Summary of

Case Study Research: Design and Methods

By Robert K. Yin

Summary Prepared by: Stephen G. Morrissette, November 7, 2002

Bibliography:

Yin, Robert K. (1994). Case Study Research Methods: Design and Methods (2nd ed.).
London: Sage Publications.

Key Take Aways:

1. Good cookbook/”how-to”; well grounded in theory and prior literature (great

bibliography)
2. Thorough but a reasonably quick read (170 pages)
3. The key insight for me, given my background in statistical research: The idea
that case studies are like experiments and the researcher should avoid the trap
of viewing the case as a sample (1 data point) from a population. (Section 3)
4. Case studies are a way to generalize, sampling and statistics aren’t the only way
(Section 3)
5. Case study is more than “interviews”; multiple sources of evidence (Section 6)
6. Key to reliability is “replication”; if one does more cases or another researcher
repeats your case would you find the same conclusions (Section 4)
7. Case studies are well suited to communicate research information to non-specialists
(pg 131) (Section 8)

Summary:

1. Table of Contents:
1.1. Introduction
1.2. Designing Case Studies
1.3. Conducting Case Studies: Preparing for Data Collection
1.4. Conducting Case Studies: Collecting the Evidence
1.5. Analyzing Case Study Evidence

Morrissette ID # 167952 Page 1 of 5

 1.6. Composing the Case Study Report

2. Definition of Case Study Research

2.1. What it is: A case study is like an experiment
2.2. What it is not: A case study is not a statistical sample. See section 4.
2.3. When to use it: to answer “how” or “why” questions or when studying
contemporary phenomenon within some real-life context (pg 1ff).
2.4. Ex. Use survey or analytical/quantitative research method if you want to know
“what” the outcomes of a new government program had been, use case study
method to know “how” or “why” the program worked (or didn’t work). (pg 7)
2.5. Types of case studies: exploratory, descriptive, explanatory (pg 3-5)
2.6. Case study research has been maligned as lacking rigor because critics state that
a single case offers a poor basis for generalization.

3. The paradigm of case study research is like experimental research, not statistical
sampling research
3.1. Experimental research is generalizable to theoretical propositions, not to
populations or universes
3.2. “…the case study, like the experiment, is not a “sample,” and the investigator’s
goal is to expand and generalize theories (analytic generalization) and not to
enumerate frequencies (statistical generalization).” (pg 10)
3.3. “A fatal flaw in doing case studies is to conceive of statistical generalization as
the method of generalizing results in this case.” (pg 31)
3.4. Cases are not samples; they are to be “selected as a laboratory investigator selects
topics for a new experiment. Multiple cases, in this sense, should be
considered like multiple experiments.” (pg 31)
3.5. Empirical results of the case study are compared to the previously developed
theory. If two or more cases are shown to support the theory, then replication
may be claimed. (pg 31)
3.6. “Replication, not sampling logic, for multiple-case studies.” (pg 45)
3.7. “…a few cases (two or three) would be literal replications, whereas a few other
cases (four to six) might be designed to pursue two different patterns of
theoretical replications.” (pg 46)
3.8. “The eight cases are sufficient “replication” to convince the reader of a general
phenomenon.” (pg 48)
3.9. On pages 47-49 Yin does a great job differentiating case studies from
statistics/sampling paradigm

4. Designing Case Studies

4.1. “A research design is the logic that links the data to be collected (and the
conclusions to be drawn) to the initial questions of a study.” (pg 18)
4.2. Components of research design (all research, not just case study) (pg20-27)
4.2.1.1. A study’s questions
4.2.1.2. Its propositions, if any
4.2.1.3. Its unit of analysis
4.2.1.4. The logic linking the data to the propositions

Morrissette ID # 167952 Page 2 of 5

 4.2.1.5. The criteria for interpreting the findings
4.3. Empirical results of the case study are compared to the previously developed
theory. If two or more cases are shown to support the theory, then replication
may be claimed. (pg 31)
4.4. “Covering these five components …will effectively force you to begin
constructing a preliminary theory related to your topic of study.” (pg 27)
4.5. “For case studies, theory development as part of the design phase is essential,
whether the ensuing case study’s purpose is to develop or test theory.” (pg 27)
4.6. “…theory development prior to the collection of any case study data is an
essential step in case studies.” (pg 28)
4.7. Validity: Design should include multiple sources of evidence (interviews,
company documents/records, external documents, etc) (pg 35)
4.8. Validity: A theory must be tested through replications of the findings in a second
or even third case. (pg 36)
4.9. Reliability: “… if a later investigator followed exactly the same procedures…
and conducted the same case study all over again, the later investigator should
arrive at the same findings and conclusions.” (pg 36)
4.10. Interesting discussion of single case with multiple units (ex. Multiple
departments within a company) (pg 39)
4.11. How determine number of cases:
4.11.1. One case if testing a well-formulated theory or the case is
rare/unique event (pg 38-40)
4.11.2. 2-3 cases when rival theories are grossly different (pg 46)
4.11.3. 5, 6 or more if want high degree of certainty or if rival theories
have subtle differences (pg 46)
4.11.4. Can modify design after some early data collection (realize design
was faulty); often use pilot studies to test and revise design

