Introduction lecture 2
Lecture 1.1 Factor analysis – Technique 6
Lecture 1.2 Factor analysis – Video clips 8
Lecture 1.3 Factor analysis – Application lecture 18
Lecture 2.1 ANOVA – Technique 27
Lecture 2.2 ANCOVA – Video clips 32
Lecture 2.3 ANCOVA – Application lecture 41
Lecture 3.1 Multiple Regression with dummies and interaction – Technique 56
Lecture 3.2 Multiple Regression with dummies and interaction – Video clips 59
Lecture 3.3 Regression analysis – Application lecture 69
Lecture 4.1 PLS – Technique 83
Lecture 4.2 Partial Least Squares – Video clips 89
Lecture 4.3 Partial Least Square – Application lecture 102
Lecture 4.1 Closing lecture 118
,Methodology in Marketing and Strategy Research
Introduction lecture
The exam is not an open book exam, the book has to be studied detailed. Exam is in January
Zoom account with RU mail.
2 assignments in block 1 and 2 assignment in block 2
Thesis is not mandatory for quantitative
- Videos in advance
- Assignments in groups + Online tutorial
- Application lecture
Book: Hair, multivariate data analysis 8th edition à 9781473756540
Theory
Theories in MSR are usually not that theoretical, they are better the more they prohibit
(following Popper), are not tautological (they explain something), ought to be empirically
testable or falsifiable, consist of constructs (concepts, phenomena, variables) and hypotheses
(on their interactions or relationships). Popper: the criterion of the scientific status of a theory
is its falsifiability, or refutability, or testability.
A theory is a proposed description, explanation, or model of the manner of interaction
of a set; phenomena, capable of predicting future occurrences or observations of the same
kind, and capable of being tested through experiment or otherwise falsified through empirical
observation.
Hypothesis
Interrelation between two constructs (consist of two parts) – a condition and a consequence.
Each of the parts contains a construct (independent/dependent variable). The independent
variable is the condition, and the dependent variable is the consequence. An example: the
higher A, the higher B or A leads to B
- You can(not) formulate a hypothesis for correlational relationships à you can but it is
more complex. Yes, you can hypothesize that two constructs are associated (e.g., using
a correlation coefficient) instead of one construct influence another construct (e.g. b-
coefficient in regression)
Construct
Conceptual term used to describe a phenomenon of theoretical interest. It is quantifiable and
directly or indirectly observable. They are defined in terms of an object, attribute and rater
entity. An indirectly observable construct is called 'latent', for example: IQ.
- How satisfied are you as a customer? The construct is then customer satisfaction
,Researchers in marketing and strategy research usually want to investigate relationships
between constructs:
- Direct causal relationship
Most often used. A (independent – exogeneous) leads to B (dependent - endogenous).
Arrow pointed at it = endogenous. Usually, but not necessarily, a linear effect is meant.
- Mediated (indirect) causal relationship (fully or partial)
Impact of A to B only exists with Z. Z is a partial mediator if there is also an effect A
and B (If predictor variable A in has a direct significant impact on response variable
B and it also has a significant impact on moderator Z, which has a significant impact
on response variable B, this is known as a case of partial mediation) à Mediation is called
partial if the effect between A and B remains significant after inclusion of the mediator
- Spurious relationship
A third variable (Z for example), influences A and B. There is a relationship between
drowning in pools and ice cream consumption. à In the summer = Z.
- Bidirectional (cyclic) causal relationship
A lead to B, and B leads to A, but note! Not necessarily at the same time.
- Unanalyzed relationship
There is a correlation between A and B
- Moderated causal relationship (interactions)
The strength and/or direct of the effect of A and B depends on the level of M. Here,
M is a moderator (variable).
Example Theory of Reasoned Action
The two-language concept
Intention = mediator in theory of reasoned
action
Theoretical language= Theory of reasoned
action
Observable language = how we measure
constructs
, Plane of theory -> latent data
Observed -> empirical real data
Measurement Model and Structural Model
= indicators are normally represented as squares. For questionnaire-based research,
each indicator would represent a particular question.
= Latent variables are normally drawn as circles or ovals. Latent variables are used to
represent phenomena that cannot be measured directly. Examples would be beliefs,
intention, motivation.
= Few researchers expect their models to perfectly explain reality. They therefore
explicitly model structural error - in the case of error terms, for simplicity, the circle
is often left off.
Measurement model Structural model
Why the multi-item measurement?
Multiple items/questions to measure a certain construct = multi-item measurement. It
increases reliability and validity of measures. It allows measurement assessment of
measurement error, reliability and validity. There are two forms of measurement models:
formative (emerging) (from indicator to construct) and does not need to correlate and
reflective (latent) (from construct to indicator), likely to correlate.
