Page 246 -

234 CHAPTER 9 Ethnography

Often, the human, social, and organizational aspects of information systems de-
velopment are the ones most critical to ensuring the success of a project (Harvey and
Myers, 2002). Ethnography can help in providing an understanding of the context in
which specific interfaces or systems are developed and implemented. While research
methods such as experimental design focus on reducing research to a small number
of hypotheses with findings that are easily generalizable to other projects, ethnog-
raphy focuses on the opposite: understanding the context of individuals in groups,
their processes and norms, at a specific point in time, without generalization as a goal
(Harvey and Myers, 2002). In addition, ethnographic approaches can be especially
good for designing technology out of a workplace context: “Designing for pleasure
demands a different approach from designing for utility” (Gaver et al., 2004, p. 53).
In a study aimed at understanding the importance of communication to multitask-
ing, researchers “shadowed” 19 workers at a large US corporation, noting all of the
workers' activities at their desks and following them around wherever possible. The
resulting 550 hours of data, including over 13,000 events, were analyzed and coded
to understand how workers switch between tasks, interlocutors, and communication
media. The finding that coordinating activities with multiple people was a stressful
and difficult activity led the authors to suggest that communication systems might be
designed to identify interruptions that might require significant coordination effort
(Su and Mark, 2008). The detailed records of communication behavior collected in
this study would have been difficult, if not impossible, to collect via other means:
observing the workers' activities at their desks, analyzing e mail transcripts, or oth-
erwise observing some subset of their activities would have given an incomplete
picture of the activities and interactions between modes of communication.
The example of the hospital information system (see Section 9.1) illustrates the
other primary goal of ethnography in HCI—to understand system requirements and
user needs. Successful design of complex or novel interfaces for use in unfamiliar do-
mains, requires researchers to build a detailed, multifaceted understanding of how the
work is done, how users interact, how tools are used, what users need, what policies
are in place, and other related questions. It comes down to understanding the context
surrounding where the information system will be used and who will be using it.
As in the case of the hospital information system, interviews, surveys, and other
simpler data collection techniques may not be up to the task. Ethnographic research
puts developers into the thick of the situation, letting them observe and study the
situation firsthand. Extending the hospital example, most computer developers would
not know how hospitals typically refer to patients. In a typical database design, data
about individual humans is often referred to by an ID number or their last and first
name. However, in hospitals, patients are often referred to by bed number. In a typical
database design, the ID does not represent anything physical or meaningful, but in a
hospital situation there is a physical meaning (the bed number or location) behind the
identifier. This is an important difference that might be uncovered using ethnographic
techniques but otherwise would not be obvious to the average researcher or developer.
The use of ethnographic investigations for understanding the requirements for a
computer system is closely related to a design philosophy known as participatory