12 conceptual model building model representation
Post on 29-Jun-2015
134 Views
Preview:
DESCRIPTION
Transcript
Conceptual Model Building, Model Representation,Conceptual Model Building, Model Representation,Construct Operationalisation Review (CORE) and Construct Operationalisation Review (CORE) and
Construct ValidityConstruct Validity
Professor Luiz MoutinhoProfessor Luiz MoutinhoUniversity of Glasgow, ScotlandUniversity of Glasgow, Scotland
SCIENTIFIC MODELLING is the process of generating abstract, conceptual, graphical or mathematical models.
•A scientific model can provide a way to read elements easily which have been broken down to a simple form.
•Modelling is an essential and inseparable part of all scientific activity.
RESEARCH MODELS
used to describe the overall framework used to look at reality, based on a philosophical stance
•E.g. empiricism, positivism, postmodernism, post-structuralism•Models identity basic concepts and describe what reality is like, and the conditions by which we can study it•Ideas identified in models are called concepts.
RESEARCH MODELSMeanings and Communication
•Concepts are used to impose some sort of coherent meaning on the world•It is through them that we can make sense of reality, and perceive order and coherence•Used to communicate our experience of the environment around us•Our perception of our surroundings is therefore highly dependent on the scale of our knowledge and our familiarity with a wide range of concepts.
RESEARCH MODELS
Concepts and Statements
The use of concepts on their own is limited in research
We expect that research should provide: A system of classification Offer explanations, Make predictions, and Acquire a sense of understanding
Concepts are only useful in providing a system of classification (typology) the remaining expectations are met by research statements.
Definitions
•Model A model is a set of variables and their inter-relations designed to
represent some real system or process, but that not fully represent that same reality
•A theoretical framework is How one theorizes or makes logical sense of the relationships
among factors that have been identified as important to the problem being investigated, which integrate one’s logical beliefs with published research
•Theory comes from ancient Greek theôria meaning contemplation, speculation. Thea – “a view” + horan – “to see”, literally “looking at a show”
•It is from the theoretical framework, then, testable hypotheses can be developed to examine whether the theory formulated is valid or not
RESEARCH MODELS
Caveats
Sometimes the word ‘model’ is used incorrectly – it is used in a much more constrained sense to indicate a set of (hopefully) related concepts
Similar use occurs with ‘framework’ – have to distinguish between the disciplinary usage of the term and the term as it exists in the philosophy of science
Also known as research philosophy, research paradigm,….
THEORIES, METHODS, DOMAINSSingle versus Multi-method Research
•In general the best advice would be to select a single paradigm (qualitative or quantitative) for your research work (after Creswell 2003)•Based on pragmatic choices: time, skills, and overall size of the project•The reason for this is that qualitative and quantitative research are based on differences in:
• Nature of reality – ontology• Relationship to that being research – epistemology• Roles of values – axiology• Use of language/words – rhetorical• Overall process of research - methodological
THEORIES, METHODS, DOMAINS
Methods and Techniques
•Methods (a.k.a techniques) are used to reveal the existence of, identify the ‘value’, significance or extent of, or represent semantic relationships between one of more concepts identified in a model from which statements can be made.
•Sometimes a distinction is made between methods and technique – one definition has technique as the way or manner in which a method is applied or deployed
Reductionism – problems as a whole are better understood if they are reduced into the simplest possible elements.
Entropy is designed to encapsulate the whole of the research measurement process as a system.
Parsimony is the avoidance of unnecessary complexity
Three purposes for modelling in management – measurement, decision support, and explanation or theory building
The purpose of the model is to provide a prediction system which can be manipulated to study the real world situation for the purpose of aiding the decision maker.
Model Benefits
Small models can offer insight
Models can identify phenomena
Operational models can provide long-term benefits
A model is a set of variables and their inter-relationships designed to represent some real system or process.
A theoretical model is a set of assumptions that describes a subject environment. Its purpose is to explain the subjects’ phenomena.
Model Representation
… it models only a very small fragment of a world.
Models in general are representations of systems and their entities in an abstract, easy to understand way, i.e., the model does not contain all the possible details of the actual system, but presents a limited parallelism.
