Anissa Tanweer is a research scientist at the eScience Institute and is focused on human-centered data science. She completed her Ph.D. research as a 3-year ethnographic fieldwork at eScience. This setup, studying the institution which funded her research, highlighted some challenges common for ethnographers but also presented some unique ethical tensions. At her QUAL Speaker Series talk, Tanweer recalled these issues and offered her approaches to addressing them.
Ethnography as a research method emerged in anthropology, but today it is used in many disciplines. Throughout the method’s development, there were issues of power between researcher and research subjects that became especially prominent for various critical scholars. Aware of the potential challenges of this power dynamic and evolving critiques of ethnography, Tanweer found works to help inform her particular experience – such as Diana Forsythe’s “Studying Those who Study Us: An Anthropologist in the World of Artificial Intelligence.”
Three Areas of Tension When the Lines Blur Between Research Participants, Colleagues, Mentors, and Sponsors
Forsythe, who conducted ethnographic studies in a similar setting to Tanweer’s, offered the idea of “mutual vulnerability” in this type of work. But this concept came with risks too – when ethnographers place themselves in power settings, they risk surveillance, coercion, and even censorship. Tanweer’s view was that mutual vulnerability can be healthy and productive. Her approach to avoiding censorship, or even self-censorship, was to focus on shared interests and concerns with the data scientists she was observing and interviewing.
Anissa Tanweer, research scientist with eScience Institute, discussed ethical tensions she faced when studying the same organization that funded her dissertation work.
This approach led Tanweer to take a more active role in shaping the practice she was studying – data science as ethical innovation. There was a unique concern that arose here as well – the worry of coopting the ethnographer’s work for the data scientists’ goals. Tanweer decided to address this tension by moving herself from the observant towards the participant role on the anthropology spectrum.
This choice, she explained, led to yet another tension – her level of immersion was threatening to become too deep and dissolve her position as a researcher. Add to this the ethical concerns with studying the institution that was funding her research and Tanweer saw the situation as a choice – she could either expand her fieldwork to include another institution or entity, or approach the issue more creatively. She decided to obtain permission to name people and projects in her published works from the study, adding transparency to her ethnography.
Her solution, however, meant that she ended up focusing only on positive examples of data science for social innovation. Since these examples were constructive and there already existed plenty of negative examples, Tanweer saw this as a worthwhile outcome for conducting transparent ethnography. And to some extent, Tanweer said, these tensions are inevitable when studying the institution you are funded by.
An Ethnographer’s More Typical Challenges
Tanweer also shared her solutions for more typical ethnography projects, including:
Research Scientist Anissa Tanweer took questions from faculty and graduate students about her ethnography in data science project.
Mountains of data – “Be kind to yourself,” Tanweer told the QUAL audience. She kept thorough notes, bulleted most important points at the top of observation notes and interviews, and grouped her data by data science project, not just by chronological observation and interview data collections. Her recommendation for deep immersive ethnography – start with the observations notes, then read the interviews, which she used to illustrate the story she wanted to tell through the observations. Starting with the interviews is not wrong, Tanweer explained, but it would have told a rather different story compared to the themes she identified in her observation notes.
Making your way through – “Grounded theory is like a blackberry bush,” Tanweer said. it’s hard to see a path through, as you venture in, “you’ll get snagged, but in the end, you look back and there is a path.” She proposed a three-stage analysis process that includes germination, distillation, and refinement (see slides 33-39).
Authentic voice – It is difficult to stay in a disciplinary silo, but if a researcher does, they can learn to speak authoritatively; if they choose to go the interdisciplinary route, Tanweer warned, researchers often suffer chronic impostor syndrome. Co-authoring studies can help with the latter, she added.