Single-Case vs. Multiple-Case Study: Which Should You Use?

Is It Better to Study One Case in Depth or Compare Several Cases?

You want to investigate how universities develop policies for generative AI in assessment. One institution offers unusually detailed access to policy records, decision-makers, faculty consultations, and implementation activities. Alternatively, three universities are willing to participate, but access to each institution is more limited.

Would examining three universities produce a stronger study because it includes more cases? Or could a single, intensively investigated university provide a more informative explanation?

Researchers sometimes assume that multiple-case studies are inherently more rigorous because they include more settings. Others select a single case because it is easier to access, without explaining why that case deserves intensive investigation.

Neither assumption is sufficient. The choice between single-case and multiple-case study designs should reflect the research question, the rationale for selecting cases, the analytical contribution sought, and the resources required to investigate each case adequately.

Choose One Case for a Strong Case-Specific Rationale and Multiple Cases for Meaningful Cross-Case Inquiry

How Single-Case and Multiple-Case Designs Differ

What Is a Single-Case Study?

A single-case study investigates one bounded case as the principal object of inquiry.

The case may be an individual, institution, program, event, community, or process. The research seeks to understand its characteristics, development, relationships, or explanatory significance in depth.

For example, a researcher might investigate how one university developed and implemented a generative AI assessment policy during its first two years.

The study could draw on interviews, policy documents, meeting records, observations, and institutional data. The number of participants or evidence sources does not change the fact that the research concerns one overarching case.

A single-case design may be particularly useful when the case is unusually informative, provides access to a previously inaccessible process, or offers an opportunity to examine a theoretical proposition.

However, the researcher must explain why the selected case is appropriate rather than treating convenience as a sufficient justification.

What Is a Multiple-Case Study?

A multiple-case study investigates two or more distinct bounded cases within one coherent research design.

Each case is examined in sufficient depth to develop a meaningful understanding of its context and evidence. The researcher then considers relationships, similarities, differences, or theoretical patterns across cases.

For example, a study might investigate how three universities with different governance arrangements developed policies for AI-assisted assessment.

Each university's policy-development process would constitute a separate case. Researchers could first reconstruct each process and then examine how governance structures, faculty participation, and institutional priorities shaped the different outcomes.

The defining feature is not simply collecting data from several institutions. The research must treat them as cases and use an explicit analytical logic connecting their investigation.

What Is the Main Difference Between Single-Case and Multiple-Case Designs?

The principal difference concerns the number of bounded cases and the analytical purpose served by investigating them.

These are typical differences rather than fixed rules. A single-case study may contain embedded analytical units, and a multiple-case study may devote considerable attention to the particularity of each case.

When Is a Single-Case Study Particularly Appropriate?

A single-case study is defensible when the selected case offers a clear analytical opportunity that would not necessarily be improved by adding more cases.

In Yin's methodological tradition, rationales for single-case designs include critical, unusual, common, revelatory, and longitudinal cases. The terminology and classifications vary across editions and methodological traditions.

A critical case may provide an opportunity to examine a theoretical proposition under especially informative conditions.

An unusual case may reveal processes or relationships that are difficult to observe in more ordinary settings.

A common or typical case may help researchers understand how a phenomenon operates under relatively ordinary conditions.

A revelatory case may provide access to a phenomenon that researchers previously had limited opportunity to investigate.

A longitudinal single-case study may examine meaningful changes within the same case across time.

These rationales should be supported by evidence about the case's characteristics. Simply declaring a case unique or critical does not establish its significance.

When Is a Multiple-Case Study Particularly Appropriate?

A multiple-case study becomes especially useful when the research question requires understanding similarities or differences across bounded settings.

For example, a researcher might investigate why AI assessment policies were implemented successfully in some universities but encountered resistance in others.

Examining institutions with different governance structures, levels of faculty participation, or implementation arrangements could help evaluate competing explanations.

Multiple cases may also be useful when researchers expect a proposed mechanism to operate similarly across settings or differently under theoretically specified conditions.

However, selecting cases simply because they are available may produce a collection of unrelated descriptions.

The researcher should explain why each case contributes to the comparative inquiry and how the resulting evidence will be analyzed across cases.

What Is Replication Logic in Multiple-Case Research?

Yin's approach emphasizes replication logic as a rationale for selecting and analyzing multiple cases.

Rather than treating cases as interchangeable members of a statistical sample, the researcher selects cases according to expectations about what their findings may reveal.

Literal replication involves selecting cases expected to produce similar findings for theoretically relevant reasons.

Theoretical replication involves selecting cases expected to produce contrasting findings for reasons anticipated by the explanatory framework.