SCIENTIFIC VISUALISATION
(Post) is the transformation, selection, or representation of data from simulations or experiments, with an implication explicit geometric structure, to allow the exploration, analysis, and understanding of the data.
(Pre) Systems visualisation provides a researcher the ability to quickly understand the complexity of a theoretical system. The need of knowledge representation through ontologies.
What is a Model?
A model is a stylized representation of reality that is easier to deal with and explore for a specific purpose than reality itself.
We use the following types of models:
Verbal
Box and Arrow
Mathematical
Graphical
An Example of a Verbal Model
Sales of a new product often start slowly at “innovators” in the population adopt the product. The innovators influence “imitators”, leading to accelerated sales growth. As more people in the population purchase the product, sales continue to increase but sales growth slows down.
Boxes and Arrow Model
Fixed Population Size
Imitators
Timing of Purchases by Imitators
Pattern of Sales Growth of New Product
Innovators
Timing of Purchases by Innovators
InnovatorsInfluenceImitators
Mathematical Model
))(( ttt xNbxadt
dx
Where:
xt = Total number of people who have adopted product by time tN = Population sizea,b = Constants to be determined. The actual path of the curve will depend on these constants
Graphical Model
FixedPopulation Size
Time
Cumulative Salesof a
Product
The model builder’s task is to identify those factors that have a significant effect on the dependent construct and to try to express those effects as “influence functions”, which are then incorporated in the model.
The process of model building is a problem-solving analysis that enables attention to be focused on the relevant factors.
The Modelling Process
Dimensions
Constructs
Sub-constructs
Variables
Items
Measures
Scales
Surveys
Case Study Research
Action Research
Grounded Theory
Mode 2 Research
Ethnographic Research
Positivism
Realism
Interpretativsm
Constructivism
Ontology
Ontology
Reality
RealityPhilosophicPerspectives
Method & Strategies of
Research
Modelling
Natu
reK
no
wled
ge
Research D I C O S U V A M (i) Mnemonic
Modelling
DIMENSIONS(DOMAINS)
CONSTRUCTS
SUB-CONSTRUCTS
VARIABLES
MEASURES
ITEMS
Construct/Construal - is a term/concept that is somehow involved in the
management problem that will be researched
Parameters - define the “state of the system”
Constants - are characteristics of the system that do not change
•The model may be considered to represent relationships among constants, parameters and variables of the business system and its environment. The constants are characteristics of the system which do not change over the period of interest.
•Parameters may be controllable or non controllable in the real world, but once given, they generally remain constant over a time period. Parameters describe the “state” of the system.
•Variables may also be either controllable or non controllable and represent characteristics of the system which change continuously or are subject to change at any time.
MotivationsSocial Groups
Event Attraction& Excitement
Cultural/ Historical Attraction
Socialization & Entertainment
Family Togetherness
Site Novelty
Group Togetherness
G1G2G3
M1M2M3
M4M5M6
M8M9
M10
M11M12M13
M14M15M16
M5M6M7
Theoretical Construct
Operational Construct
Empirical Concept (Variable)
Proposition
Correspondence Rules
HypothesesOperationalDefinitions
Soil of Observation (Experience)
Theoretical framework. Adapted from Bagozzi (1980)
Effects of Group Membership and Motivation on Attendance
Model High-Level Abstraction
•Epistemological (meaning)
•Ontological (nature of being)
•Etymological (roots)
•Nomological validity (theories)
•Semiotics(signs and symbols)
•Epidemics (map out the scope)Causal Network of Concepts
•Identification
of concepts and relationships
of direction of impact
of intensity
of direct/indirect effects
•The notion of the concept “stretching” relates to a construct that is defined at a high level of abstraction and has both a large bandwidth (i.e., broad coverage) and a wide connotation (i.e., lumping of too many classes of things with little attribute precision).
