2025
Lecture summaries
‘Quantitative and
Design Methods in
Business Research’
QUANTITATIVE AND DESIGN
METHODS IN BUSINESS RESEARCH
FIDDER, B.R. (BASTIAAN, STUDENT M-BA)
, 1
Table of contents
Lecture 1: Introduction session and video ...................................................................................... 2
Theory of lecture .............................................................................................................................. 2
Video on Design & Behavioural Research (see Canvas) ...................................................................... 2
Lecture 2: Exploratory factor analysis (EFA) and principal component analysis (PCA)....................... 3
Lecture 3: Regression analysis ...................................................................................................... 5
Lecture 4: Analysis of (co-)variance .............................................................................................. 10
Lecture 5: Structural equation modelling ...................................................................................... 13
, 2
Lecture 1: Introduction session and video
Theory of lecture
Multivariate analysis comprises all statistical methods that simultaneously analyse multiple
measurements.
Measurement scales can be metric (…) or non-metric (…), this is important when deciding
on for an appropriate multivariate technique.
Measurement error: all variables have some error; these make the multivariate techniques less
powerful. Error term, random error or systematic error (epsilon) is used to describe the error
within a certain linear equation. Two important characteristics of measurement: reliability
(precision of observation) and validity (accuracy of measurement).
Statistical significance and power
• Type 1 error: you say there’s NO effect, but there is (wrongly reject null hypothesis,
falsely positive)
• Type 2 error: you say there’s an effect, but there isn’t (failing to reject null hypothesis
when you should’ve done so, falsely negative)
• Power: probability of rejecting the null hypothesis when it is false.
Dependence techniques identify dependent variables that are explained or predicted by a
set of independent variables (e.g. multiple regression, ANOVA, SEM). Interdependence
techniques simultaneously analyse all variables in the set without this distinction (e.g. EFA,
PCA…).
See slide 28-41 for explanation of each technique.
Video on Design & Behavioural Research (see Canvas)
, 3
Lecture 2: Exploratory factor analysis (EFA) and
principal component analysis (PCA)
Both EFA and PCA are interdependent methods, no distinction between independent and
dependent variables. PCA is a design technique, using existing data/variables in order to reduce
number of variables. Main goal of EFA is to identify underlying dimensions that explain
correlations among the variables.
EFA = technique within factor analysis to discover underlying relationships between (large set)
measured variables.
PCA = data reduction technique to explain variance in the dataset
Steps in EFA and PCA
1. First step is the selection of variables, e.g. based on theory, past research or judgement
of researcher.
2. Data must fulfil certain requirements, e.g. sample size, no missing data (…). Outliers
heavily influence outcomes, so they need to be considered. Factorability means that
variables are somewhat correlated and can be tested using Bartlett test of sphericity.
Kaiser-Meyer-Olkin criterion should always be 0.5 as a base minimum.
3. Extraction method: PCA and EFA are examples of extraction methods. PCA is a data
reduction technique which is used to reduce variables. Rotation method steps of PCA:
a. Standardize the observations.
b. Random lines can be drawn. After this, the space between observations and the
line should be reduced by turning this line (P1) as shown in figure 2.
c. Lastly, a second line (= component) is drawn which is the opposite of P1,
resulting in P2.
d. Components can be summarized in a table or linear combination.
4. Determine number of factors using one of the methods: prior knowledge, eigenvalues,
percentage of values (…). Scree plot is also one of the methods, which uses eigenvalues
and factors to plot a graph.
5. Factor rotation helps with the interpretation of factors. There are two types of factor
rotation: orthogonal and oblique rotation. These and its examples shown in figure 3.
Lecture summaries
‘Quantitative and
Design Methods in
Business Research’
QUANTITATIVE AND DESIGN
METHODS IN BUSINESS RESEARCH
FIDDER, B.R. (BASTIAAN, STUDENT M-BA)
, 1
Table of contents
Lecture 1: Introduction session and video ...................................................................................... 2
Theory of lecture .............................................................................................................................. 2
Video on Design & Behavioural Research (see Canvas) ...................................................................... 2
Lecture 2: Exploratory factor analysis (EFA) and principal component analysis (PCA)....................... 3
Lecture 3: Regression analysis ...................................................................................................... 5
Lecture 4: Analysis of (co-)variance .............................................................................................. 10
Lecture 5: Structural equation modelling ...................................................................................... 13
, 2
Lecture 1: Introduction session and video
Theory of lecture
Multivariate analysis comprises all statistical methods that simultaneously analyse multiple
measurements.
Measurement scales can be metric (…) or non-metric (…), this is important when deciding
on for an appropriate multivariate technique.
Measurement error: all variables have some error; these make the multivariate techniques less
powerful. Error term, random error or systematic error (epsilon) is used to describe the error
within a certain linear equation. Two important characteristics of measurement: reliability
(precision of observation) and validity (accuracy of measurement).
Statistical significance and power
• Type 1 error: you say there’s NO effect, but there is (wrongly reject null hypothesis,
falsely positive)
• Type 2 error: you say there’s an effect, but there isn’t (failing to reject null hypothesis
when you should’ve done so, falsely negative)
• Power: probability of rejecting the null hypothesis when it is false.
Dependence techniques identify dependent variables that are explained or predicted by a
set of independent variables (e.g. multiple regression, ANOVA, SEM). Interdependence
techniques simultaneously analyse all variables in the set without this distinction (e.g. EFA,
PCA…).
See slide 28-41 for explanation of each technique.
Video on Design & Behavioural Research (see Canvas)
, 3
Lecture 2: Exploratory factor analysis (EFA) and
principal component analysis (PCA)
Both EFA and PCA are interdependent methods, no distinction between independent and
dependent variables. PCA is a design technique, using existing data/variables in order to reduce
number of variables. Main goal of EFA is to identify underlying dimensions that explain
correlations among the variables.
EFA = technique within factor analysis to discover underlying relationships between (large set)
measured variables.
PCA = data reduction technique to explain variance in the dataset
Steps in EFA and PCA
1. First step is the selection of variables, e.g. based on theory, past research or judgement
of researcher.
2. Data must fulfil certain requirements, e.g. sample size, no missing data (…). Outliers
heavily influence outcomes, so they need to be considered. Factorability means that
variables are somewhat correlated and can be tested using Bartlett test of sphericity.
Kaiser-Meyer-Olkin criterion should always be 0.5 as a base minimum.
3. Extraction method: PCA and EFA are examples of extraction methods. PCA is a data
reduction technique which is used to reduce variables. Rotation method steps of PCA:
a. Standardize the observations.
b. Random lines can be drawn. After this, the space between observations and the
line should be reduced by turning this line (P1) as shown in figure 2.
c. Lastly, a second line (= component) is drawn which is the opposite of P1,
resulting in P2.
d. Components can be summarized in a table or linear combination.
4. Determine number of factors using one of the methods: prior knowledge, eigenvalues,
percentage of values (…). Scree plot is also one of the methods, which uses eigenvalues
and factors to plot a graph.
5. Factor rotation helps with the interpretation of factors. There are two types of factor
rotation: orthogonal and oblique rotation. These and its examples shown in figure 3.
