Page 231 -

8.10 Analyzing interview data 219

in traditional surveys, possibly because responding directly to a computer (with-
out the presence of a human) encourages more openness in responses (Walston and
Lissitz, 2000). In comparison, in one study of behavior relating to HIV/AIDS in
sub-Saharan Africa, interviews were administered using both paper and computers.
In both cases, interviewers asked the questions to the interviewees and recorded the
answers on the paper or computer. Analysis of the results indicated that participants
who were interviewed with the computers were more likely to provide socially desir-
able answers regarding risky behaviors. Although more study would be needed to
understand these responses, the researchers conjectured that interviewers who used
computers may have appeared to have been either affluent or outsiders. Participants
may simply have been trying to impress the interviewers (Cheng et al., 2008).
Online conferencing tools provide additional context that can narrow the gap
between online and in-person interviews. The combination of real-time audio, which
is obviously necessary for a conversational flow, and video, which can restore some
of the visual cues associated with face-to-face conversation, can be almost as good
as being there. As capabilities vary across service providers, you might want to ex-
periment with multiple tools to find those that work best. Screen sharing and re-
cording tools are often useful for conducting contextual inquiries online, as they
allow detailed exploration of the participants' use of systems, with recordings cap-
turing each interaction with the system. Other tools may provide remote mouse and
keyboard input, allowing users to interact with software running on your computer.
Although perhaps not as personal as in-person interviews, such approaches, like all
electronically mediated interviewing techniques, can be a cost effective means of
reaching a broader range of participants.

8.10 ANALYZING INTERVIEW DATA

Having conducted a series of interviews or focus groups, you'll find yourself faced
with the daunting task of interpreting your data. Countless pages of written notes and
hours of video or audio recordings pose a significant challenge—how do you make
sense of it all? Your goal in analyzing interview data is to generate an accurate repre-
sentation of interviewee responses. Usually, your analysis works towards a general,
holistic understanding: the analysis of answers to individual questions are combined
to form general models of user needs for a particular task, reactions to a proposed
design, or other focus of the interview. This may not be possible—you may find that
there are no consistent patterns. This is interesting as well.
Whichever techniques you choose to use, you should try to analyze your data
as soon as possible. When the interview is fresh in your mind, you will be well-
positioned to remember details and nuances that you may not have captured in your
notes. As time passes, you will find it increasingly difficult to remember potentially
important nonverbal cues or comments. Your notes will also become less useful over
time, as hastily scribbled cryptic comments will be hard to interpret weeks or months
later.