---
name: asgard-ai-platform/grad-case-study
source: https://app.decimal.ai/s/asgard-ai-platform-grad-case-study@1/SKILL.md
source_sha256: 1dae19d534c9
---

# Case Study Research (Yin)

## Overview

Case study research is an empirical inquiry that investigates a contemporary phenomenon in depth and within its real-life context, especially when the boundaries between phenomenon and context are not clearly evident. Yin's framework provides systematic design choices — single vs. multiple cases, holistic vs. embedded analysis — and emphasizes triangulation to strengthen construct validity.

## When to Use

- Research questions are "how" or "why" questions about contemporary events
- The researcher has little or no control over behavioral events
- Contextual conditions are highly pertinent to the phenomenon of study
- Boundaries between phenomenon and context are blurred

## When NOT to Use

- When frequency or incidence data is needed (use survey or experiment)
- When context is irrelevant and can be controlled (use experiment)
- When the goal is statistical generalization to a population

## Assumptions

```
IRON LAW: Case study answers HOW and WHY questions in context — if you
need frequency or incidence data, use a survey or experiment instead.
Applying case study to "how many" or "how much" questions misuses the
methodology.
```

Key assumptions:
1. The case is a bounded system — define temporal, spatial, and conceptual boundaries
2. Multiple sources of evidence are essential for construct validity
3. Multiple cases follow replication logic (not sampling logic) — each case is an experiment, not a survey respondent
4. A case study protocol and database enhance reliability

## Methodology

### Step 1: Design the Case Study

Define the research question (how/why). Select the case type using Yin's 2x2 matrix:

| | Single Case | Multiple Case |
|---|---|---|
| **Holistic** (single unit) | Critical, unique, or revelatory case | Literal or theoretical replication |
| **Embedded** (multiple units) | Multiple units within one case | Multiple units across cases |

Develop propositions to guide data collection.

### Step 2: Collect Evidence from Multiple Sources

Gather data from at least three of six source types: documents, archival records, interviews, direct observation, participant observation, physical artifacts. Maintain a chain of evidence linking questions to data to conclusions.

### Step 3: Apply Triangulation

| Triangulation Type | Description |
|---|---|
| **Data** | Multiple data sources converge on the same finding |
| **Investigator** | Multiple researchers independently analyze the same data |
| **Theory** | Multiple theoretical perspectives applied to the same data |
| **Methodological** | Multiple methods (qual + quant) address the same question |

### Step 4: Analyze and Report

Use pattern matching, explanation building, time-series analysis, or cross-case synthesis. Report the chain of evidence transparently.

## Output Format

```markdown
## Case Study Analysis: [Context]

### Research Question
- Question: [the how/why question]
- Case type: [single/multiple] x [holistic/embedded]
- Unit of analysis: [what constitutes the "case"]

### Case Selection Rationale
| Case | Rationale | Expected Pattern |
|------|-----------|-----------------|
| [name] | [why selected] | [literal/theoretical replication] |

### Evidence Matrix
| Source Type | Data Collected | Key Findings |
|------------|---------------|--------------|
| [documents/interviews/etc.] | [description] | [finding] |

### Triangulation Results
- Convergent findings: [where sources agree]
- Divergent findings: [where sources disagree]
- Explanation: [how divergence is resolved]

### Pattern Matching
- Predicted pattern: [from propositions]
- Observed pattern: [from evidence]
- Match assessment: [strong/moderate/weak]

### Conclusions
1. [Key finding with chain of evidence]
2. [Analytical generalization — how findings extend theory]
```

## Gotchas

- Case studies generalize to THEORY (analytical generalization), not to populations (statistical generalization)
- A single-case design requires explicit justification: critical, extreme, unique, revelatory, or longitudinal
- Replication logic in multiple cases means each case must independently confirm or disconfirm a proposition
- The chain of evidence must be traceable from question to protocol to database to report
- Do NOT confuse case study with case history or case report — Yin's case study is a research strategy with formal design
- Rival explanations must be addressed explicitly, not just dismissed

## References

- Yin, R. K. (2018). *Case Study Research and Applications: Design and Methods* (6th ed.). Sage.
- Stake, R. E. (1995). *The Art of Case Study Research*. Sage.
- Eisenhardt, K. M. (1989). Building theories from case study research. *Academy of Management Review*, 14(4), 532-550.