Page 189 -

176 CHAPTER 7 Case studies

results. Ideally, a research protocol will be clear enough that it can be used by other
researchers to replicate your results.
Consider running a pilot case study. Pilot tests will help you debug your research
protocols, identifying questions that you may have initially omitted while potentially
exposing flaws in your analysis plans. For some studies, a pilot may not be possible
or desirable. If you have a unique case, this may not be possible. If your study is
exploratory, you may find that a single case will provide you with sufficient data to
generate an informative analysis.

7.9 ANALYSIS AND INTERPRETATION

As qualitative data is a key component of case study research, your analysis will use
many of the techniques and strategies discussed in Chapter 11. You should start plan-
ning your data analysis early in the process, before you collect any data. Grounded
theory, content analysis, and other techniques from Chapter 11 are commonly used
to analyze case study data.
Perhaps the largest challenge in the analysis of case study data involves the lim-
ited range of data samples. Unlike controlled quantitative experiments, which use
large numbers of participants to generate statistically significant results, case studies
rely on a few samples, which may be idiosyncratic. This may present challenges if
you are interested in building general models: how can you be confident that conclu-
sions drawn from experience with your cases generalize to others?
To some extent, these validity concerns are inherent in case study research. No
matter how carefully you choose your cases, collect your data, or conduct your anal-
ysis, your case study may lead to interpretations that are not valid or do not general-
ize to other cases. You should always keep in mind that case study results may not
generalize. Even if yours seems to point to trends that hold in all cases, you should
avoid assuming that those trends are truly general.
Careful attention to the strategies described in Chapter 11 can help increase the
rigor of your analysis and confidence in your conclusion. Triangulation, documenta-
tion of chains of evidence, and consideration of rival theories are all appropriate tools
for case study analysis.
Case study analysis generally proceeds in a bottom-up fashion, using techniques
from grounded theory to code and categorize data (see Chapter 11). In Sara's case,
the analysis might have involved examining all of her descriptions of previous inter-
actions with technology, her current approaches and speculative desires for solving
a specific task. Any conclusions that were supported by all three of these approaches
might be seen as being reasonably valid. These analyses could then be used to form
an integrated description of the unit of analysis—the specific task.
After analyzing individual units of interest, you are likely to want to push your
analysis to help you understand larger trends that describe your case as a whole and
(for multiple-case studies) can be used to support comparison of similar cases. The
goal here is not necessarily to make everything agree: there may be fundamental