“Case study” is one of the most widely used terms in research, and one of the most loosely used. Some students call any study of a single organisation a case study. Others use the term for a teaching example or a short illustration. In research methodology, however, a case study is a specific research design with its own logic, strengths, limitations and quality criteria.
If you are considering a case study for your thesis, you need to be able to explain what your “case” is, why a case study design fits your research question, how you selected your case or cases, what data you will collect, and how you will ensure rigour. This post covers all of these points.
What is a case study research design?
A case study is an in-depth investigation of a particular case, or a small number of cases, within its real-life context. Robert Yin, one of the most cited authors on case study methodology, describes it as an empirical inquiry that investigates a contemporary phenomenon in depth and within its real-world context, especially when the boundaries between the phenomenon and its context are not clearly evident.
Other influential writers, such as Robert Stake and Sharan Merriam, emphasise the case as a “bounded system” and focus on understanding the particular case in its complexity.
Key features of case study research include:
- Depth: detailed understanding of one or a few cases rather than broad coverage of many.
- Context: the case is studied in its natural setting, and context is treated as important, not as noise.
- Multiple sources of evidence: interviews, documents, observations, archival records, surveys or artefacts.
- Boundaries: the case is clearly defined in terms of who, what, where and when.
What counts as a “case”?
A case can be almost any bounded unit, for example:
- An individual, such as a school principal leading a turnaround.
- A group or team, such as a hospital ward adopting a new protocol.
- An organisation, such as a small manufacturing firm introducing automation.
- A programme or policy, such as a national scholarship scheme.
- An event or process, such as a merger or a community's response to a flood.
- A community or place.
The crucial step is bounding the case. You need to state clearly what is included and excluded: which organisation, which time period, which people and which activities. Without clear boundaries, a case study can expand endlessly and lose focus.
When should you use a case study?
A case study design is especially appropriate when:
- Your research question asks “how” or “why”. Yin notes that case studies are well suited to explanatory and exploratory questions about processes and mechanisms, such as “How did this organisation implement sustainability reporting?” or “Why did this policy succeed in one district but not another?”
- You have little or no control over events. Unlike experiments, case studies examine situations as they occur naturally.
- The focus is on a contemporary phenomenon in a real-life context (although historical case studies also exist).
- Context is essential to understanding. If you cannot separate the phenomenon from its setting, a case study lets you study both together.
- You need rich, holistic understanding rather than statistical generalisation.
A case study may be less suitable if your question asks “how many” or “how much” across a population, or if you need to test causal effects under controlled conditions. In those situations, surveys or experiments may fit better.
Types of case study
By purpose
- Exploratory: investigates a phenomenon that is not well understood, often to generate ideas or hypotheses.
- Descriptive: provides a detailed account of a phenomenon in context.
- Explanatory: seeks to explain how or why something happened, often by tracing processes or testing rival explanations.
By the role of the case (Stake)
- Intrinsic: the case itself is of primary interest, for example because it is unique.
- Instrumental: the case is used to understand a broader issue or theory.
- Collective: several cases are studied together to understand a phenomenon.
By number of cases (Yin)
- Single-case design: appropriate when the case is critical (testing a theory), unusual or extreme, typical (representative of many others), revelatory (previously inaccessible), or longitudinal (studied over time).
- Multiple-case design: two or more cases, often chosen to allow comparison. Yin describes this in terms of replication logic: cases are selected either because they are expected to show similar results (literal replication) or contrasting results for predictable reasons (theoretical replication).
Holistic vs. embedded
A holistic case study examines the case as a whole. An embedded case study includes sub-units within the case, such as several departments within one organisation, each analysed and then integrated.
Selecting cases
Case selection in case study research is purposive, not random. Choose cases because they can best inform your research question. Consider:
- Is the case information-rich?
- Will you have sufficient access to people, documents and sites?
- For multiple cases, what variation or similarity do you need?
- Are there practical or ethical constraints, such as confidentiality in a small organisation?
Explain your case selection criteria clearly in your methodology. Examiners will often ask, “Why this case?”
Data collection in case studies
A strength of case study research is the use of multiple sources of evidence. Common sources include:
- Interviews with people involved in the case.
- Documents, such as reports, policies, minutes, emails and websites.
- Archival records, such as databases, performance data or historical files.
