Kathleen Eisenhardt on Theory-Building from Case Studies 107.0

Innovation pertains to new things by definition. As such, typical methods of research like statistical hypothesis testing using random samples, which data are available now and thus come from history, cannot be used without qualification.

Kathleen Eisenhardt’s Building Theories from Case Study Research, 1989, Academy of Management Review, an inductive and iterative approach using cases by ‘theoretical sampling’, is therefore a valuable tool in the research on innovation, itself.

“This paper describes the process of inducting theory using case studies – from specifying the research questions to reaching closure.”

Eisenhardt defines the case study as “research strategy [that] focuses on understanding the dynamics present in single settings.” Eisenhardt’s clinical case studies may involve one or multiple cases in comparison.

The table below paraphrases Eisenhardt’s as a linear process but one must read the discussion in the original paper at Vol 14, Academy of Management Review to truly appreciate how new grounded theory is validated. Just following the steps in a line below will not work.

The good Professor describes the process as starting from traditional problem definition and construct validation similar to hypothesis testing research.

Then the Eisenhardt process segues into a priori specification of constructs and new steps like triangulation of multiple investigators, within-case and cross-case analyses, and the role of existing literature that are unique to her clinical case method.

“The resultant theory is often novel, testable and empirically valid,” the good Professor writes.

The general approach of inductive logic used in the Eisenhardt process is often criticized in basic scientific research and in philosophy as not leading to useful conclusion because of the bias from the selection of the cases themselves.

For me, this criticism also applies in statistical sampling when researchers do not identify the right population and use the incorrect plans and methods.

For social science research, I believe the resultant theory from the Eisenhardt process is grounded enough for application.

The process does indicate how cases and how many are to be chosen, ie by ‘theoretical sampling’ and as many to achieve ‘theoretical saturation’ without inundating the researcher with data.

The method seeks to correct for the bias by (a) constant iteration between the steps and (b) constant validation between theory and data and (c) challenging findings with existing literature.

The Eisenhardt process is very suitable “in new research areas or research areas where the existing theory seems inadequate.”

In new theory setting, there is also the reminder that the original research question should be written as open ended but directional. This allows some leeway for the iterative process to fine tune this question itself and let the data yield the theory.

Eisenhardt notes many situations when the process resulted in changes but still yielded verifiable and generalizable theory.

When the hypothesis has been shaped as above, another screening and possibly iterating step is done by comparison to similar and conflicting literature.

Finally, closure is attained when further iteration yields minimal improvements. And like any new theory, the hypothesis must also pass the tests of parsimony, repeatability and generalizability.

At the Asian Institute of Management, we use the case method to teach management principles. Teaching cases have been used as inputs into the Eisenhardt process but only if written and formatted with consideration to within-case and across case analysis required in the process; most teaching cases are not.

As inspired by Professor Patarapong’s paper (See my Post #103 – Innovation as Intensive Learning and Emulating the East Asian NIEs) on the inability of current international innovation surveys to yield the correct status for latecomer countries, I intend to try out the Eisenhardt process to enlighten such issue.

Maybe, with suitable insight, I can add to the knowledge bank of the world in this ‘long-tail’ area of innovation research.