Title: Role of Frameworks, Role of Models, Types of Model, Research Criteria
1MIS 650Research Frameworks and Models
- Role of Frameworks, Role of Models, Types of
Model, Research Criteria
2Basic Definitions
- Framework
- Model
- Variable/Surrogate/Proxy
- Concept/Construct
3Framework
- A FRAMEWORK is a matrix within which research
efforts can be classified - The dimensions of the framework are limitless,
but often include roles, effects, technology,
users, etc. - Frameworks are informal, cannot be proven and
are used to compare or situate papers and ideas
4Awop bopalumbop A
?
Yaddayadda yadda Ya dadad yadad
A Framework is a pair of intellectual spectacles
to see with and a set of cubbyholes to display
what you want to see.
?
?
5Uses of Research Frameworks
- Classification of phenomena
- Comparison of phenomena
- Generation of research questions
- Creation of research models
- Evaluation of field efforts
?
?
?
?
?
6Model
- A simplification of some aspect of reality
- Simplification occurs through
- Lumping (Categorization)
- Elimination
- Models contain elements with relations among
those elements. These reflect relationships
among elements of the base reality
7Hierarchy of Theoretical Entities and
Relationships
Theory
Concept
Con-struct
A b s t r a c t i o n
Variable
Proxy/ Surrogate
Data
8Concept
- Conceive ME conceiven lt OFr conceveir lt L
concipere to take in, receive lt com together
capere to take - An idea or thought, esp. a generalized idea of a
thing or class of things abstract notion. - Concepts serve both to sum up (generalize)
thoughts as well as organize and relate them.
Concepts are sophisticated forms of experience.
9Construct
- Construct ltL constructus lt com together
struere to pile up, build rel. to strew - Something put together systematically
- An idea or perception resulting from a synthesis
of sense impressions, etc. - A concept or theory devised to integrate in an
orderly way the diverse data on a phenomenon.
10Variable/Surrogate/Proxy
- A variable is a theoretical concept (actually
construct) that varies in value part of a
theory. - A surrogate or proxy for that variable is a
measurement taken in the real world of that
variable part of testing or using a theory. - All proxies present problems of validity,
reliability and pragmatic measurement difficulties
11Models Level of Explanation
Physical makes things happen in the modeled
system
Causal asserts that some phenomena in
the modeled system cause other phenomena
Logical Asserts logical dependencies
among elements of the modeled system
Classificational groups phenomena within the
modeled system together
Descriptive describes in translated terms the
modeled system
Depictional depicts or reproduces aspects of
the modeled system
Atheoretical
12Types of Models Causal / Associational
IV is sufficient to cause change to DV of
strength a
c
a
IV
DV
EV1
CV
b
d
Direct Cause IV-DVIndirect Cause
IV-EV-DVCommon Cause EV-(IV,DV)Effectiveness
a/(ab)Loss cEfficiency CV-IV-DV
EV
EV is sufficient to cause change to both IV and
DV of strengths r(d,b)
13Types of Models Process
IV4
IV6
IV2
b4
b6
DV
b2
a
b1
b3
b5
IV3
IV5
IV1
Each IV is necessary at a certain level of
strength b to allow change to DV of strength a
14Types of Models Multidomain Modeling
DOMAIN 3 Statements / Interpretations
DOMAIN 1 Sources - Industry characteristics-
DOMAIN 2 Industry Experience/Action - Outcomes
for Industry
15Types of Models Rationalising
- Description of the use of a model
- Description of a phenomenon for whom
- Prediction of a phenomenon for when
- Control of a phenomenon why, how much
- Operation of a phenomenon ?
16Types of Models Contingency
X
Y
a
X
Y
b
X
Y
X
In different regions, X and Y have different
relationships at different strengths. A region
may or may not be describable in terms of a set
of variables itself
17Types of Models Classificational
The phenomena within a region or system are
arbitrarily classified as being similar and
dissimilar based on a scheme of controlling
variables (A, B) which are in turn justified
theoretically (in terms of another theory)
A
B
18Types of Models Time Series
Treatments across time allow changes in (a)
criterion value(s). There is no control group
the treatment effectiveness is compared to
non-treatment effectiveness.
Baseline
Criterion Value(s)
0 1 2 3 4 5
6 7 8 9 10
Important times (in series)
19Types of ModelsEvaluation
- hkhjhjk - hkjhjh - ghgj - gjhgh
Action Domain
Evaluative Environment
20Types of Models Layered
Level 1
Level 2
Level 3
21Research Criteria
- Are the results meaningful?
- Validity
- Are the results trustworthy?
- Reliability
- Significance
22Validity
This relationship has no correspondence in reality
These elements have no counterpart in reality
These elements have no counterpart in the model
This relationship is not mirrored in the model
23Threats to Validity
- History
- Testing
- Instrumentation
- Mortality
- Maturation
- Ambiguity about causal direction
- Generalisability
- Measuring the wrong thing
24Types of Research Errors
Type I Type II Type III
Seeing something that isnt there in general,
someone else will disprove you Not seeing
something that is there waste of time and
effort puts a chill on creativity Not looking
in the right place severe methodological
problems or sampling or theoretical problems