Page 190 -

7.9 Analysis and interpretation 177

differences between individual units of analysis or cases. That's fine. The point is
to facilitate understanding of the differences and similarities between the individual
elements.
The multiplicity of data sources used in case study research can support data
source triangulation (Chapter 11). If you can use artifact, interview, and observa-
tion data together to provide a consistent interpretation of certain aspects of the case
under examination, you will have a strong argument in favor of the validity of your
interpretation.
Appropriate data displays can prove invaluable in this process. If you have mul-
tiple units of analysis that can be described in many ways, you may create a matrix
display (Miles and Huberman, 1994) that lays out the data in a tabular format. With
one unit of analysis per row and a specific aspect of the analysis in each column,
these displays can easily be used to understand an individual unit (reading along a
row) or to compare some aspect of each unit (reading down a column), see Table 7.1.
The relationship between the theory behind the case study design and the anal-
ysis of Sara's individual tasks provides an opportunity for the use of an impor-
tant case study analysis technique—pattern matching. In this approach, case study
observations can be matched to predictions from the theory behind the design.
Matches between the observations and the theory provide support for the theory
(Yin, 2014). The specific pattern that is being matched in Sara's study can be found
in the researchers' discussion of their study: they initially believed that Sara would
use a wide range of technological approaches and creative workarounds to solving
problems, and that these practices would help provide a greater understanding of
factors influencing the success or failure of tool designs. The description of each
task in terms of the situation that led to the difficulty and the characteristics of the
individual workarounds allowed each task to be matched directly to the theoretically
proposed model.
A final level of analysis takes the comparisons between the units or cases and
combines them to develop a model or framework that communicates the results of
your case study and the over-arching themes that emerged from your analysis. As
you analyze the individual pieces and their relationships, you may identify higher-
level patterns, common concerns, or recurring ideas that may help explain, catego-
rize, or organize your results. These explanations might cut across individual units
of analysis or multiple cases, forming the basis of a case description (Yin, 2014),
which might organize your case study into specific areas of interest. In Sara's case
study, the researchers identified several criteria that technologies must meet to sat-
isfy her needs, including efficiency, portability, distinguishability of similar items,
and suitability for socially appropriate use in a sighted community (Shinohara and
Tenenberg, 2007). As always, you should be very careful to consider rival explana-
tions (see Chapter 11).
Although case studies may rely heavily on qualitative data, quantitative
data is often vitally important. A study of massive, multiplayer, online games
(Ducheneaut et al., 2007) used quantitative analysis to address questions left un-
resolved in the qualitative analysis. By defining measures of activity such as the