Case studies

1
Chapter 14

• Case studies

2
Case studies vs Surveys

• Survey
• Get a little bit of info about a large population
• Case studies
• Specific cases in more detail
• The challenge is to make a case study useful
• Who says info from few cases has any bearing on
  other?

3
Case studies - techniques

• Structured and unstructured interviews
• Direct observation
• Direct measurements
• Participant observations
• Group discussions
• Quantitative Qualitative information

4
Case studies

• Range from
• Technical systems
• To social systems
• Case studies in IT is relatively scarce
• Structured approach vs Exploratory

5
Case study design
Holistic
Embedded
Single-case
Multiple-case
6
Case study design
Holistic
Embedded
Single-case
Comparisons
Multiple-case
7
Case study design
Holistic
Embedded
Critical case Extreme Unique Revelatory
Single-case
Multiple-case
8
Case study design
Holistic
Embedded
Single-case
Focus?
Multiple-case
9
Case study design
Holistic
Embedded
Single-case
Generalization?
Multiple-case
10
The protocol

• Sets out precisely what is to be studied and how
  its to be done
• Helps in retaining focus
• Spell out
• What you expect to learn (the theory/hypothesis)
• Questions to be asked
• Aspects to be observed
• Commit to paper before study commences
• Multiple-cases pilot study?

11
Data collection

• Quantitative
• Surveys
• Measurements
• Qualitative
• Observations
• Feelings
• Variety of methods used in a case adds strength

12
Example

• P 100
• If your theory is that some system improves
  patient care
• Select a few hospitals that use the system.
• And

13
Causation

• Surveys could only show correlation
• Case study (can) collect evidence of causation
• Remove other reasonable explanations
• Get similar results in multiple cases

14
The role of the researcher

• Participant-observer to unobtrusive observer
• Remember
• Hawthorne effect?
• Unobtrusive difficult to collect info

15
Record keeping

• Keep complete records of observations
• If something become apparent later you should be
  able to go back to previous observations
• Reconsider correctness
• If available for others to scrutinize can add
  scientific merit
• How? capture data in its original form
• Tape recordings of interviews (transcribed)
• Printouts
• Photographs

16
Case selection

• Literal replication
• Random?
• Cases where theory least likely to apply
• Theoretical replication
• X iff Y, select a couple where X applies and not
• X causes Y, select case where X applies and cases
  where Y applies

17
Data analysis

• Pattern matching
• Predicts observations

18
Generalizing results

• Number of successes in a row
• If chosen probably the odss are stacked in your
  favor
• See p 102-103