Suppose researchers hypothesize that sustained faculty consultation supports acceptance of institutional AI assessment policies.

They might select universities with comparable consultation arrangements and expect similar implementation patterns. They could also select universities with different governance structures and anticipate theoretically meaningful contrasts.

The objective is to evaluate the proposed explanation across cases, not to obtain a statistically representative sample of universities.

Replication logic is one influential approach to multiple-case design, but it is not the only legitimate comparative tradition. Interpretive and constructivist case study researchers may prioritize contextual understanding, particularity, and cross-case interpretation without adopting Yin's formal replication framework.

Is Replication Logic the Same as Statistical Sampling?

No. Statistical sampling and case-study replication address different inferential purposes.

Probability sampling selects units according to a known sampling mechanism, often to support population estimation and associated uncertainty calculations.

Replication logic selects cases according to theoretically meaningful expectations about similarities and contrasts.

For example, selecting four universities because they represent different governance arrangements does not make those universities statistically representative of all higher education institutions.

Likewise, finding similar patterns across four cases does not establish the population prevalence of those patterns.

Multiple-case studies may contribute to analytical generalization by evaluating whether a proposed explanation remains informative across contexts.

Researchers should distinguish theoretical generalization in case study research from statistical generalization to a population.

Does a Single-Case Study Have to Be Holistic?

No. A single-case study may be holistic or embedded.

A holistic design investigates the overarching case primarily as one analytical unit.

An embedded design includes identifiable subunits within that case.

For example, a researcher might investigate one university's AI governance process while systematically examining how four faculties implemented the same policy.

The university-wide implementation process would remain the overarching case, while the faculties would serve as embedded units.

This does not automatically become a four-case study. The classification depends on how the units are defined and the level at which the research draws its principal conclusions.

Researchers should clarify the unit of analysis and the relationship among embedded units before naming the design.

Can a Multiple-Case Study Contain Embedded Units?

Yes. Each case in a multiple-case design may contain embedded analytical units.

For example, a study could examine three universities, with two faculties investigated systematically within each institution.

The design would contain three overarching university cases and faculty-level embedded units.

Such a structure may allow researchers to examine variation within institutions and compare patterns across institutions.

However, it introduces analytical complexity. The researcher must distinguish faculty-level findings from university-level conclusions and avoid treating embedded units as independent cases without justification.

Data collection and analysis should preserve the hierarchical relationships among units.

How Many Cases Are Enough for a Multiple-Case Study?

There is no universally accepted number of cases that makes a multiple-case study adequate.

The appropriate number depends on the research question, case-selection rationale, expected variation, analytical framework, depth of evidence, and available resources.

Two carefully selected cases may be sufficient for a focused comparison when the contrast is theoretically meaningful.

More cases may be needed when the research aims to examine several contextual configurations or investigate whether a proposed explanation holds across varied conditions.

However, adding cases without adequate evidence may weaken the study by reducing analytical depth.

Researchers should not automatically apply participant sample-size rules, statistical power formulas, or qualitative saturation claims to determine the number of cases.

Case adequacy should be justified through the analytical purpose and the evidence needed to support it.

Does a Multiple-Case Study Automatically Provide Stronger Evidence?

No. Multiple cases can strengthen an argument when they allow meaningful comparison or examination of theoretically expected patterns.

However, a multiple-case study may be weak if cases are poorly defined, selected without a rationale, or investigated superficially.

For example, conducting three interviews at each of five universities may provide less contextual understanding than a carefully designed single-case study with extensive documentary, observational, and interview evidence.

The number of cases does not compensate for inadequate evidence within each case.

Conversely, a single-case study may provide rich explanation but offer fewer opportunities to examine whether the same interpretation applies under different contextual conditions.

The relevant trade-off is between analytical depth, comparative leverage, and the research purpose.

How Should Cases Be Selected in a Multiple-Case Study?

Case selection should follow an explicit analytical rationale.

Researchers may select cases that are similar in important respects but differ in a theoretically relevant condition, or cases that vary across contexts expected to influence the phenomenon.

For example, a study of AI policy implementation might select universities with centralized, decentralized, and collaborative governance arrangements.

The researcher should explain why these arrangements are relevant and what patterns the comparison is intended to investigate.

Selection based only on accessibility may be necessary in some circumstances, but the resulting limitations should be acknowledged.

The principles of selecting cases without relying only on convenience apply to both single-case and multiple-case designs.

How Does Data Collection Differ Between the Designs?

A single-case study may permit intensive engagement with one setting, including repeated interviews, prolonged observation, and detailed examination of institutional records.

