While ethnography generally draws on qualitative data, it does not not mean that quantitative approaches shouldn’t be employed in the research process. Combining the two leads to a “mixed-method approach” that can take various forms: data collection and analysis can be either separated or addressed together, and each of them can be used in service of the other. Of course, this isn’t new in academic circles and corporate ethnography but there seems to be a renewed interest lately in this topic.
One of the driving forces of this renewed interest is the huge amount of information produced by people, things, space and their interactions — what some have called “Big Data“. The large data sets created by people’s activity on digital devices has indeed led to a surge of “traces” from smartphone apps, computer programs and environmental sensors. Such information is currently expected to transform how we study human behavior and culture, with, as usual, utopian hopes, dystopian fears and *critical sighs* from pundits.
Although most of the work of Big Data has focused on quantitative analysis, it is interesting to observe how ethnographers relate to it. Some offer a critical perspective, but others see it as an opportunity to create innovative methodologies to benefit from this situation. See for instance the notion of “Ethnomining” described by Aipperspach et al. (2006) in their insightful paper Ethno-Mining: Integrating Numbers and Words from the Ground Up:
Ethno-mining, as the name suggests, combines techniques from ethnography and data mining. Specifically, the integration of ethnographic and data mining techniques in ethno-mining includes a blending of their perspectives (on what interpretations are valid and interesting and how they should be characterized) and their processes (what selections and transformations are applied to the data to find and validate the interpretations).
For these researchers, such integration is meant to highlight new understandings and potentially inspire design in human-computer interaction research. We recently featured a series from Wendy Hsu, an ethnographer who uses data mining and GIS techniques along with ethnographic research.
Another driving force underpinning the renewed interest in mixed-methods may also be the result of cross-disciplinary programs in different universities. This hypothesis is based on my personal experience visiting different places in Europe and in the US where I have noticed an increasing number of courses about ethnography and qualitative research in computer science departments and design schools. Graduate students are being asked to understand the value of different approaches and to strive to integrate them for their own purposes. Hence the work on “ethnomining” by researchers from a Computer Science division at UC Berkeley.
That being said, the mixed-method approach, whether involving large data sets or not, is not so straight-forward. There are potential problems worth exploring. The most important issues lies in the fact that qualitative and quantitative methods do not necessarily mix easily at the epistemological level: how do positivist assumptions embedded in quant research mix with more interpretive standpoints? Another problem also consists in the triangulation process between data: should they only be to the service of one another? Or is it possible to collect and analyze both types of data in a more integrative way? Then what does all this mean in a practical sense? Finally, as discussed by danah boyd and Kate Crawford, the large data sets we can use have their own challenges around what is considered to be “truth.” They point out that “what is quantified does not necessarily have a closer claim on objective truth“.
Building on these discussion, this month’s “Combining qualitative and quantitative data” theme will give an overview of current opportunities and issues. The post series will not focus only on ethnomining, but it will show various case studies and perspectives on the implications of mixed-methods approaches. Here are some posts that we have coming up in this edition:
- Rebekah Rousi (@RebekahRousi) will describe how she combined questionnaire results with on-site observations to investigate how people experienced their interactions with elevator designs.
- My colleague Fabien Girardin (@fabiengirardin) will show how he used sensor data to yield field observations in a study for Le Louvre in Paris.
- Rachel Shadoan (@RachelShadoan) and Alicia Dudek (@aliciadudek) will describe the results from their research on Plant Games, an online Role Playing Game.
- Alex Leavitt (@AlexLeavitt) will discuss his research on Tumbler using a computational ethnography perspective.
- Tricia Wang (@triciawang) is going to share her thoughts about the opposite of Big Data, in what she calls “thick data”.
- David Ayman Shamma (@ayman) from Yahoo! Research will describe his personal perspective on the topic.
For each of these blogposts, beyond the results and the authors’ viewpoints, I think the most fascinating bit concerns the motivation to combine qualitative and quantitative methods, as well as the role played by the research focus in this choice. Many of the issues were brought up in contributing editor Jenna Burrell’s series, “The Ethnographer’s Complete Guide to Big Data.” We hope this month’s edition continues to extend the conversations around ethnography and big data.
We’re looking for guest contributors for Jenna Burrell’s May edition on talking to companies and organisations about ethnographic fieldwork. Check out the upcoming themes to see if you have something to submit!