How to Build Theory from Case Studies
Based on Building Theories from Case Study Research by Kathleen M Eisenhardt 1989
In the world of organizational research, developing new theory is a central but often mysterious activity. Traditionally, researchers have relied on combining previous literature, common sense, and experience to dream up theories. However, theories created this way often suffer from a lack of connection to the real world their tie to actual data is tenuous.
In her landmark 1989 paper, Kathleen Eisenhardt provides a rigorous roadmap for a different approach Inductive Theory Building. Instead of starting with a hypothesis and testing it, this method immerses the researcher in real world case studies to let the theory emerge directly from the evidence. The result is theory that is novel, testable, and empirically valid.
This article details the complete 8 step process outlined by Eisenhardt, serving as a comprehensive guide for researchers venturing into the black box of inductive analysis.
The Background
What is this Method
This approach is a specific type of research strategy that focuses on understanding the dynamics present within single settings. It sits at the intersection of three different traditions.
Qualitative Data Analysis
Techniques for handling descriptive data for example Miles and Huberman
Grounded Theory
The comparative method of developing theory solely from evidence for example Glaser and Strauss
Case Study Design
The logic of using cases as experiments for example Yin
Eisenhardt synthesizes these into a positivist approach meaning the goal is not just to describe a case but to create generalizable testable theory.
The 8 Step Roadmap
The process is highly iterative. Researchers constantly move backward and forward between steps but the overall trajectory follows this distinct path.
Step 1 Getting Started
Before entering the field, the researcher must have a plan. It is impossible to start with a clean theoretical slate so it is better to be intentional.
Define the Research Question
You must identify a broad research focus. Without this, investigators can easily become overwhelmed by the sheer volume of data.
A Priori Constructs
It is helpful to identify some tentative variables constructs to measure at the start. This allows for better measurement grounding.
No Hypotheses
Crucially, you should not think about specific relationships between these variables yet. Retaining theoretical flexibility allows you to spot serendipitous findings that you did not expect.
Step 2 Selecting Cases
Unlike survey research where you pick a random sample to represent a population case study research relies on theoretical sampling.
The Logic
You choose cases for theoretical reasons to replicate a finding extend a theory or fill a conceptual category.
Polar Types
A powerful strategy is to choose polar types extreme examples such as one highly successful organization and one failing one. This makes the phenomenon you are studying transparently observable.
Controlled Variation
Selecting cases from a specific population for example large British corporations helps control extraneous variation making the limits of the findings clear.
Step 3 Crafting Instruments and Protocols
Good theory building requires robust evidence. Eisenhardt advocates for a triangulation approach combining multiple methods to view the phenomenon from different angles.
Qualitative and Quantitative
Case studies can and should use both. Quantitative data numbers can keep researchers from being misled by vivid but false impressions while qualitative data stories explain the why behind relationships.
Multiple Investigators
Using a research team rather than a solo investigator enhances creativity and confidence. One person might interview while another observes capturing different perspectives and preventing premature closure on a theory.
Step 4 Entering the Field
A striking feature of this method is the overlap of data analysis with data collection. You do not wait until the end to start thinking.
Field Notes
Researchers should keep a running commentary of impressions constantly asking what am I learning.
Flexible Data Collection
This overlap allows for controlled opportunism. If a new theme emerges during an interview researchers are free to alter their instruments adding questions or even new data sources in real time to probe that theme in future cases.
Step 5 Analyzing Data Within Case
The sheer volume of data in case research can lead to death by data asphyxiation. To cope researchers must first analyze each case in isolation.
Write Ups
Investigators should create detailed write ups for each site. These are often pure descriptions but they are central to generating insight.
Intimate Familiarity
The goal is to become intimately familiar with each case as a stand alone entity before trying to generalize across them. This helps unique patterns emerge.
Step 6 Searching for Cross Case Patterns
People are notoriously poor processors of information. We tend to leap to conclusions based on limited data or vivid anecdotes. To counter this researchers must use structured tactics to look at the data through divergent lenses.
Select Pairs
List similarities and differences between pairs of cases. This forces you to see subtle differences you might otherwise miss.
Divide by Source
Look at the data from one source for example interviews separately from another for example questionnaires to see if they corroborate each other.
Dimensions
Select categories for example high versus low performance and look for within group similarities and intergroup differences.
Step 7 Shaping Hypotheses
This is where the theory solidifies. It involves two distinct parts.
Sharpening Constructs
You must refine your definitions and build evidence for them. For example if you claim CEO Power is a key variable you must show a table of evidence quotes scores observations that proves you measured it accurately in every case.
Verifying Relationships Replication Logic
You treat each case as a distinct experiment. If a hypothesis holds true in Case A does it replicate in Case B. If it works in multiple cases you have replication. If it fails in one you must modify the theory to explain why. This confirms the internal validity of the theory.
Step 8 Enfolding Literature
Inductive researchers must not ignore existing literature. They must explicitly compare their new concepts with what has been written before.
Conflicting Literature
If your findings contradict past research you must explain why. This forces you into a more creative framebreaking mode of thinking and strengthens your theory.
Similar Literature
If your findings agree with similar contexts it strengthens confidence that your results are generalizable.
Reaching Closure When to Stop
Ideally the process stops when theoretical saturation is reached. This is the point where analyzing one more case provides minimal marginal improvement you are learning nothing new.
The Magic Number
While there is no strict rule a number between 4 and 10 cases usually works best. Fewer than 4 makes it hard to generate complex theory more than 10 becomes too complex to manage.
Evaluating the Outcome
How do we know if the resulting theory is good Eisenhardt suggests three main criteria for evaluation.
Parsimonious
The theory should be simple and stripped of unnecessary detail.
Testable
It must include constructs that are measurable and hypotheses that can be falsified.
Logically Coherent
The internal narrative must make sense.
Strengths and Weaknesses
Strengths
This method is excellent for generating novel theory because it is not tied to old perspectives. It produces theory that is empirically valid because it is built directly from real world evidence.
Weaknesses
The intensive use of data can produce theory that is overly complex trying to explain everything or idiosyncratic too narrow to apply generally.
Conclusion
Eisenhardts framework offers a powerful alternative to traditional deductive research. It is most appropriate in new topic areas where existing theory is inadequate. By following this architects blueprint from the cornerstone of the research question to the final inspection of the literature researchers can close the gap between data and theory creating insights that are both profound and practical.
Leave a Reply