my commitment to working with communities of color in Portland, OR, that have been alienated by and from the technology and application development industry because of intersecting and juxtaposed systems of gender, racial, and access inequality. Perhaps one of the central epistemic problematics for the project, then, can be posed as a question: Can we decolonize design anthropology, and to what success ordegree? What might this entail?

[Image no longer available.] Screenshot of the title deck for Tinn (by author).

Decolonization is slippery. My “academic anthropology” is a historical ethnography of the ways media scientists after World War II shaped (and continue to shape) the gendered contours of global media flows in the British Empire, with a particular focus on sound, listening, and citizen-subject formation in Gibraltar. My work on Tinn gave me the opportunity to transform my critique of the violence done by media scientists into the starting point for a more ethical approach to user experience research with marginalized Native American, Latin@, migrant, and African American Oregonians living with tinnitus in Portland. Yet what I thought of as decolonizing and what my collaborators thought of as decolonizing was at odds in some instances. For one, while decolonizing anthropology attempts to re-balance the scales of recognition and redistribution in the research process, it is much more difficult to reconcile the colonial histories of designer and designed for. Yet, for my collaborators, this division didn’t actually matter. As Nephi, one of my politically astute collaborators, put it, “the ethics are too heady when we need material help. Someone has to do that work. It’s you today.” While Tinn began with my commitment to making the project open source (as resistance to the privatization and commoditization of collaboration — it’s not that simple), Nephi protested. “My labor is there, too. You’d give it away for free? There’s a long history of white men giving away the work of my people for free.” I said it wasn’t simple.

While there were times where my collaborators and I didn’t agree on what constituted decolonization, we did agree on one thing: data poses a particularly tricky sociohistorical landscape for thinking about recognition, redistribution, and reconciliation. The rest of this post is dedicated to the complications of tracking data about tinnitus, and tracking data about vulnerable and/or marginalized users with tinnitus.

Tinnitus and technology

In general terms, tinnitus is a condition where people experience varying degrees of “buzzing” or “ringing” in their ears — often called “ghost sounds” — accompanied by varying levels of discomfort and pressure. It is often either the result of intense and repeated hearing trauma or the complication of another type of physiological or psychological trauma, although 11% of those living with tinnitus have a genetic version of the condition. One of the challenging things about tinnitus is that, like Bowker and Star say of tuberculosis, flare-ups and stimuli are hyper-individualized and difficult to categorize: diet, technology use habits, weather, stress, social activity, health states, and so forth, all affect people living with tinnitus differently.

[Image no longer available.] Image: Tinn running on a Samsung Galaxy (before the recent Material Design Android update). Photo by author.

Tinn is collaboratively designed to enable (some collaborators say empower) people living with tinnitus to track, quantify, and visualize their daily habits to better understand what might cause their own flare-ups. The app pings users, asking them to enter data about their daily habits at different times of the day. When users experience particularly intense discomfort, they press a button and Tinn gathers all of the past data and visualizes it, helping users recognize habitual and behavioral patterns that may be the cause of their tinnitus. Perhaps most importantly for this blog post, it is produced through an ethnographic engagement, a collaborative and iterative process of building that places a feminist ethics of reducing (as much as possible) the histories of domination at the core of both anthropological and technological innovation. This is, indeed, a crucial element of the project, as Tinn seeks to disrupt the common practice of what Melissa Gregg has described as the universalization of white, heterosexual, executive men’s use habits through design practices devoid of cultural critique. All of my collaborators are people living with tinnitus from marginalized and under-served, intersecting communities — migrant laborers, African Americans, Latinas, and Native American and Aboriginal Oregonians. Tinn, in short, is an experiment in using design ethnography to decolonize (insofar as it is an aspirational and never fully accomplished political process) both design and and ethnography. This is undoubtedly a fraught process when we talk about ethnographic “fieldwork” and data tracking. To start with, urban “fieldwork” in the United States is a form of knowledge production “straight out of plantation slavery, plantation management, and plantation geographies that were laboratories for black subjection and black resistance.” For Tinn, this methodological history of domination is exacerbated by the historical content of the project itself. If the entangled histories of self-help and racial uplift weren’t fraught enough in a project that revolves around health improvement via data tracking, then the tracking of user’s phone use habits and behaviors as a supplement if not replacement for participant-observation pushes us into ethically and politically dangerous territory (as I’ll describe below).

