Case studies can be classified as designed cases or naturalistic cases, depending on whether
the situation under investigation is manipulated in any way by the researcher. Formative
research is a designed case if the researcher instantiates the theory (or model) and then
formatively evaluates the instantiation. Alternatively, it is a naturalistic case if the researcher
(a) picks an instance (or case) that was not specifically designed according to the theory but
serves the same goals and contexts as the theory, (b) analyzes the instance to see in what ways it
is consistent with the theory, what guidelines it fails to implement, and what valuable elements it
has that are not present in the theory, and (c) formatively evaluates that instance to identify how
each consistent element might be improved, whether each absent element might represent an
improvement in the instance, and whether removing the elements unique to the instance might
be detrimental. Furthermore, for naturalistic cases, the methodology varies depending on
whether the observation is done during or after the practical application. This makes three major
types of formative research studies:
• designed cases, in which the theory is intentionally instantiated (usually by the
researcher for the research,
• in vivo naturalistic cases, in which the formative evaluation of the instantiation is done
during its application, and
• post facto naturalistic cases, in which the formative evaluation of the instantiation is done after its application.
And within each of these three types, the methodology also varies depending on whether the
study is intended to develop a new design theory (one which does not yet exist) or to improve an
existing theory. Table 1 shows these variations.
Designed Case To Improve an Existing Theory
While there is often much variation from one such case study to another, the following is a
fairly typical process for conducting this type of formative research study.
1. Select a design theory. You begin by selecting an existing design theory (or model)
that you want to improve.2. Design an instance of the theory. Then you select a situation that fits within the
general class of situations to which that design theory (or model) applies, and you design a
specific application of the design theory (called a "design instance") . This instance may be a
product or a process, or most likely both. It is important that the design instance be as pure an
instance of the design theory as possible, avoiding both of the two types of weaknesses
(omission: not faithfully including an element of the theory; and commission: including an
element that is not called for by the theory). This is an issue of construct validity, and its
counterpart in experimental design is ensuring that each of the treatments is a faithful
representation of its corresponding independent-variable concept.
The design of the instance can be done either by the researcher (as participant) or by an
expert in the theory (with the researcher as observer), preferably with the help of a subjectmatter
expert (usually the teacher for the course used in the instance). In either event, it is wise
to get one or more additional experts in the theory to review the instance and ensure that it is a
faithful instance of the theory. If you find yourself or the expert in the theory having to make
decisions about which the theory offers no guidance, make special note of all such occurrences,as areas of guidance that should be added to the design theory later. It is also wise to get one, or
preferably several, additional subject-matter experts to review the instance for content accuracy.