Theorizing in a Practical Science

1

• Theorizing in a Practical Science

Shirley Gregor Simon Fraser University 27
October, 2008
2
Outline

• Part A - Recap on design approaches
• Part B - Work-in-progress
• Initial Problem
• Which paradigm?
• Theorizing in general
• Principles for theory building in I.S.
• Questions?

3
Part A Recap On Design Approaches

• Good Reference
• Keuchler Vaishnavi (2008) The emergence of
  design research in information systems in North
  America. J. of Design Research, 7,1,1-16.
• Also see
• Gregor and Jones (2007)

4

• Examples of design theories/knowledge
• Structured systems analysis, Codds relational
  database theory, technological rules for building
  DSS and other applications
• 30-45 of work in leading journals of this type
  but may not be recognized as such
• Recent popularity of topic - ISWorld site,
  DESRIST conference, Hevner et al (2004) MISQ
  article on design science

5
Philosophy of Science

• Some ideas on design-type knowledge date back to
  Aristotle built upon by Heidegger
• Techne distinguished from episteme

6
the Artificial
Sciences of the Artificial
Design is the core of all professional training
engineering, architecture, business, education,
law, medicine - and information technology Need
a science of design intellectually tough,
analytic, partly formalizable, partly empirical
and teachable. Herbert Simon - Sciences of the
Artificial
7
Constructive Research Design Science

• Iivari (1991)
• Kasanan et al (1993)
• Nunamaker et al (1991)
• March Smith (1995)
• Hevner at al (2004)
• Tends to focus on the research activity and the
  artifact (instantiation) as the main product
  less on the theory that results

8
Hevner et al (2004) MISQ

• Guidelines
• Design research in IS (ISDR) produces artifacts
• ISDR must be relevant
• The design of the ISDR artifact must be
  rigorously evaluated
• ISDR must provide a novel contribution
• ISDR must balance rigour and relevance
• An ISDR contribution must be functional
• ISDR results must be communicated to technical
  managerial audiences
• Note Focus on research that produces an artifact.

9
Work in other disciplines

• Architecture (Alexander 1977)
• Management (Van Aken, 2004)
• Management accounting (Kasanan 1993)
• Education (Samuelson 2003)

10
Part B - Work in Progress

• Need a wider view of nature of research in I.T.
  fields
• Need to understand it more fully against the
  philosophy of science

11
Previous work

• Gregor (2006). Nature of Theory in IS, MISQ
• Five different types of interrelated theory
  distinguished.
• Type V is theory for design and action other
  theory contributes
• Gregor and Jones (2007). The Anatomy of a Design
  Theory. JAIS.
• Showed structural nature of design theory, with 8
  components (1) purpose and scope (2)
  constructs (3) principles of form and function
  (4) artifact mutability (5) testable
  propositions (6) justificatory knowledge (kernel
  theories) (7) principles of implementation and
  (8) an expository instantiation.

12
1. The Initial Problem

• Information Technology disciplines are new
• Questions to do with theory not dealt with
• Little help from philosophy of science does not
  deal with technology

13
Problems of other science paradigms in IT

• Pre-occupation with idealized views of science
• Aim for legitimacy through being scientific
• Dont regard design theory as legitimate form of
  theory (as in wheres the theory?)
• In Inf Sys
• pre-occupation with debates about positivism vs
  interpretivism (largely irrelevant)
• Grab-bag choice of theory to underpin Inf Sys
  work
• Lack of focus on outcomes and things in the real
  world that can be manipulated to achieve these
  outcomes when doing normal science-type work

14

• Teaching inappropriate research methods (not
  including ones to do with design approaches)
• Theories/models investigated that do not lead
  well to design knowledge (eg TAM in Inf Sys)
• Lost opportunity for identifying what really
  defines ICT disciplines
• Lack of relevance of much IT work to real
  problems
• Insufficient thinking about what practical
  science research really means and how to
  practice it

15
2. Which paradigm?
16
Natural Sciences Physics, chemistry, geology,
astronomy
Practical Sciences (Sciences of the
Artificial) Agriculture, engineering, medicine,
law, IT/IS
Human Sciences History, archaeology, linguistics,
cultural anthropology
Note Follows Strasser (1985). There are also
human sciences with prcatical aspects eg
economics, sociology, psychology
17
Approaches change with phenomena of interest

• Pre-science
• Greek philosophers all branches of knowledge
• Age of enlightenment/Age of Reason
• 18th century
• Advances in natural sciences (Newton, Hooke,
  Boyle)
• Word scientist coined (Coleridge/Whewell)
• Galileo /Bacon scientific method (experiments)
• Focus on human sciences
• Sociology (Comte, 1838), Psychology (James,
  1890)
• human science distinguished by Dilthey (1883)
  (First interpretivist?)
• Focus on artificial sciences
• Increasingly complex artifacts with invention of
  computers
• Herbert Simon Sciences of the Artificial (1969)
• Strasser practical sciences (1985)

18
Argument
The IT research paradigm should be one of a
practical science1 which concerns the design
and construction of IT-related artifacts and
interventions in the world. We should not
uncritically adopt our models of research from
the science paradigms, either natural or human
sciences. Herbert Simon made a good start, but
much more work is needed. Cant separate out
design work and then pass other theorizing back
to the natural sciences domain. Need to look at
practical science in totality
main Focus
19
3. Theorizing in general

• Theory building in general poorly understood
• Have inherited hypothetic-deductive method
• Step 1 Generate conjectures (result of
  observations?)
• Step 2 Formulate hypotheses
• Step 3 Test hypotheses
• Step 4 Reject or support hypotheses modify or
  extend theory
• Back to Step 1
• Does this apply in practical sciences?

