Ethnography
What is ethnography?
Ethnography is an observational technique that is well established in the fields of sociology and anthropology. It involves immersing an individual researcher or research team in the everyday activities of an organization or society, usually for a prolonged period of time. The ethnographic approach is very naturalistic, and provides the kind of information that is impossible to gather from laboratory or "clean room" observational studies.
What does ethnography have to do with HCI?
The task of designing an interactive system and developing viable requirements specifications can often overlook the most important piece in the entire development puzzle - the users. It is becoming increasingly evident that a good interactive system must take into account the social factors involved with those who use the system on a day to day basis. If a new system adversely affects or drastically changes the way that a user base interacts in order to complete a task, then it will be very unlikely that this design will gain acceptance.
The more formal methods of design evaluation, such as Task Analysis, are usually lacking in the areas that ethnography excels in. The information gathered from an ethnographic study allows developers to design systems that take into account the sociality of interactions that occur in the "real world". Ethnography has proven especially effective in domains in which the interactions are confined to a small scale, such as an industrial plant control room or an air-traffic control room. By allowing the ethnographic researcher access to the subtleties of the everyday work environment it allows the interactive systems designer more perspective on how the system should be really be used.
The information gained from an ethnographic study allows us to see how a prospective user base interacts with one another in order to complete tasks. Thus, the main goal of ethnography when applied to HCI appears to be a kind of synergy between the interactive system and the way a work society is organized to carry out everyday work tasks. >
Drawbacks to ethnography
Like many of the more qualitative evaluation methods, ethnography does not mesh easily with the rigorously structure areas of design and software engineering. The information gathered from an ethnographic study is often very discursive and difficult to translate into tangible design requirements. However, this kind of information is never without value, and it points to the growing need to combine formal methods and quantitative evaluation with the more socially aware qualitative evaluation methods.
Also, it is still unclear whether or not allowing an ethnographic research team access to the day to day workings of an organization will be acceptable. The researcher must be as unobtrusive as possible, and must gain acceptance within the community he or she is observing. This has proved feasible within small social domains, but whether or not this will scale into larger organizations is still somewhat of an open question.
Finally, the sheer cost of supporting an ethnographic study that can often span several years is prohibitive to its acceptance as a viable evaluation technique. There are, of course, areas where this is not as much of a factor (such as the evaluation of an air-traffic control system), but it simply isn't possible to do this kind of in depth study for more mundane applications.
Case studies
The role that ethnography has played in interactive system design has thus far been extremely limited. The main success stories that advocates point to have usually been research projects.
One example of successful application of ethnography to systems design can be found in the paper The Role of Ethnography in Interactive Systems Design (ACM, Interactions - Hughes, King, Rodden, Anderson - April 1995) in which the authors detail an ethnographic study of the London Air Traffic Control. The three year systems design project included eighteen months of ethnographic study. From the information gathered by the researchers, they were able to evaluate what information was essential to the controller, what was less important but needed to be "ready to hand", and what was inessential. They also found it very important to keep the controller "geared into" his work, and they found that this could be done by not automating some aspects of the controller's work that were most familiar in the current setting. For example, they found that the ordering of the screen-based flight strips should not be automated, in order to retain some of the functionality of the current paper flight strips that the controllers were used to working with.
Another application of ethnography to HCI can be found in the paper Collaborative Activity and Technological Design: Task Coordination in London Underground Control Rooms (ESCSW '91 - Heath, Luff). The researches realized, from ethnographic study, that it was essential for the technicians in the subway control rooms to be able to effectively monitor and communicate information that they and their colleagues were observing on the panels and displays that they were assigned to. Thus, they were able to communicate to the system designers who were doing the evaluation and redesign of the current system these needs, and the new design focused on making the interactions and information sharing between controllers more efficient and more natural.
The future of ethnography in interactive systems design
Most advocates of ethnography realize it's shortcomings and often stress that it is a complimentary evaluation technique that should not attempt to replace formal methods, but rather augment them.
The areas in which ethnography seems to be most strongly suitable for are not in the initial design process, but in the iterative process of refining and evaluating a design once it has been deployed. The cost prohibitions to full-scale ethnographic study can be compensated for by doing "quick and dirty" studies in which a field researchers is put in place for several weeks to assess the usability of a design prototype.
Most of ethnography's advocates stress a need for the application of ethnography throughout the lifetime of the design process, not just at a certain critical juncture.
Here are some examples of the role that ethnography can play in the system design process:
Ethnography on the net
Wendy Newstetter has done ethnographic research here at Georgia Tech, and there are some pointers to papers on this topic at her website.
Papers on ethnography
The information presented above was gathered from the following sources
The Role of Ethnography in Interactive Systems Design. Hughes, King, Rodden, Anderson. ACM Interactions April 1995.
Making Work Visible. Suchman. Communications of the ACM Vol. 38 No. 9 - September 1995
Collaborative Activity and Technological Design: Task Coordination in London Underground Control Rooms. Heath, Luff. Proceedings of the Second European Conference on Computer-Supported Cooperative Work. 1991.
Transforming Work: Collaboration, Learning, and Design. Sachs. Communications of the ACM Vol. 38 No. 9 - September 1995