- Direct observation of meetings, work practices or events.
- Participant observation, where the researcher takes part in activities.
- Physical artefacts, such as tools, products or spaces.
- Surveys, sometimes used within a case to add quantitative data.
Using multiple sources allows triangulation: checking whether different sources lead to the same conclusions. When sources converge, findings become more credible. When they diverge, the differences themselves can be informative.
Analysing case study data
Analysis depends on your approach, but common strategies include:
- Within-case analysis: building a detailed description and explanation of each case.
- Cross-case analysis: comparing cases to identify patterns, similarities and differences.
- Pattern matching: comparing observed patterns with patterns predicted by theory.
- Explanation building: developing an explanation of how or why something happened, refined through the evidence.
- Time-series or process analysis: tracing events over time.
- Thematic analysis of interview and document data.
Keep a case study database, an organised collection of all your data, notes and analytic memos, separate from your final report. This supports transparency and reliability.
Ensuring quality and rigour
Yin describes four tests for case study quality, adapted from quantitative research traditions:
- Construct validity: use multiple sources of evidence, establish a chain of evidence, and have key informants review draft case reports.
- Internal validity (for explanatory studies): use pattern matching, explanation building, and address rival explanations.
- External validity: use theory in single-case studies and replication logic in multiple-case studies to support analytic generalisation.
- Reliability: use a case study protocol and develop a case study database.
Researchers working in more interpretivist traditions often use alternative criteria such as credibility, transferability, dependability and confirmability, drawing on Lincoln and Guba. Choose criteria that match your philosophical position and explain them.
Generalisation: the common criticism
The most frequent criticism of case studies is: “How can you generalise from one case?” The answer is that case studies aim for analytic generalisation rather than statistical generalisation. Findings are generalised to theory, not to a population. A well-conducted case study can support, refine, challenge or extend theoretical propositions, which can then be tested in other settings.
Rich description of the context also allows readers to judge transferability: whether findings may apply to other situations they know.
A worked example: planning a multiple-case study
Suppose your research question is: “How do small and medium-sized manufacturers in Southeast Asia adopt Industry 4.0 technologies, and why do some adopt faster than others?”
- Why a case study? The question asks “how” and “why”, adoption happens in a real business context you cannot control, and organisational context (leadership, skills, supply chains) is likely to be central to the explanation.
- Defining the case: each case is one manufacturing firm with 50–250 employees, and the time boundary is the period from the first adoption decision to the present.
- Number of cases: four firms, selected using theoretical replication logic: two firms known as fast adopters and two known as slow adopters, similar in size and sector.
- Data sources: interviews with owners, production managers and technicians; internal documents such as investment proposals; site observations of production lines; and publicly available information.
- Analysis: first, a detailed within-case narrative for each firm; second, cross-case comparison to identify factors that differ between fast and slow adopters; third, comparison with existing technology adoption theories.
- Quality: a case study protocol guides data collection across firms; a case database stores all materials; key informants review draft case descriptions for factual accuracy; rival explanations, such as access to government grants, are explicitly examined.
This kind of plan shows examiners that the case study design is deliberate, structured and linked to the research question, rather than a convenient label for “I studied a few companies”.
Writing a case study protocol
A case study protocol is a document that guides your fieldwork, especially important in multiple-case studies. It usually includes an overview of the study and research questions, field procedures (access, scheduling, ethics), the questions you need to answer in each case (for yourself, not only for interviewees), and an outline of the case report. A protocol increases consistency and reliability, and it is an excellent appendix for your thesis.
Common mistakes to avoid
- Failing to define and bound the case clearly.
- Choosing a case only because it is convenient, without justification.
- Relying on a single data source.
- Producing a long description with little analysis or link to theory.
- Claiming statistical generalisation from a small number of cases.
- Not addressing rival explanations in explanatory studies.
Final thoughts
A case study design is a powerful choice when you want to understand how or why something happens in its real-world context. It demands clear case boundaries, a strong rationale for case selection, multiple sources of evidence, systematic analysis and careful attention to rigour. Used well, it can produce deep insights and meaningful contributions to theory. Used loosely, it risks becoming an unfocused description. Plan carefully, document everything, and link your case back to your research questions and theory.
Are you planning a single or multiple case study? Share your research question in the comments and we can discuss your design.