5. Preparing for Data Collection

5.1. Not everyone is good at this! Researcher needs good questioning skills, listening
skills, must know subject, be unbiased (this is tough!) and be able to see
contradictory evidence (pg56)
5.2. If using multiple investigators, need to train them to get uniformity/consistency
(pg 59-63)
5.3. Develop a case study protocol (procedures and rules) (pg 63-73)
5.4. Interesting: case study questions are posed to “you,” the investigator, not to the
subject being interviewed. You interview the subject or review the material
so you can answer the question. (pg 69-73)

6. Collecting the Evidence

6.1. Six sources of evidence (pg 81-90)
6.1.1. Documentation (memos, newspaper clippings, etc)
6.1.2. Archival records (service records, surveys they’ve conducted)
6.1.3. Interviews (this isn’t the only source!!!) (use tape recorders?)
6.1.4. Direct Observation (watch and note)

Morrissette ID # 167952 Page 3 of 5

 6.1.5. Participant-Observation (great, but high risk of bias)
6.2. Three principles of data collection (pg 90-99)
6.2.1. Triangulation: use multiple sources of evidence (Yin cites Michael
Patton!)
6.2.2. Create Case Study Database (one for data, one for investigator
reports)
6.2.3. Do not edit the data
6.3. “The investigator is the respondent, and his or her goal is to cite the relevant
evidence -- whether from interviews, documents, observations, or archival
evidence -- in composing an adequate answer. (pg 97)
6.4. Maintain the chain of evidence (like a criminologist, how trace from conclusions
back to data)

7. Analyzing Case Evidence

7.1. Four dominant analytic techniques
7.1.1.1. Pattern matching: compares pattern observed in case (empirical)
with pattern predicted by theory (pg 106)
7.1.1.2. Explanation building: “To ‘explain’ a phenomenon is to stipulate a
set of casual links about it.” (pg 110)
7.1.1.3. Time-series (pg113 ff)
7.1.1.4. Program logic
7.2. Interesting discussion of the difference between data within a case versus data
across cases. A case study, uses the data within the case to draw conclusions
about the case. Do not pool sub-data across cases. For instance, if you were
doing 5 case studies, and as one source of evidence you administered a
survey, you would not pull out just the survey and compare it across the 5
cases. (pg 120) (I’m not sure I understand this). However there is a technique
called “cross case analysis.” (pg 121-122)
7.3. High Quality Analysis (pg 123-124)
7.3.1.1. Analysis should show that it relied on all the relevant evidence
7.3.1.2. Include/discuss all major rival interpretations
7.3.1.3. Address the most significant aspect of your case
7.3.1.4. Bring your own prior, expert knowledge to your case study

8. Composing the Case Study Report

8.1. Identify the audience
8.2. Case studies are well suited to communicate research information to non-
specialists (pg 131)
8.3. Review previous case study reports that have been successful in communicating
with your target audience.
8.4. In addition to the obvious single case report or multi-case (with each case
covered in a separate chapter), there is an interesting format where “…there
may no separate chapters or sections devoted to individual cases. Rather, the
entire report may consist of the cross-case analysis, whether purely descriptive
or covering explanatory topics.” (pg 135) “The individual cases, in a sense,
serve only as the evidentiary base for the study …” (pg 137) “…with

Morrissette ID # 167952 Page 4 of 5

 pertinent examples drawn from each of the four “cases”… there is no attempt
to present the single revolutions as individual case studies.” (pg 137)
8.5. Good advice: Decide the type of report during design, before doing case studies,
not after.
8.6. 6 Structures/organization for report: Linear-Analytic, Comparative,
Chronological, Theory Building, Suspense (“answer” or conclusions not
disclosed until end of report), Un-sequenced (often used for descriptive case
studies) (pg 140-141)
8.7. Recommend start writing early (can write bibliography and methodology before
start) (can draft descriptive section before you’ve done the analysis and
conclusions) (pg 141)
8.8. Good discussion of keeping case identities anonymous or disclosing (pg 143-
144). Note of caution: if change names to create anonymity, easy for writer to
get confused and make mistakes!
8.9. Have participants/subjects of the case review draft.
8.9.1.1. Not only a matter of courtesy, but helps catch mistakes
8.9.1.2. Some even include their responses as part of the final document
(Here’s how they reacted…) (pg 144)
8.9.1.3. Okay if subject disagrees with your conclusions, but they should
not disagree with your facts (pg 145)
8.9.1.4. Their review often uncovers additional information
8.10. High Quality?
8.10.1.1. Case study must be significant: before starting study you should
know why it matters
8.10.1.2. Must be complete: convince reader that investigator included all
relevant evidence, little relevant evidence remains untouched (pg
148)
8.10.1.3. Consider alternative perspectives (if not do so, it raise reader’s
suspicions) and show evidence to refute (pg 149)
8.10.1.4. Display evidence to prove conclusions (without making it a data
dump) (pg 150)
8.10.1.5. Avoid bias where one case gets most/more attention
8.10.1.6. Use an engaging style that keeps the reader interested

9. Varia
9.1. Book is part of a great series on Social Research Methods by Sage (40
books!). Some other books in series that interested me: Survey Research
Methods (F Fowler), Participant Observation (D. Jorgensen),
Standardized Survey Interviewing (F Fowler and T Mangione),
Applications of Case Study Research (R Yin).

9.2. Literature review is to develop sharper questions about your topic, not to
find answers or to simply find out what is known. (pg 9).

Morrissette ID # 167952 Page 5 of 5