by Danny Lämmerhirt
The COVID-19 pandemic and the mediatization of social life raised lively discussions and concerns among academic and industry ethnographers. Rather than adding to discussions how to cope with the disappearance of ‘on-site’ research, I would like to speak about how COVID-19 prompted me to rethink some common conceptions of ethnographic research, and the productive effects this had on my research design as well as the theoretical stance I took towards ethnographic fieldwork of data infrastructures. While COVID-19 closed off several opportunities for field research, it not only opened up entirely new field sites, but also forced me to rethink concepts like ‘everyday practices’, ‘presence on-site’, or the ‘situatedness’ of people, but also adjacent topics such as the role of ‘mobility’.
My dissertation project was in a fairly early stage when COVID-19 put public life to a halt. Originally, I planned to study health data sharing platforms and visit various actors – from patient groups to researchers, to understand the meaning of these platforms in their lives. I had envisioned to spend time in a bounded social milieu or organisation and to follow people around who are involved in it, not unlike laboratory studies where I could follow people around and observe how they make these infrastructures work for them.
Opening new sites as society turns into a living laboratory
Many commentaries have been written on how COVID-19 turned society into a ‘living laboratory’ and how all of a sudden everything seemed to relate to medical and health concerns. These experimental setups saw the rise of consumer companies like Apple and Google influencing public health policy through the management of app releases via their app stores (Sharon 2020), but also raised the public profile of otherwise rather sector-specific topics such as using digital biometric data for passive remote health monitoring, or legislation to enable patients to donate their clinical data to research (Radin et al. 2020).
‘Attuning’ to COVID-19 meant to follow these societal responses for a while to see if they might bring up a relevant case. It required a monitoring protocol to compile and make sense of publicly available information. For this, I captured topical tweets via the Twitter Capture and Analysis Tool (DMI-TCAT) and monitored the evolving debates by looking at topics, keywords, hashtags, and Twitter actors. Following Marcus’s elaborations on multi-sited ethnography – my research first had to ‘pass through zones of specialized, technical knowledge before it define[d] the traditional fieldsite’ (2011).
As someone studying health data infrastructures, I naturally followed the passionate and important debates over contact tracing apps, their potential ‘function-creep’, and the new roles of Apple and Google as public health regulators. While important, they spoke to concerns of epidemic response, or programmable infrastructures rather than to my interest in studying health data exchange platforms. Prior to contact tracing, Germany’s CDC had rolled out a data donation app, inspired by earlier research in the US, asking people to donate biometric data to build an early warning system against influenza.
Following this project for a while, I learned that besides being a culturally and ethically charged mode of exchange (a donation), the project is much more interesting as it raises the question: under what circumstances digital data counts as ‘health-related’ in the first place? Suddenly many actors had a stake in flu predictions with companies like Oura Ring promising to be able to predict COVID-19 with 90% accuracy, people claiming that influenza detection is a ‘huge step’ to remote monitoring of patients, and others criticizing that such claims may erode public trust in science. Here it was, a full-fledged debate around the value and societal uses of biometric data, addressing economic, epistemic, and cultural concerns. COVID-19 not only opened new case studies but expanded my original focus from data exchanges to translations of consumer biometric data into pathological insights.
‘Attuning’ ethnographic methods to COVID-19 – or: How to write ethnographically about data exchange platforms in a pandemic?
Still, I asked where to observe ‘everyday experiences’ or ‘practices’ involved in data donations. Prior to COVID-19, I envisioned my cases to somehow be organizationally bounded – perhaps because I initially described them in organizational metaphors like ‘cooperatives’. With COVID-19 basically cancelling all physical presence, and my cases being mostly distributed, automated, mediated, and data-based infrastructures, I started looking into infrastructural ethnographies (Leigh Star and Bowker 1999), multi-sited ethnography (Marcus 2009), and more generally work on mobility ethnographies, and how webs of relations and systems shape people (such as the work of Appadurai, Tsing or Fortun).
Here, Marcus’s ideas of multi-sited ethnography were useful as they suggested moving from the study of situated subjects and their practices to the relations they are involved in, and to focus the field site with the subjects being co-investigators or para-ethnographers (2011). As my cases -– data exchange platforms – had specific kinds of mobility, movement, and flows at their core (e.g. of data across devices, sensors, databases, exchange protocols), I saw an opportunity to study how these databases mobilize data in the first place, and with what purposes they have been designed. This drew my attention to the specific ‘centres of calculation’ (to use STS language) that exchange data, and to lead interviews with the people involved. As Marcus would say, it helped to bound and appropriately account for my strategically partial research sites (2011, see also Strathern 1996).
The Corona-Datenspende app gladly received quite some public attention and also publicly communicated its research process. The RKI had the intention to share back for people’s willingness to give their data and documented its analyses, and assessments whether the data donation model ‘works’. I could follow my original intention to study data practices at least in part by seeing publicly available fever visualizations. In addition, publicly available user comments on app stores not only pointed out hopes, fears or frustrations with the app but also sometimes documented how people interacted with the app. The challenge was now to thicken such thin data.
That being said, two of my cases are data platforms that remained largely closed to access. For now, I could primarily access them through websites, technical whitepapers, or other public-facing traces documented in online environments like GitHub. Besides using remote methods such as self-disclosed participant diaries, or screen shares, I’m currently exploring methods to study distributed organisations ethnographically (see for instance Geiger’s and Ribes’ trace ethnography, 2011) and materially (Ribes 2019), or to study how organizations relate to their environments (for instance via technical documents, patents, and other devices, proposed by Lepage-Richer 2019).