Data in its many guises

Use data, in fact, has been essential for this project. For design anthropologists, who are often situated in corporate atmospheres with multiple simultaneous projects, the idea of doing extended “fieldwork” is more a fantasy than reality. Situated within an atmosphere of meetings, project management, budget meetings, planning the future, and other business tasks, it is often impossible to dedicate a year — or a month, or even a week — to following one or a community of users. So, many of the design-focused anthropologists I’ve met in the past year supplement their fieldwork with something else. For me, the supplement was use, or “secondary” data. Where users create primary data (text messages, photos, browser history), phones create secondary data about users such as when they use applications to produce primary data, in what sequence they use them, and how they use them (hence, use data — which is wholly separate from tertiary cached data). Like participant-observation, use data allowed me to look at when, where, and in what succession my collaborators were using Tinn. Perhaps the most important part of participant-observation from a research (not necessarily ethical or political) perspective, use data allowed me to see the distinction between what people said they do and what they actually do.

[Image no longer available.] Image: A User-Centered Design approach to transforming categories of experience into information architecture using stickynotes. Photo by author.

For example, in the early stages of developing Tinn, I used paper prototypes: a combination of printed storyboard designs and worksheets that allowed my collaborators and future users to participate in a paper version of the digital application. These paper prototype worksheets were developed out of interviews where we discussed and played with the categorization and organization of data and their daily lives. By the end of the paper phase, everyone seemed to agree that they enjoyed a data logging schedule that divided the day into seven parts: Sleep/Wake up, Morning, Mid-Morning, Noon, Afternoon, Evening, Late Evening. But when we look at the usage data, we see that while people do log in the granularity of these times, they don’t actually log at these times. Nephi, for example, would typically log when she woke up, then again in the afternoon, and finally in the late evening (before she would go to bed and/or stop using her phone for the evening). Use data, then, gave me a critical insight otherwise missing from interviews — and which greatly affects what I might perceive at a rational level as the efficacy of data tracking.

But just because participant-observation, that pre- and post-Malinowskian surveillance and management technique, is gone does not mean that design anthropology is not haunted by the spectres of colonialism. That’s because it’s not necessarily or just counting and surveillance — that is, method — that makes anthropology a colonial practice, but also the historical conditions in which those practices pretend not to exist. That is to say, anthropology of all flavors is epistemically a colonial project until it faces and attempts to mitigate that. And even then, anthropology remains a colonially-inflected intellectual project (at least in its American form). Indigenous, Black, Feminist, and allied anthropologists have and continue to theorize how we might contend with the discipline’s involvement in colonial domination and administration.

Contemporary colonial inflection is certainly present in my use of my collaborator’s mobile use data. Use data is typically done in the background of mobile applications, uploaded to a database, and disseminated to researcher(s). This made one of my collaborators, Adolpho, particularly nervous. An undocumented laborer in his 40s, his tinnitus is the result of years in construction (where undocumented laborers are rarely afforded adequate health precautions and protections) and has made him functionally deaf. Adolpho was suspicious of data collection and research (it was only through a common connection with a friend in the housing authority that he agreed to meet with me). We communicated primarily through writing on a notepad, passing it back and forth, since it made it easier for him to communicate. “You see where I am at any time” he wrote on his notebook. It was a question he pointed at many times throughout our first and second interview. It became such a frequent question that we created a shortcode for it, which he could write any time he had a concern about surveillance. After our third meeting, Adolpho’s daughter walked me to the door, explaining Adolpho’s nervousness. “I don’t know if it’s true,” she said, “but my aunt thinks that my cousin was caught by INS through his Facebook. My father is nervous the same could happen to us.”

