Case study - Designing and Selecting Cases

Types of Case Study Research Designs and Case Selection Strategies Introduction Case studies are a powerful research method, but they can serve very different purposes. Understanding the different types of case study designs helps you recognize what a particular case study aims to accomplish—whether it's describing a phenomenon, testing a theory, or generating new ideas. Equally important is knowing how to select cases strategically. A common misconception is that case selection should be random, but in small-N research, this approach often backfires. Instead, researchers should deliberately choose cases that will provide the most insight for their research question. Types of Case Study Research Designs Case studies can be classified by their primary purpose. Each design serves a different role in the research process. Atheoretical (Configurative Idiographic) Design An atheoretical case study aims to richly describe a specific case in depth without explicitly contributing to broader theory development. The goal is to understand the case on its own terms, capturing its complexity and uniqueness. Think of this as "telling the full story" of a single case. A researcher might conduct an atheoretical case study of a particular historical event, organization, or community to create a comprehensive narrative account. This design is valuable when little is known about a phenomenon and descriptive understanding is the primary goal. Interpretative (Disciplined Configurative) Design Interpretative case studies do something more focused: they use established theories to explain a specific case in depth. Rather than generating new theory, the researcher applies existing frameworks to understand why a particular case unfolded the way it did. For example, a researcher might use organizational theory to interpret why a specific company failed despite having strong market conditions. The existing theory provides a lens for understanding the case. Hypothesis-Generating (Heuristic) Design Hypothesis-generating case studies take the opposite approach: they work inductively to identify new variables, hypotheses, causal mechanisms, and causal pathways that weren't previously recognized. This design is exploratory—it's about discovery. A researcher might deeply examine an unusual case and discover new factors that explain how outcomes occur. These findings then become hypotheses that can be tested in larger studies. This is particularly valuable when studying emerging phenomena that existing theories don't yet address. Theory-Testing Design Theory-testing case studies are designed to assess the validity and scope conditions of existing theories. They ask: "Does this theory actually hold up? Under what conditions does it apply?" A researcher selects a case and examines whether the causal mechanisms predicted by theory actually operate in that case. This can confirm a theory's applicability or reveal that it only works under certain conditions. Plausibility-Probe Design Plausibility-probe case studies occupy a middle ground—they evaluate the plausibility of new hypotheses and emerging theories before investing in large-scale testing. They're preliminary investigations to assess whether a new idea is worth pursuing further. Think of this as a feasibility check. A researcher examines a case to determine whether preliminary support exists for a new theoretical claim. Building-Block Study Design Building-block studies are designed to identify common patterns across multiple cases or subtypes. By examining several cases systematically, the researcher identifies which variables consistently matter and which causal pathways recur. This design bridges single-case and comparative methods, collecting enough cases to see patterns while maintaining the depth that case study research provides. Case Selection Strategies Once you've decided on a research design, the next critical choice is which cases to study. This decision directly affects what you'll learn. Why Case Selection Matters The fundamental goal of case selection is to find cases that are representative of theoretical dimensions and that provide high expected information gain. In other words, choose cases that will teach you something meaningful about your research question. A counterintuitive finding in case study methodology is this: selecting unusually informative or atypical cases often yields richer insights than choosing average or typical cases. This is because extreme or unusual cases highlight causal mechanisms more clearly—they provide contrast that makes patterns visible. The Problem with Random Selection Here's where many researchers make a critical mistake: random selection can generate unrepresentative and uninformative cases, creating serious bias when the number of observations is small. Why? Imagine a researcher studying organizations. If they randomly select cases, they might accidentally end up with a sample that doesn't capture the full range of the phenomenon. With a small sample (which is typical in case study research), this isn't just unfortunate—it seriously compromises what you can learn. Random selection optimizes for representativeness of a large population, but in small-N research, you should instead optimize for information gain. Information-Rich Cases One strategy is to deliberately choose outlier, extreme, or deviant cases that can reveal more information than typical cases. Why? Because these cases highlight unusual circumstances that clarify causal mechanisms. For example, if you're studying what causes organizations to innovate, a company that dramatically transformed itself despite unfavorable conditions is more informative than a typical successful innovator. The extreme case forces you to articulate the mechanisms that made success possible. Selection Based on Researcher Knowledge Another approach is to select cases based on deep local knowledge, allowing researchers to "soak and poke"—spending time in the field to generate nuanced explanations grounded in intimate familiarity with the context. This strategy acknowledges that researchers often have valuable tacit knowledge about which cases will be most informative. A researcher familiar with a particular industry, community, or field often knows better than random selection which cases will reveal important insights. A Formal Typology of Selection Strategies To make case selection more systematic, researchers use a set of formal categories: Typical cases exemplify a stable cross-case relationship and are representative of the larger population. These are the "textbook examples" of a phenomenon. Diverse cases contain variation on relevant independent and dependent variables, representing the full range of the population. Rather than selecting similar cases, you select cases that differ systematically to explore how different conditions produce different outcomes. Extreme cases have an extreme value on either the independent or dependent variable relative to other cases. An example would be the wealthiest versus poorest neighborhoods in a city, or the most versus least successful schools. These highlight the boundaries of a phenomenon. Deviant cases are particularly valuable—they defy existing theories and common sense, showing extreme values and contradictory causal patterns. A case that "shouldn't happen" but does is incredibly informative. For instance, if theory predicts that a country should descend into civil war but it doesn't, studying that case might reveal protective factors theory overlooked. Influential cases are central to a model or theory. For example, a regime that epitomizes a theoretical concept (like a "totalitarian state" or "liberal democracy") is influential because it defines what the category means. Most similar cases are alike on all independent variables except the one of interest to the researcher. By holding other factors constant, you isolate the effect of the variable you care about. This is a case study version of experimental control. Most different cases differ on all independent variables except the one of interest to the researcher. This strategy works in reverse: if cases that differ dramatically still show the same relationship between your variables of interest, that relationship is likely robust and important. Connecting Selection Strategy to Research Design Notice how your case selection strategy should align with your research design. If you're conducting a hypothesis-generating study, you might deliberately select deviant or extreme cases to spot new patterns. If you're doing theory-testing, most-similar or most-different cases help you isolate whether the theory's mechanisms actually operate. A building-block design might systematically select diverse cases to identify which patterns recur. The key insight is this: in case study research, strategic selection isn't a limitation—it's a strength. By deliberately choosing informative cases, you extract far richer insights than random selection could ever provide.