Overview Methodology Marketing & Strategic Management (Weblectures)
1. Introduction
Four main techniques, elaborations and add-ons:
- Explorative factor analysis
o Reliability analysis
Confirmatory factor analysis
- AN(C)OVA
o Interaction effects
Multiple AN(C)OVA
- Multiple Regression
o With dummies and interaction effects
Logistic regression
- Partial Least Squares
o Intro Structural Equation Modeling (SEM)
Measurement Models and/or Path Modeling
A hypothesis consists of two parts:
- A condition
- A consequence
Each of the parts contains a construct
- The independent variable (condition)
- The dependent variable (consequence)
A construct is a term used to describe a phenomenon or theoretical interest
- Is quantifiable and directly or indirectly observable
- Indirectly observable constructs are called ‘latent constructs’
o Ex. IQ
Different relationships between constructs:
Direct Causal relationship
o A is an exogenous variable; B is an endogenous variable
o
(Fully or partially) mediated (indirect) causal relationship
o Z is the mediator, mediation is partial if effect between A and B remains significant
after inclusion of mediator
o
Spurious relationship
o A third variable influences both A and B
o
Bidirectional (cyclic) causal relationship
o A lead to B, and B leads to A
o
, Unanalysed relationship
o There is a correlation between A and B
o
Moderated causal relationship (interactions)
o The strength and/or direction of the effect of A on B depends on the level of M
o Here, M is the moderator
o
Measurement Model and Structural Model
X11 = Indicators that represent a particular question
E1 = The construct
n1 = Latent variables that are used to represent phenomena that cannot be measured directly (e.g.
beliefs, intentions, motivations)
E21 = Researchers often not expect their models to perfectly explain reality, therefore sometimes
they model structural error, which is in for simplicity often left off (looks at if there are any systematic
biases that might influence how we measured these items).
Multi-item measurements – why?
Increase reliability and validity of measures
Allows measurement assessment (measurement error, reliability, validity)
Two forms of measurement models:
o Formative (emerging)
Direction of causality is from measure to construct
No reason to expect indicators to be correlated
, Dropping an indicator from measurement may alter meaning of the construct
Based on multiple regression (beware of multicollinearity)
Typical for success factor research
o Reflective (latent)
Direction of causality from construct to measure
Indicators expected to be correlated
Dropping an indicator from the measurement model does not alter the
meaning of the construct
Takes measurement error into account at item level
Similar to factor analysis
Typical for consumer research constructs (e.g. attitudes)
2. Factor Analysis
2.1 Introduction
Factor analysis = Estimate a model which explains variance/covariance between a set of observed
variables (in a population) by a set of (fewer) unobserved factors & weightings
Ex. Collect data (how fair you find the grading and how satisfied you are)
What is factor analysis?
- Interdependence technique
- Define structure among variables
- Interrelationships among large number of variables to identify underlying dimensions (=
factors)
- Data summarization and reduction
, Reflective measurement models
- Direction of causality from construct to measure
- Correlated indicators
- Takes measurement error into account at item level
- Validity of items is usually tests with factor analysis
2.2 conducting Factor Analysis
Analysis Process:
Problem formulation
- Which variables
o Best on past research, theory, and judgement of the researcher
o Measurement properties (ratio, interval)
o Sample size (4-5 respondents per variable)
Distinguish between Exploratory Factor Analysis and Confirmatory Factor Analysis
- Exploratory Factor Analysis
o There is not so much knowledge about the subject yet
o Find an underlying structure
o Assumptions that superior factors cause correlations between variables
o Reveal interrelationships
o Purpose is to generate hypotheses
- Confirmatory Factor Analysis
o A priori ideas of underlying factors, derived from theory
o Relationships between variables and factors before conducting the factor analysis
o Purpose is to test hypotheses
Example: You want to conduct a research among consumers on how they perceive toothpaste
- First you collect data and put them in a data set
- Image you have 6 variables (items/questions) that you ask consumers (how important is it
that the toothpaste prevents cavities, that it gives shiny teeth, etc.)