One of my other collaborators, Nephi, put the infopolitics of data accrual in historical context. Nephi is in her mid-20s, and identifies as a Black and Native American woman. She is currently an undergraduate “non-traditional” student at university, majoring in business administration with a minor in history. Her tinnitus is the complication of her experiences with eating disorders as an adolescent woman trying to “fit in.” “Whites have always tracked my ancestors and relatives somehow,” she described to me in our third meeting. “You don’t get to control my data. I know you get it.” In her final phrase, she referenced a long conversation — approximately two hours — she and I had about the history of Native and African Americans in Oregon and world history. Anthropologists, she had described quite correctly, have long been implicated in the monitoring and tracking of indigenous populations, a practice that lives on today through the Indian Registry and Identification Systems that “legitimate” Native and Aboriginal Americans as sovereign-yet-American. The identificatory system has historically been used to track Native American “subversives,” who were detained with limited-to-no rights under international espionage laws.

De-colonizing data

[Image no longer available.] Image: Tinn’s home screen in prototype (by author).

In this sense, even without participant observation, Tinn is implicated in anthropology’s history — and present, if we are to think about the U.S. Human Terrains Science and British GCHQ – and haunted by the historical weaponization of information and knowledge in the service of protecting racialized borders. To decolonize design anthropology and the production of Tinn required a mitigating design that is rarely afforded to our research subjects in vanilla, academic ethnography: that is, the possibility of being forgotten, of going unseen for all or certain swaths of time. Tinn, once made into a high fidelity prototype, was uploaded onto some of my collaborators phones directly from my computer. The application did not communicate with anything outside of itself directly — insofar as I could control (carrier-committed phones are notorious for collecting information). The application also recorded researcher use data and personal health data separately; what data was available to me was kept separate from the data they were logging. If the project was successful in any regards, it was with the inclusion of a button that erased all of the researcher data while maintaining users’ personal data. At any time, in any place, my collaborators could delete all of the data that I would upload from their phones every week (at least since the last time I had either uploaded or they had pushed the button). This empowered my collaborators not only to use their apps with less anxiety about being seen and tracked by me, while empowering them to do data collection. That is also why I can only say typically log data three times a day, rather than definitively.

In this way, however, Tinn sheds light back onto anthropological methodology itself: while we constantly think of data science and use data as a scientific endeavor, which implies some sort of totality, it can be bound to the same strictures as participant observation. Just as we cannot “hang out” with our collaborators/informants/interlocutors/subjects/etc. 24 hours a day (nor do we actually want to in most instances), data science is bound by the strictures of time, space, and permission. When people forget their phones at home, they do not produce data (even while the phone does). When people’s phones die, they do not produce data. And, in the opposite direction, while the button on Tinn allowed my collaborators to metaphorically “un-write” themselves from my pseudo-fieldnotes, no such opportunity is afforded to interlocutors, subjects, informants, or whatever name we call people in a lexical attempt to hide the history of spying we’re participating in as anthropologists.

So in the end, beyond the application itself, Tinn forces us to ask complicated questions about the history and evolution of the field in its various academic and corporate pathways — academic, corporate, design, innovation, business, intellectual, and so on. For me as a privileged designer and builder, Tinn revealed the ways in which histories of my privilege, training, and professional trajectory cannot be undone in the research process, and that decolonization is not an overcoming so much as a collaborative mitigation of violence that, always imperfect and difficult, must be done in the baby steps we make towards more equitable futures. As an intellectual intervention, then, Tinn has been kind of successful but not perfect — a classic story of political work within historically colonial epistemic frameworks.

Acknowledgments: This research has been possible because of a Julie & Rocky Dixon Doctoral Research Fellowship in Graduate Innovation from the University of Oregon. This fellowship allowed me to work with Dawn Nafus and her colleagues at Intel Labs, and expand on UXR research that I had started with Carol Stabile and Karen Estlund building Fembot. Associate Editor Jamie Sherman and Editor-in-Chief Jordan Kraemer provided amazing direction for and feedback on this essay, which has also benefited from ongoing conversations with my colleagues and friends Gennie Nguyen, Jessica Hardin, and Mark Auslander. Errors are my own, as is the gratuitous linking throughout the piece. 

Bryce Peake is a media anthropologist and historian of communication technology. In fall 2015, he will be an Assistant Professor of Media & Communication Studies at the University of Maryland, Baltimore County (UMBC).