(Osigweh 1989)
Construct Operationalisation Review
(CORE)
CORE•NOMOLOGICAL VALIDITY
•THEORETICAL UNDERPINNINGS
•DISSECTION OF CONSTRUCT VALIDITY
•DEFINITIONS, DIVIDES AND DILEMMAS
•MEASUREMENT SCALING
+•CONTEXT OF RESEARCH
•MANAGERIAL IMPLICATIONS
Review of measurement scales used in previous studies
Loyalty -Repurchase probability (Andreassen and Lindestad, 1998)
-Word of mouth
-Prefer the particular bank to other financial institutions (Nguyen and LeBlanc, 1998; Ruyter, Wetzels and Bloemer, 1998)
-INTEND TO USE THE BANK IN THE YEARS TO COME (Ruyter and Bloemer, 1999; Ruyter, Wetzels and Bloemer, 1998; Pedersen and Nysveen 2001)
-Consider the bank the first choice for following purchases (Ruyter and Bloemer, 1999; Wetzels and Bloemer, 1998; Yu and Dean, 2001)
-Propensity to switch (Parasuraman et al., 1994 – cited in Rosen and Surprenant, 1998; Zeithaml et al., 1996; Mittal and Lassar, 1998; Moutinho and Brownlie, 1989)
-Cognitive loyalty
- provide positive word-of-mouth to potential buyers (Andreassen and Lindestad, 1998; Yu and Dean 2001)
-intend to recommend the financial institution to others (Nguyen and LeBlanc, 1998; Yu and Dean, 2001; Pedersen and Nysveen, 2001)
-Say positive things about the bank to other people (Ruyter, Wetzels and Bloemer, 1998; Yu and Dean, 2001)
-Encourage friends and relatives to do business with the bank (Ruyter, Wetzels and Bloemer, 1998; Yu and Dean, 2001)
Review of measurement scales used in previous studies
Service attributes
Inertia
-Location (Moutinho and Brownlie, 1989; Leonard and Spencer, 2991)
-New and improved services (Moutinho and Brownlie, 1989)
-Helpfulness of staff (Moutinho and Brownlie, 1989)
-Banking fees (Moutinho and Brownlie, 1989)
-Queueing (Moutinho and Brownlie, 1989)
-Courteous personnel (Leonard and Spencer, 1991)
-Modern facilities (Leonard and Spencer, 1991)
-Loans money easily (Leonard and Spencer, 1991)
-No established scale in the existing body of literature
Definition of variables and measurement used in the study
Construct Type of variables
Definition of variables Key relevant research
Measurement
Control Dependent Mechanism of control: the method by which control is exercised
Schann (1988) Measuring the appointment of high-level managers made by whom
Extent of control: the degree of control over a JV is exercised by parents
Killing (1983)Wu (1994)Cheng (1996)
Measuring the score and classifying into 4 type of control
Focus of control: the control is exercised in which area of joint venture’s operation
Killing (1983)
Ownership Independentdependent
The share of equity held by the Taiwanese and foreigner parents
Measured as a percentage
Bargaining Power
IndependentDependent
The power that can be used to affect the outcome of the negotiation process
Fagre & Wells (1982)Yan & Gray (1994)
Using 5-points scale to measure 10 items
Contribution IndependentDependent
The resources contributed by the Taiwanese and foreigner parents
Grant (1991)Barney (1991)Chatterjee & Wernerfelt (1991)
Using 5-points scale to measure 14 items
Motivation Independent The motives of companies cooperate with other firms to establish a joint venture
UNEC (1988) and relevant literatures
Using 5-points scale to measure 26 items
Development of Conceptual Framework
Identify constructs/ concepts
Specify and define constructs/concepts
Decide on a model
Operationalise constructs/concepts
(operational definitions) Identify and explore
relationship between constructs/concepts
Operationalisation of the Constructs
•A model ties together various constructs and their relationships
•A model connects constructs in ordered relationships
•A relationship is a meaningful link believe to exist between two constructs
•The researcher translates the construct into an operational definition, which describes how a researcher measures the construct. That is, an operational definition implies a specific question format that will be used in a survey. To capture the essence of each research construct. Specific measurement scaling and statistical analyses.
Operationalisation of the Constructs
•Decide on a model – once you take a set of constructs and order their relationships with some understandable logic, you create a model
•The amount of effort the researcher spends in gaining an understanding for a research problem, pays off in model specification
What is an operational definition?
•A construct is an abstract concept (the building blocks of a theory and represent the points around which business research is conducted.
•The actual measure of the construct is the construct operationalisation
•Operationalisation refers to the operations by which a concept is measured (example: IQ is a measure for intelligence)
•Operationalising or operationally defining a concept to render it measurable, is done by looking at the behavioural dimensions, facets or properties denoted by the concept
•These are then translated into observable and measurable elements so as to develop an index of measurement of the concept
It is important to be accurate in order to have valid measures.
