Page 187 -

174 CHAPTER 7 Case studies

on their own as research results, providing insights into the larger group of re-
spondents not selected for closer examination in your case study (Yin, 2011). See
Chapter 5 for advice on conducting surveys.

7.8 DATA COLLECTION
Having defined your research questions, chosen the number of cases and the units of
analysis, and determined whether your study is embedded or holistic, you are ready
to plan your data collection. Specifically, you need to define the types of data you
will collect and the specific procedures you will follow for collecting those data.

7.8.1 DATA SOURCES AND QUESTIONS
As described in Section 7.2, case studies often, if not always, rely on multiple data
sources. Data sources for case studies in other fields include documentation, archival
records, interviews, direct observation, participant observation (similar to ethnog-
raphy), and physical artifacts (Yin, 2014). For HCI research, you may find yourself
adapting and adding to this list as appropriate. If you are trying to understand some-
one's use of existing computer tools, e-mail messages, web history logs, and related
data sources may be considered archival. Logs of specific activities with applications
of interest might be available or you might be able to use a variety of technical ap-
proaches for collecting such data (see Chapter 12).
Your research questions and hypotheses will play a significant role in determin-
ing which of the available data sources you will use. Documentation and archival
records are likely to be most interesting if you want to understand past and current
practices and use of existing software tools. Interviews are helpful for understand-
ing perceptions, concerns, needs, and other user reactions. Direct observation can
help you understand what people do in circumstances of interest, while participant
observation can be a powerful tool for understanding complex organizational dynam-
ics. For HCI researchers, artifacts can be used to provide valuable examples of how
people bridge the gap between computer work and the rest of their lives. Classic
examples include paper notes stuck to the edge of computer monitors.
Your choice of the types of data that you will collect should be guided by the goal
of using multiple sources that address your questions from different perspectives.
Sara's case study took this approach, combining interviews about early technology
use, demonstrations of various physical and software artifacts, and speculation about
desired designs.
By using your research goals to guide a careful selection of data sources and spe-
cific questions, you will increase your chance of generating the multiple sources of
evidence that form the backbone of data triangulation. A design that makes clear and
explicit links between each of the data sources and your research questions will help
you understand which questions are addressed by multiple data sources, and which
are not. If you find that you have questions that are only represented in one of the