Page 179 -

166 CHAPTER 7 Case studies

to be a bit of a contradiction in terms, but there is nothing strange about doing case-
study research with two or more cases.
To understand why you might use multiple cases when one might seem to do
just as well, we must consider one of the important goals of many instrumental case
studies: generalization. An in-depth discussion of one individual (such as Sara) is
interesting, but the real value in a study of this sort lies in generating insights that can
be applied to a broader class of design challenges. We might be pleased if Sara's case
study led to some suggestions for the design of assistive devices that would help Sara
with her daily challenges, but we would often like to go further. If the case study led
to insights that apply to many blind people, any resulting designs might be useful to
a much broader range of blind users.
If our goal is to generalize, we would ideally argue that our cases are somehow
representative. They must be similar to the members of the broader group that is the
focus of our generalization, at least in ways that are relevant to the study at hand. A
single case may or may not be representative, and we may not have any way of evalu-
ating whether or not any single case provides a basis for generalization. From the
description of her tasks and challenges, we might infer that Sara is a reasonably rep-
resentative blind college student, but we really can't say for sure. She may be more
(or less) experienced with computers than other blind college students, more (or less)
willing to try new technologies, and so on. Casting a broader net, we might wonder
if insights gained from interviewing Sara can apply to blind people of different ages
or education levels, such as working professionals or elementary school students.
Just as scientific experiments of all sorts rely upon replication to provide in-
creased confidence in observed results, case studies can use multiple cases to pro-
vide critical support for confidence in the generality of any results. Suppose another
college student had been interviewed, following the same protocol that was used with
Sara. If the observations and insights gained from the two studies were similar, we
might be more inclined to believe that these results were applicable to blind college
students in general. This use of closely comparable cases to demonstrate consistency
of results is known as literal replication (Yin, 2014).
The analogy between case studies and other scientific experiments can lead us to
another useful form of multiple-case studies. Experimentation relies upon contrasts
between situations that are similar but differ in specific, controlled ways. When these
situations are created correctly, observed differences in experimental outcome can
be attributed to the differences between the groups. Multiple-case studies might use
cases with specific differences in much the same manner. Imagine an extension of
Sara's study that involved a blind executive instead of another student. Differences
between Sara and the executive in terms of how they use technology might be due
2
to differences in their occupations. The use of comparable cases to generate results
that differ in ways that can be explained by differences between the cases is known
as theoretical replication (Yin, 2014). The International Children's Digital Library
sidebar describes a multiple-case study involving theoretical replication.

2
Differences in age and economic resources might also play a role. Strictly controlling for differences
is difficult with cases involving human participants.