Case Study Research

Yin's approach is more systematic and design-oriented, emphasizing case selection logic, multiple data sources, and explicit research questions. Cases can be exploratory, descriptive, or explanatory. Yin's framework is common in applied fields (business, policy, evaluation). Stake's approach is more interpretive and emphasizes the intrinsic interest in understanding a particular case (intrinsic case study) or using a case instrumentally to understand a broader issue (instrumental case study). Stake emphasizes thick description and readers' ability to make naturalistic generalizations. Both are rigorous; Yin is more systematic and structured; Stake is more flexible and interpretive.

Single-case studies are appropriate when the case is unique, extreme, or particularly informative. Multiple-case studies (typically 2–6 cases) offer greater credibility through replication logic. A good case is: (1) bounded—has clear boundaries; (2) accessible—you can gain sufficient access to data; (3) relevant—illuminates your research question; (4) informative—is typical, extreme, or revealing. Purposively select cases rather than randomly; sampling is strategic, not probabilistic.

Duration depends on case complexity and research questions. Some case studies last weeks; others span years. Sufficient time in the field is essential to understand context, observe processes, and build relationships with participants. A minimum is typically several weeks; longer engagement (months to years) reveals change and deeper patterns. Document the time spent; short engagement may limit depth.

Triangulation means using multiple data sources (interviews, observation, documents) to verify findings and converge on truth from different angles. If interviews, observations, and documents all confirm a finding, confidence increases. Triangulation also reveals phenomena from different perspectives; each source has strengths and weaknesses. In case studies, triangulation is essential because a single source (e.g., only interviews) may provide incomplete or biased understanding. Use at least three data sources; ideally four or more.

Case studies do not produce statistical generalization (applying findings to populations). Instead, they enable transferability: readers assess whether findings apply to their context. Present cases with sufficient detail so readers understand context and can judge relevance. Write for transferability by clearly describing the case, context, and findings; readers in similar settings can decide whether your findings apply to them. This is not the same as statistical generalization but is a legitimate and valuable form of knowledge contribution.