The Case Study Method

Of all the methods of enquiry used by human geographers, economists, policy studies and the humanities, the Case Study method is the least explained and articulated.   Yet many students in particular, use an implicit Case Study method, using the cases to symbolise or metaphorise the phenomenon they wish to study.   It is also frequently used, implicitly or explicitly in Evaluation Studies and Policy Studies.  The following lecture, combined with the Methods Chapter from my own dissertation (to be distributed to class) will give students a sound basis to consider whether the Case Study method is suitable for their purposes, and if so, and understanding of  how it can be applied with rigour.

Much of this is taken from the excellent work of Robert K. Yin.

Why Use Case Studies?

The Case Study method is most useful when a new or contemporary phenomenon is being investigated.  One which is situated within a mass of constantly changing relations and influences.   It is used to provide understanding, not prediction.   It is one of the most difficult methods to do well, however, because it contains few hard and fast rules, and thus is not a method that can be applied by rote.   The Case Study Method (CSM) must be constructed in relationship to the theoretical principals which will be tested and its rigour, appropriateness and validity must constantly be guarded.   It is easy to apply a Case Study method carelessly and in such a way as to not prove your case, difficult to do it well.

Unlike the scientific experimental method, it has no controls and does not reduce the phenomenon studied to a set of constants and a predictable relationship between a dependent and independent variable. – because it cannot control the situation or behaviour of individuals.  However, the depth and breadth of useable data sources and rigour of construction provide understanding and context which the experimental method cannot deliver with its distance from the phenomenon under study.

Survey methods provide description, and may provide prediction, but do not provide sufficient context or detail for explanation.  The superior depth of the case study method makes it more appropriate than the breadth of coverage of a survey method or accuracy of prediction of a scientific experiment  Therefore, it is most suitable to use this method when research is being taken inside a multi-variate environment which is constantly shifting, where the researcher cannot exercise control over the phenomenon or population of interest

Case studies, like experiments, are generalizable to theoretical propositions, and not to populations or universes.

Case studies, like experiments, do not represent a ‘sample’ and are oriented towards expanding theories (analytical generalization) and not towards enumeration (statistical generalization).

Case study design is not the same as traditional ethnography or other qualitative methods.  The difference lies in the theoretical point which drives the research.  The “role of theory development, prior to the conduct of any data collection, is one point of difference between case studies and related methods such as ethnography and ‘grounded theory’.  Typically these methods deliberately avoid specifying any theoretical propositions at the outset of an inquiry”  (Yin, 27).

A case study is an empirical inquiry that investigates a contemporary phenomenon within its real life context, particularly where the boundaries are blurred between phenomenon and context.

When To Use Case Studies?

The prime determining factor regarding the use of the CSM revolves around the research questions, the type of phenomenon under study, and the type of theory-building being undertaken.

Case Studies are used when asking How? and Why? questions about contemporary phenomena over which the researcher has no control [see table below].   It also can be used to answer What? questions where exploratory research is being undertaken into a new and unexpected phenomenon.

It is used to study phenomena which are recursive, iterative, never-fixed and emergent.  In other words, it is most useful for studying human relations, power, politics and society.

Case studies are useful for only a small range of purposes.   The following table illustrates the various types of research methods and where they can be applied:

Source COSMOS Corporation (cited in Yin)

Case Study as a Comprehensive Research Strategy

The chief advantage of CSM is its comprehensiveness.   Given a rigorous application, the CSM offers access to the most comprehensive set of data of any method.

CS inquiry offers
-more variables of interest than data points.
-multiple sources of evidence, using triangulation
-is best when theory-driven (especially in terms of data collection and analysis)

Qualitative vs. quantitative?

The Case Study method sits entirely within neither camp.   It will use statistical data and sampling data if it is appropriate to the research question, yet it will also use qualitative data too.   What matters in the CSM is the evidence, not the research methodology used to produce it.   Is the evidence relevant to theory-building?  If so, it’s use is appropriate under the CSM.

