Analytic Induction

1
MSc by Research in Leading, Learning and Change
Case Study Research
Dr Heather Skipworth Research Fellow, Supply
Chain Research Centre heather.skipworth_at_cranfield.
ac.uk
2
Who am I
1989 BSc Mechanical Engineering, Leicester
University 1989 1991 Project Engineer, Metal
Box 1991 1995 Technical Manager, Field
Packaging 1995 1996 MSc Manufacturing Systems,
Cranfield University 1996 1998 Senior
Manufacturing Systems Engineer, BICC Cables
Limited 1998 2003 PhD Programme, Cranfield
University Application of Form Postponement in
Manufacturing Industry 2004 to date Research
Fellow, Cranfield University
3
Survey of Cranfield Doctoral Thesis Submissions

• Out of 156 thesis submissions between 1987
  2007,
• 65 were case-based,
• 32 used statistical methods,
• 10 used repertory grid
• We major on in-depth research thats relevant
  to practice

4
What Case Study Research is not...

• an aid to teaching
• an interesting story
• promotion of a new fad
• a basket of unconnected observations
• your views with illustrations
• someone elses views with illustrations

5
What is a Case Study?

• investigates a contemporary phenomenon within its
  real life context...
• ...when the boundaries between phenomenon and
  context are not clearly evident

Yin, 2003
6
Prejudices...

• lack of rigour
• biased views, data collection, link conclusions
  to evidence
• lack of generalisability
• n 1, narrow relevance, context specific
• too complex
• data asphyxiation

7
Case Studies in Operations M. Research
Modelling, Experiments
Abstraction
Large Population Surveys
Case Studies, Action Research
Accuracy / Repeatability
8
Charles Ragin
9
Variable-oriented Research

• a true statement about a population...
• may not apply to any individual case
• generalising impedes true understanding
• properties shared by all organisations are
  obvious
• averages show how organisations are the same
• what matters is how they are different
• large samples statistical significance...
• generate significant findings that have no
  meaning
• large sample statistics...
• deflect from individuality, complexity variety

Bill Starbuck
10
How Case Studies can be Used...

• explore social processes as they unfold
• understand social processes in context
• internal, external
• explore new processes or behaviours
• explore extremes
• capture emergent properties
• explore informal or secret behaviour
• cross-national comparative research

Hartley, 1994
11
Applications of Case-Based Research
Exploratory Descriptive Explanatory
Testing
Theory
Generation Testing
12
Research Strategy - induction v deduction?
13
Research Design Considerations

• research questions
• not just a journey into the unknown
• hypotheses
• balance between induction deduction
• data collection
• triangulation (data source, method, investigator)
  for construct validity
• researcher involvement, identity and biase
• data analysis
• within case and cross-case analytic strategies
  for internal validity (Yins research designs and
  Pettigrews framework)
• interpreting the observations
• explaining variation

14

• Can we learn anything from a sample of one?

15
The case of Phineas Gage
16
Yin, 2003
17
Pettigrews meta- level analytical framework
CONTEXT
Business environment, product/manufacturing
process types
CHANGE CONTENT
Reasons for applying FPp its application in a
MTO and MTS environment
OUTCOME VARIABLES
MTS Unit of Analysis
Internal Variables
MTO Unit of Analysis
External Variables
Internal Variables
FPp Unit of Analysis
Internal Variables
Skipworth 2003
18
Example of Case Study Scope
Production Equipment
Product Specs.
Process Specs.
Manufacturing Planning
Production Scheduling
Duration, frequency, capacity plan
Bills of Material
Production line schedules
Process routings
Product
Data
Project Boundary
Project Boundary
Replenishment factory orders
Production line records
Ex-works records
Delivery schedule
Customer Order Processing
Stock Control
Outbound Logistics
Production
Facilities
Mode of transport
Skipworth, 2003
19
Selection in Case Study Research

• Case selection for external validity analytic
  generalisation
• - clarify domain
• - sampling using replication logic
  theoretical or literal
• - extremes and polar types
• Selecting the Unit of Analysis
• - differences in outcome
• - coming to terms with time - snapshot /
  longitudinal / retrospective
• Selecting the data sources/methods
• - informants - opponents / supporters /
  doubters
• - methods - databases / documents / observations
  / interviews

20
Example of different outcomes...
21
Analysing Case Studies

• data collection and analysis iterative process
• - theory data
• within case analysis
• - between units of analysis or establishing
  links between observations
• - qualitative and quantitative data
• cross-case analysis
• - search for patterns
• - similarities differences

22
Eisenhardts Roadmap assumes inductive

• getting started
• selection of cases
• selection of research methods
• entering the field
• analysing data
• shaping hypotheses
• enfolding literature
• reaching closure

Eisenhardt, 1989
23
Eisenhardts Roadmap
24
Analysing Case Study Evidence

• Analysing case studies is always challenging
  because of the detail. It is helped by
• being clear about research objectives
• being clear about the unit of analysis study
  questions
• coming to terms with time
• making your research method explicit
• making your meta level framework explicit
• making your hypotheses explicit
• identifying themes that cut across the data
• using techniques of data reduction display

25
Cross-Case Comparisons