What an operational definition is not
•Does not describe the correlates of the concepts
•Does not delineate the reasons, antecedents or consequences of the concept
An example: Operationalising the concept of “learning”
Delineate the dimensions and elements of the concept learning
Learning
ApplicationRetentionUnderstanding
Integrate with other relevant
material
Solve problems applying concepts
understood and recalled
Recall material after some
lapse of time
Give appropriate examples
Answer questions correctly
Modelling + Measurement (M2)
•SUBSTANTIATE
•DEFINE
•MEASURE
(SDM)
Cross-Examination of Model Constructs
•Comparability - (CL) – construct, operational measures and interrelationships between
constructs
•Relativity - time frame and amount
•Referencing - attribute by attribute equivalence or valence
•Variability - contextual and group
•Equilibrium - a balanced perspective of the interplay between equivalent research constructs
•Multidimensionality - an assessment of the degree/level and amount of dimensions related to each construct
•Measurability - Construct Measurement Robustness
Construct Validity
The vertical correspondence between a construct which is at an unobservable, conceptual level and a purported measure of it which is at an operational level.
Validity
•Refers to the issue of whether an indicator that is devised to gauge a concept really measures the concept
Face validity: if the measure apparently reflects the content of the concept in question
Concurrent validity Predictive validity Construct validity Convergent validity
The measurement process
•Validity Ensured the ability of a scale to measure the intended concept Types of validity
• Content validity
• Ensures that the measure includes an adequate and representative set of items that tap the concept
• Criterion-related validity
• Is established when the measure differentiates individuals on a criterion it is expected to predict
• Concurrent validity
• Predictive validity
The measurement process
•Validity Types of validity
• Construct validity
• Testifies to how well the results obtained from the use of the measure fit the theories around which the test is designed
• Convergent validity: correlation between two different instruments measuring the same concept
• Discriminant validity: when empirical evidence about correlation between two variables support the initial prediction, based on theory, that the variables are unrelated
• Techniques
• Correlation analysis
• Factor analysis
• Multitrait analysis
For the explanatory-predictive validity of concepts, it is useful to distinguish two evaluative discussions, namely the accuracy and the range of the explanation-prediction. The dimension of accuracy refers to the extent of agreement between the predicted and the observed value for each dependent variable. The range refers to the number of dependent variables that can be explained–predicted with the help of the independent variables. The accuracy is indicated by the criterion-related validity.
Validity of Measurements
•Content validity – the extent to which the measure appears to measure the characteristic it is supposed to measure
•Criterion/pragmatic/empirical validity – the extent to which a measure can be used to predict individual’s score on some other characteristic (the criterion)
•Construct validity – the extent to which the measure behaves in a theoretically sound manner
Different types of equivalence
•Operationalisation equivalence – “best capture the latent construct” (Malhotra and McCort 2001, p. 237)
•Construct equivalents (Singh 1995, Ulijn, et al. 2000,
Weinfurt and Mohgaddam 2001), construct which will …Serve the same function and will be expressed similarly (functional and conceptual equivalence),
Be interpreted similarly (instrumental equivalent), andThe construct equivalency (measurement equivalence) across cultures
Validity Assessment Approaches
Approach Description Procedure
1. Content Validity
The extent to which a measure appears to measure the characteristic it is supposed to measure.
Subjective assessment of the appropriateness of the measure for the task at hand.
1.1 Face Validity The extent to which a measure seems to capture the characteristic of interest.
Agreement between expert and/or non-expert judges as to the suitability of the measure.
1.2 Sampling Validity
The extent to which a “content population” of situations/behaviours relating to the characteristic of interest (i.e. the characteristic’s conceptual domain) is adequately represented by the measure concerned.
As above
2.Criterion Validity*
The extent to which a measure can be used to predict an individual’s score on some other characteristic (the criterion).
Examination of the relationship between the measure and a criterion.
2.1 Concurrent Validity
The extent to which a measure is related to another measure (the criterion) when both are measured at the same point in time.
Comparison of the scores obtained on the measure concerned and those obtained on the criterion.