Inductive vs. deductive

The Case Study Method is both inductive and deductive.  Being theory-driven, it begins inductively – a theoretical proposition allows the research to induce situations where it might be found, and what it might look like.   Once the Case Study Method has been begun, however, the depth of information means that the researcher can deduce back to generalised principles.   As a result, the method is circular or cyclical in that it requires the researcher to keep a dialogue constantly going between the inductive and the deductive.   The strength of this, if done well, is that it assures that the research undertaken both directly addresses and validates/improves the theoretical construct being undertaken.

How To Use Case Study Methods?

The actual choice of data sources, cases and analysis depends upon the theoretical questions being asked.  The first consideration of CSM must be an interrogation of the appropriateness of each data source, and a theoretical rigour to consider confounding cases.

Designing Case Studies

Consider the following:
-a study’s questions
-its propositions
-the units of analysis
-the logical links between data and propositions
-the criteria for interpreting findings

At some point the researcher will need to consider how many cases to include.   This number will likely be less than they originally envisaged, as the depth of detail in any individual case will limit the number which can be effectively undertaken.     In some cases, of course, depending upon the question, only one case will be studied, as in Evaluation studies, or in enquiry into an event or entity.   However, even then, the event or entity can be broken into a number of cases, all of which have a logical and constructive relationship with the whole which remains the focus of research.

Rigor and Validity

-construct validity
-internal validity
-external validity
-reliability.

Construct Validity:   the actual model which is driving the research must be reflected in the construction of method.  Method must reflect model, not contradict it.
Internal Validity:  care must be taken to assure that all propositions relate to the research question and that choices made regarding data and analysis have an internally consistent relationship with the propositions.
External validity:  a logical and perceptible case must be able to be made for the contingent relationship between each component of the CSM and the external world which is being investigated.
Reliability:  the method does not purport to deliver predictability, but it must be reliable in the sense that data and analysis choices are made for good and transparent reasons, and that the components and structure of their relationship in the argument inspire confidence.  In particular care must be taken not to select only those cases which demonstrate a point, without considering some which do not.

Material which disagrees or disproves parts of the propositions is very important because it can set practical limits on the theoretical aspirations, and will contribute to the maintenance of both internal and external validity.

Range:
Case study design offers the opportunity to examine a wide range of evidence.  Of all the methods of social enquiry it allows the researcher the chance to determine context and sift through the detail of the empirical situation.  Yin lists at least six common sources of case study evidence:
documents,
archival records,
interviews,
direct observation,
participant observation,
physical artefacts

These can take the shape of  the press, government records, interviews with protagonists or significant local figures, direct observations, and analysing the physical landscape for evidence of the presence of the theoretical construct.  It can include secondary and primary data.  The case study method, by incorporating such a wide range of data sources, allows the investigator to address a broader range of historical, attitudinal, and behavioural issues.

Triangulation of data

The strength of the case study method lies in its ability to develop converging lines of enquiry, in other words, to engage in triangulation.

Patton (1987) discusses four types of triangulation: 
data triangulation --the triangulation of data sources ;
investigator triangulation --among different evaluators;
theory triangulation -- of perspectives on the same data set;
methodological triangulation -of methods. 
By applying data triangulation across a number of sources one can begin to build confidence in the theoretical proposition. 
Triangulation, through its careful use of multiple sources of evidence offers a way to assure construct validity, as the multiple sources provide multiple measures of the same phenomenon, ensuring confidence in both the theoretical constructs and the reliability of the interpretation.   Whilst the evidence will vary from one source to another, the logic chain tying it to the theoretical construct illustrates the construct’s presence in the case and in multiple cases the same applies across cases, demonstrating external validity by illustrating the scope and breadth of the phenomenon.

Units of Analysis and Scale

One way of structuring the use of differing data sets is to look for consistency of the presence of the phenomenon of interest across a variety of scales.   By selecting cases across a number a scales, the ubiquity and presence of a single phenomenon can be examined although it is expressed differently at each scale.   A possible schema would look at a micro, meso, and macro scale.