- Interested in how these variables related to each other and whether they provide you with
higher order dimensions
Now construct the correlation matrix
1. Introduction
Four main techniques, elaborations and add-ons:
- Explorative factor analysis
o Reliability analysis
Confirmatory factor analysis
- AN(C)OVA
o Interaction effects
Multiple AN(C)OVA
- Multiple Regression
o With dummies and interaction effects
Logistic regression
- Partial Least Squares
o Intro Structural Equation Modeling (SEM)
Measurement Models and/or Path Modeling
A hypothesis consists of two parts:
- A condition
- A consequence
Each of the parts contains a construct
- The independent variable (condition)
- The dependent variable (consequence)
A construct is a term used to describe a phenomenon or theoretical interest
- Is quantifiable and directly or indirectly observable
- Indirectly observable constructs are called ‘latent constructs’
o Ex. IQ
Different relationships between constructs:
Direct Causal relationship
o A is an exogenous variable; B is an endogenous variable
o
(Fully or partially) mediated (indirect) causal relationship
o Z is the mediator, mediation is partial if effect between A and B remains significant
after inclusion of mediator
o
Spurious relationship
o A third variable influences both A and B
o
Bidirectional (cyclic) causal relationship
o A lead to B, and B leads to A
o
, Unanalysed relationship
o There is a correlation between A and B
o
Moderated causal relationship (interactions)
o The strength and/or direction of the effect of A on B depends on the level of M
o Here, M is the moderator
o
Measurement Model and Structural Model
X11 = Indicators that represent a particular question
E1 = The construct
n1 = Latent variables that are used to represent phenomena that cannot be measured directly (e.g.
beliefs, intentions, motivations)
E21 = Researchers often not expect their models to perfectly explain reality, therefore sometimes
they model structural error, which is in for simplicity often left off (looks at if there are any systematic
biases that might influence how we measured these items).
Multi-item measurements – why?
Increase reliability and validity of measures
Allows measurement assessment (measurement error, reliability, validity)
Two forms of measurement models:
o Formative (emerging)
Direction of causality is from measure to construct
No reason to expect indicators to be correlated
, Dropping an indicator from measurement may alter meaning of the construct
Based on multiple regression (beware of multicollinearity)
Typical for success factor research
o Reflective (latent)
Direction of causality from construct to measure
Indicators expected to be correlated
Dropping an indicator from the measurement model does not alter the
meaning of the construct
Takes measurement error into account at item level
Similar to factor analysis
Typical for consumer research constructs (e.g. attitudes)
2. Factor Analysis
2.1 Introduction
Factor analysis = Estimate a model which explains variance/covariance between a set of observed
variables (in a population) by a set of (fewer) unobserved factors & weightings
Ex. Collect data (how fair you find the grading and how satisfied you are)
What is factor analysis?
- Interdependence technique
- Define structure among variables
- Interrelationships among large number of variables to identify underlying dimensions (=
factors)
- Data summarization and reduction
, Reflective measurement models
- Direction of causality from construct to measure
- Correlated indicators
- Takes measurement error into account at item level
- Validity of items is usually tests with factor analysis
2.2 conducting Factor Analysis
Analysis Process:
Problem formulation
- Which variables
o Best on past research, theory, and judgement of the researcher
o Measurement properties (ratio, interval)
o Sample size (4-5 respondents per variable)
Distinguish between Exploratory Factor Analysis and Confirmatory Factor Analysis
- Exploratory Factor Analysis
o There is not so much knowledge about the subject yet
o Find an underlying structure
o Assumptions that superior factors cause correlations between variables
o Reveal interrelationships
o Purpose is to generate hypotheses
- Confirmatory Factor Analysis
o A priori ideas of underlying factors, derived from theory
o Relationships between variables and factors before conducting the factor analysis
o Purpose is to test hypotheses
Example: You want to conduct a research among consumers on how they perceive toothpaste
- First you collect data and put them in a data set
- Image you have 6 variables (items/questions) that you ask consumers (how important is it
that the toothpaste prevents cavities, that it gives shiny teeth, etc.)
- Interested in how these variables related to each other and whether they provide you with
higher order dimensions
Now construct the correlation matrix