2.2 Predictive Validity
The extent to which current scores on a given measure can predict future scores of another measure (the criterion).
As above
*Also know as pragmatic or empirical validity
Validity Assessment Approaches
Approach Description Procedure
3. Construct Validity
The extent to which a measure behaves in a theoretically-sound manner.
Investigation of the relationships between the measure concerned and measures of other concepts/characteristics within a theoretical framework.
3.1 Convergent Validity
The extent to which a measure is positively related to other measures of the same concept obtained by independent methods.
Examination of the relationships between measures of the same concept generated by different methods.
3.2 Discriminant Validity
The extent to which a measure is not related to measures of different concepts with which no theoretical relationships are expected
Examination of the relationships between measures of different concepts that are theoretically unrelated.
3.3 Nomological Validity
The extent to which a measure is related to measures of other concepts in a manner consistent with theoretical expectations.
Examination of the relationships between measures of different concepts that are theoretically unrelated
Types of Concept Validity
1. Observational validity The degree to which a concept is reducible to observations
2. Content validity The degree to which an operationalisation represents the concept about which generalisations are to be made.
3. Criterion-related validity
3a. Predictive validity
3b. Concurrent validity
The degree to which the concept under consideration enables one to predict the value of some other concept that constitutes the criterion.
A subtype of criterion-related validity in which the criterion measured is separated in time from the predictor concept.
A subtype criterion-related validity in which the criterion and the predictor concepts are measured at the same time.
4. Construct validity
4a. Convergent validity
4b. Discriminant validity
4c. Nomological validity
The extent to which an operationalisation measures the concept which it purports to measure.
The degree to which two attempts to measure the same concept through maximally different methods are convergent. It is generally represented by the correlation between the two attempts.
The extent to which a concept differs from other concepts.
The extent to which predictions based on the concept which an instrument purports to measure are confirmed.
5. Systemic validity The degree to which a concept enables the integration of previously unconnected concepts and/or the generation of a new conceptual system.
6. Semantic validity The degree to which a concept has a uniform semantic usage.
7. Control validity The degree to which a concept is manipulable and capable of influencing other variables of influence.
As criteria for the descriptive validity of concepts, the following can be proposed: reliability, convergent validity, nomological validity, content validity and discriminant validity. All of these criteria, with the exception of discriminant validity, focus on the relationship between a concept and its real referent.
Discriminant validity
•Discriminant validity describes the degree to which the operationalisation is not similar to (diverges from) other operationalisations that it theoretically should not be similar to.
•The concept of discriminant validity refers to the extent to which an operationalisation of a construct measures just the intended construct itself and not other constructs.
Discriminant validity
•Campbell and Fiske (1959) introduced the concept of discriminant validity within their discussion on evaluating test validity. A successful evaluation of discriminant validity shows that a test of a concept is not highly correlated with other tests designed to measure theoretically different concepts.
•In showing that two scales do not correlate, it is necessary to correct for attenuation in the correlation due to measurement error. It is possible to calculate the extent to which the two scales overlap by using the following formula where is correlation between and is the reliability of , and is the reliability of :
rxy x,y rxx ryy
y
yyxx
xy
rr
r
,
x
Discriminant validity
•Although there is no standard value for discriminant validity, a result less than .85 tells us that discriminant validity likely exists between the two scales. A result greater than .85, however, tells us that the two constructs overlap greatly and they are likely measuring the same thing. Therefore, we cannot claim discriminant validity between them.
Discriminant validity – an example
•Consider a researcher developing a new scale designed to measure Narcissism. They may want to show discriminant validity with a scale measuring Self-esteem. |Narcissism and Self-esteem are theoretically different concepts, and therefore it is important that they researcher show that the new scale measures Narcissism and not simply Self-esteem.
•First, we can calculate the Average Inter-Item Correlations within and between the two scales:
-Narcissism – Narcissism: 0.47
-Narcissism – Self-esteem: 0.30
-Self-esteem – Self-esteem: 0.52
Discriminant validity – an example
•We then use the correction for attenuation formula:
•Since 0.607 is less than 0.85, we can conclude that discriminant validity exists between the scale measuring narcissism and the scale measuring self-esteem. The two scales measure theoretically different constructs.
607.052.047.0
30.0
top related