Building a Case Study Protocol

A Protocol is the actual designed structure which guides the research.   It needs to be complete and detailed – much like the structure of an essay or dissertation.   It should mention each step of the process from formation of theory, literature review, operationalisation of basic concepts, procedures, and selection of cases, plus the operations which will be performed on each one.

A protocol should contain a series of questions – not to the respondents, but questions to the researcher.  These can be used to guide inquiry and to assure the research remains within the bounds of the theoretical construct. They must be explicit.

The protocol must be complete enough that it can be interrogated on its own – that is, the internal, external, and construct validity can be assured before operations begin.   Of course, once research is underway it is likely to throw up information which requires a change in the structure of the protocol as new ideas or concepts enter the argument.   Again, the protocol provides a solid frame against which to work with these.

Another factor in building a Case Study project is the form of the eventual report.   A Dissertation will have one set of considerations, questions and formats.   An Evaluation Study will have others.   Objectives, aims and outcomes are germane to driving the selection of cases and data.

Depending upon the size of the project, it may be useful to employ a pilot case study to test the model and the protocol.   Confirmation or rejection of the original ideas allows the researcher to proceed with confidence in the suitability of the path they have chosen.

Analysing Case Study Evidence

Data analysis consists of examining, categorizing, tabulating or otherwise recombining the evidence to address the initial propositions of a study.

-there must some consistency between the phenomena under examination across the range of scales and cases.

Four basic analytic techniques are important to the case study method.

Pattern-matching

-can patterns be found across scales?
-are similar events, phenomena, people, etc. present throughout the case or cases?
-a pattern of non-equivalent dependent or other specific variables can be specified from the inductive consequences of the theory-building.   If the pattern is found in the empirical data, the internal consistency of the model is strengthened.
-independent variables can also be similarly arranged in a pattern to test the model against the empirical data.

Explanation-building

-this is a special type of pattern-matching
-the goal is to analyse the case study data by building an explanation about the case.  This is particularly relevant to explanatory case studies.
-this involves stipulating a set of ‘causal’ links between parts of a phenomenon and looking to see if they can be traced in the empirical data.
-this is an iterative process (and recursively inductive/deductive) where explanations are built again and again until all the empirical data and phenomena are accounted for in a set of strong links.

Time-series analysis,
-simple time-series:  this involves looking at multiple changes in a single variable over time.  It may have no clear start or ending points.  The ability to trace change over time, and relate it directly to causal factors is the strength of this type of analysis.
-complex time-series considers effects which are two-way, and not nomothetic, that is, they do not move in a single direction or dimension.  This may include alternating periods of increase or decrease, and different variables may move at different rates, or even different directions over time.  In each case the link between the specific movement must be accounted for in the theoretical propositions. These patterns can be predicted and searched for in the empirical data.
-Chronologies are a useful type of time-series.  The analytic goal is to compare the chronology with that predicted by some explanatory theory.  It will specify some of the following kinds of conditions:
-some events must always occur before other events, with reverse sequence being impossible
-some events must always be followed by other events, on a             contingency basis.
-Some events can only follow other events after a pre-specified       passage of time.
-certain time periods may be marked by classes of events that differ substantially from those of other time periods.

Programme logic models.
Stipulates a complex chain of events (pattern) over time (time series), covering dependent and independent variables.  The case studies are used to demonstrate the presence or absence of such patters over time.

Where large numbers of case studies are available, it is possible to employ a secondary analysis across a wide range of cases, demonstrating the presence of an event or other variable across them all.

Conclusion

The Case Study Method is a legitimate and very useful research methodology.  It is more useful in answering some types of questions, and some types of situations or phenomena, than others.  Specifically, the breadth and depth of allowed data sources, the constituent relationship between theory and method and the rigour of its application make the Case Study Method most appropriate for use in situations where an empirical inquiry is used to investigate a contemporary phenomenon within its real life context, particularly where the boundaries are blurred between phenomenon and context, in order to understand what, why and how something new is happening.

Sources:

Patton, M.Q.  (1987) How To Use Qualitative Methods in Evaluation.  London:       Sage.

Yin, R.K. (1994)  Case Study Research: design and methods.  2nd edition.              London: Sage.