Page 209 -

8.4 Who to interview 197

such as system requirements or overall evaluation of system operation, a broader
pool of interviewees drawn from all categories of stakeholders might be more
informative.
A stakeholder is anyone who is affected by the use of a system. Relatively simple
applications, such as games or Internet chat clients, might have only one type of user
that you would want to interview. Then again, simplicity might be illusory—novice
game players may have different perspectives from experts. Enterprise information
systems, such as university course registration and management tools may have mul-
tiple types of users (or stakeholders), ranging from administrators who approve pur-
chases of the tool and rely upon it for high-level reports, faculty who use it to manage
course enrollment and grades, and students who must register for courses and pay
tuition. In some cases, stakeholders may not be users at all: patients and their fami-
lies may have valuable insights regarding hospital information systems, even if they
never use them directly. For any reasonably complex system, you can expect that dif-
ferent groups of stakeholders will have very different perspectives on requirements,
necessary functionality, and usability. Interviews with representatives of all of the
stakeholder groups—or, at least, as many as possible—will provide a more complete
picture of the situation.
Particularly when you are involved in an ongoing, long-term project with an orga-
nization or a group of users, you may find that there are certain individuals who are
particularly good sources of information. These people may be particularly knowl-
edgeable about how relevant work is done, they may play pivotal roles in the organi-
zation in question, or they may simply be unusually forthcoming. These individuals
may play the role of key informants: individuals who are repeatedly called upon to
provide important insights, usually over an extended period of time. Key informants
can provide invaluable perspectives, if your interactions and relationship are well
managed. Key informants must be selected carefully and their insights must be vali-
dated by external confirmation from other sources. A disgruntled employee with an
axe to grind would not make a good key informant. Particularly in a workplace situ-
ation, you should take care not to abuse any information that might be used against
the informant (Gilchrist and Williams, 1999). Working with key informants is closely
related to—and, indeed, can be the first step in—ethnographic research, a topic dis-
cussed in detail in Chapter 9.
As with any other research, interviews should be conducted in a manner that
respects the participants (see Chapter 14). For studies that involve populations
of participants facing special needs or challenges, this may require extra care in
planning and execution. Le Dantec and Edwards' study of the information prac-
tices of homeless people illustrates some of these challenges. Noting that simply
looking for homeless people on the street could be problematic, they worked with
outreach groups who provided feedback and acted as mediators between the re-
searchers and the homeless interviewees. In consultation with caseworkers, they
offered participants a choice of store gift cards or public transportation cards as
incentives. Staff at the centers worked with the researchers to identify appropriate