A multiple-case study requires a plan for collecting sufficiently comparable evidence across cases while remaining sensitive to each case's distinctive context.

For example, researchers might use a common set of guiding questions and evidence categories across universities while allowing additional interviews or documents to address case-specific developments.

Strictly identical data collection is not always necessary or desirable. However, unexplained differences in evidence availability may complicate cross-case comparison.

Researchers should distinguish substantive differences between cases from differences arising because one case was investigated more thoroughly than another.

How Should Cross-Case Analysis Be Conducted?

Cross-case analysis should build on adequate within-case understanding.

A common approach is to develop an individual case account before examining patterns across cases.

For example, researchers might reconstruct each university's policy timeline, governance arrangements, faculty consultation process, and implementation outcomes.

They could then compare how different governance conditions relate to implementation decisions, while examining contradictory evidence and alternative explanations.

Cross-case matrices, thematic comparisons, chronological analyses, and theoretically informed pattern matching may be useful depending on the research tradition.

However, cross-case analysis should not reduce complex cases to simplistic labels merely to make comparison easier.

The researcher should preserve contextual differences that materially affect interpretation.

When Does a Single-Case Study Offer Greater Analytical Depth?

A single-case design may be preferable when the phenomenon involves a complex sequence of decisions, relationships, or institutional changes that require sustained investigation.

For example, studying how one university negotiated a controversial AI assessment policy may require examining several rounds of consultation, governance disputes, revisions, and implementation decisions.

Such a process may be difficult to reconstruct adequately across several institutions within the same research resources.

Intensive within-case investigation can help identify turning points, competing interpretations, and evidence relevant to proposed mechanisms.

However, greater depth should not be assumed merely because only one case was selected. It must be demonstrated through the quality and adequacy of the investigation.

When Does a Multiple-Case Study Offer Greater Analytical Leverage?

Multiple cases may provide stronger analytical leverage when the research question concerns variation across contexts or the conditions under which a process unfolds differently.

Suppose researchers want to understand why AI governance policies receive faculty support in some universities but encounter resistance in others.

Examining institutions with contrasting consultation processes may help assess whether faculty involvement is relevant to the observed differences.

Nevertheless, cross-case patterns do not automatically establish causation. Other institutional characteristics may explain the contrasts.

Researchers should investigate alternative explanations and avoid interpreting a small number of case comparisons as if they were controlled experiments.

What Are the Main Practical Trade-Offs?

Multiple-case studies can offer comparative opportunities, but they generally require additional institutional access, coordination, analysis, and reporting. Their advantages depend on the quality of the comparison rather than the number of cases alone.

When Should You Reconsider the Initial Design Choice?

Researchers should reconsider the design when the selected case or cases no longer provide the evidence needed to answer the research question.

For example, a proposed single-case study may reveal that the central research question actually concerns why implementation differs across institutions. A multiple-case design might then be more appropriate.

Conversely, a proposed multiple-case study may become infeasible when access to institutional records is limited and only superficial interviews can be conducted.

Refining the design may be preferable to maintaining an ambitious comparison that lacks adequate evidence.

Substantial changes should be documented and reflected consistently in the research questions, sampling, analytical procedures, and conclusions.

Choosing Between One University and Three Universities for an AI Governance Study

Suppose the researcher selects all three universities solely because reviewers may regard multiple cases as more rigorous. That decision would be difficult to defend if the additional institutions contribute only superficial evidence.

Alternatively, choosing University A solely because it is nearby would also be insufficient without explaining its analytical significance.

The defensible choice depends on the question, the value of the selected cases, and the depth of investigation that can realistically be achieved.

Common Misconceptions About Single-Case and Multiple-Case Studies

How to Choose the Number of Cases for Your Research

Begin with the analytical purpose rather than a target number of cases. Determine whether the research question can be answered through intensive investigation of one bounded case or whether comparison across cases is essential.

Then evaluate the case-selection rationale, evidence requirements, and feasibility of achieving adequate depth.

Before committing to a multiple-case design, write down what each additional case is expected to contribute. If the justification is simply that more cases look more convincing, the comparative rationale needs further development.

For a single-case design, explain why the selected case offers meaningful theoretical, contextual, or evidentiary value.

The most defensible design is the one that produces sufficient evidence for the intended analytical contribution, not necessarily the one with the larger number of cases.

Before Choosing a Single-Case or Multiple-Case Design, Verify These Conditions

Frequently Asked Questions About Single-Case and Multiple-Case Studies

Choose the Number of Cases According to the Explanation You Need to Develop

How to Cite This Guide

This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.