20

• Some relevant ideas
• Merton (1968) - mid-range theory grounded in
  empirical enquiry useful in disciplines
  concerned with practice
• Grounded theory approach (Glaser Strauss 1967)
• Van Aken (2004, 2005) - in management proposes
  technological rules If you want to achieve Y in
  situation Z, then something like action X will
  help
• Other work in management helpful but not
  practice-oriented
• Weick (1989) disciplined imagination
• Poole Van De Ven (1989) using paradox to build
  theory
• Lewis Grimes (1989) bridging bracketing

21
Practical Science Theorizing
22
Eight Principles for IT/IS Theorizing

• Artifact centrality
• Artifact purposefulness
• Artifacts as systems
• Design research variants
• Differing logics/reasoning
• Differing types of theory
• Mid-range theorizing
• Interior and exterior modes

23
1) Artifact centrality

• By definition, discipline concerns IT artifacts
• Distinguishes our theory from reference theory
• Practical application
• More likely to be accepted in IS journals

24
2) Artifact purposefulness recognized

• Artifacts have a purpose/goal/aim/end
• Theorizing unsatisfactory if goal not studied
• Practical application
• Answers the so what question for journal
  editors/readers
• Be careful with research questions.
• NOT What types of knowledge intermediaries are
  there?
• BUT What types of knowledge intermediaries are
  more effective?

25
3) IT artifacts treated as systems(or concerned
with systems)

• Need general system theory concepts
  system-environment boundary, input, output,
  processes, state, hierarchy, goal-directedness
• Need for different conceptual tools
• Teleological explanations
• Properties of system as a whole
• Practical applications
• Systems theory as underlying theory (as in Weber
  )
• Levels of analysis issues
• Need for dynamic (not static) models
• Trend towards longitudinal analysis analysis
  reverse causality ??

26
4) Range of design research approaches

• Design activities many and varied (discovery,
  art, problem solving .)
• Design/Utility Theory (Type V) can be built in
  many ways
• Developing case study Design Science
  design/build/evaluate
• Extracting case study systemize design
  principles from artifacts already built (van
  Aken)
• Practical application
• Researchers should realize range of approaches.
• Some have been very influential (eg Davenport
  Short)

27
5) Different logics/reasoning needed

• Simon (1996) said special logic of imperatives
  not needed but did not look at open-ended
  design situation
• In practice, range of potential designs huge
• Deductive logic can not lead to imperative
  statements for action drawn from natural/human
  science (Edgley 1969)
• Induction from past cases useful
• Practical application
• Building IT/IS theory from reference theories
  suspect unless some grounding (eg as in TAM
  extensions)

28
6)Different types of theory

• Recognize 5 types of theory
• Type V theory Design and Action is the
  ultimate aim in a practical science
• Reference theory nice as explanation but not
  essential
• Practical application
• Not atheoretic if reference theory missing
• Can build on prior design theory/knowledge (and
  best practice)
• Still questions? (single artifact valuable but
  may not be a theory)

29

• Design theory is linked to theory of other types
  only where possible. Art Theory
• In some cases
• supporting explanatory
• knowledge
• from other reference paradigms is
• little or non-existent

30
7) Mid-range well-grounded theory of value

• Need to design artifacts for particular contexts
• means unlikely to get very high-level generalized
  theory
• Grounding in practice needed because use of
  inadequacy of deductive logic
• Practical application
• Aim at mid-range theory
• Accept inductive methods
• Question of theory (model) testing ???

31
8) Theorize in interior and exterior mode

• Idea from Simon
• In interior mode, focus on design that attains
  some goal/end (design theory)
• In exterior mode, observe artifact in operation

32

• Practical applications
• Can only work in one mode at a time. Recognize
  this. Dont expect researcher in interior mode to
  also to large-scale evaluation.
• Design Science approaches for interior mode (Type
  V theorizing)
• See Principle 4
• Work in exterior mode similar but different from
  traditional science (Types I to IV)
• focus on outcomes, manipulable variables,
  systems concepts, mid-range theorizing

33

• Exterior mode
• Common approach
• Design a feature X (eg explanation facility) to
  bring about some outcome Y (eg user learning).
• Normal science test Conduct experiment to test
  proposition System with X brings about outcome
  Y
• Problem Hard to test properly all design
  decisions in largish systems

34

• Conclusions questions