PRACTICE OF
EMPIRICAL
RESEARCH
Summary
,Inhoudsopgave
Lecture 1: from research problem to analysis ....................................................................................... 2
Lecture 2: data inspection......................................................................................................................... 6
Lecture 3: introduction to multiple regression .................................................................................... 14
Lecture 4: power, effect sizes and ethics ............................................................................................. 21
College 5: multiple regression in practice ............................................................................................ 28
Lecture 6: statistical toolkit..................................................................................................................... 36
Lecture 7: assumptions & data problems............................................................................................. 42
1
, Lecture 1: from research problem to analysis
This course prepares you for your master thesis and future career.
• Preparations: before sessions
• Allowed to google answers
• Material in bachelor should be enough
• Make mistakes, but always attempt
To pass the course…
1. Participate in all tutorials and computer labs
2. Review assignment (no grade)
3. End paper (grade 5.5 or higher)
Statistical techniques from the bachelor program
• Z-test
• T-test
• Correlation & regression
• ANOVA
• Non-parametric tests
Multivariate analysis: the basics
The arrow -> = means causality
Bivariate analysis: relation of two variables.
• X ->Y
Multivariate analysis: relations between multiple variables at a time.
• X1 & X2...? -> Y – multiple predicters and one outcome
• X -> Y1 & Y2...? – one predicter and more outcomes
• Both
More variables create complexity. Makes it more realistic, like the real world.
Complexity of research questions
• Univariate descriptive: one variable in the research question. Describes
data/sample.
• Bivariate: two variables in the research question. Different forms:
- Symmetric: don’t difference between predicter and outcome. Not
interested in a direction. Just want to know if there is a relation.
- Asymmetric: one is the predicter and the other one is the outcome. But
not always the case. Can be directional but not causal. Not able to prove
causality -> need an experimental technique. Don’t use the word
INFLUENCE when you can’t prove causality!
Ø Non-causal
Ø Causal
• Multivariate: more than two variables in the research question.
- (more variables = more complexity)
- Indirect effects, interactions
2
EMPIRICAL
RESEARCH
Summary
,Inhoudsopgave
Lecture 1: from research problem to analysis ....................................................................................... 2
Lecture 2: data inspection......................................................................................................................... 6
Lecture 3: introduction to multiple regression .................................................................................... 14
Lecture 4: power, effect sizes and ethics ............................................................................................. 21
College 5: multiple regression in practice ............................................................................................ 28
Lecture 6: statistical toolkit..................................................................................................................... 36
Lecture 7: assumptions & data problems............................................................................................. 42
1
, Lecture 1: from research problem to analysis
This course prepares you for your master thesis and future career.
• Preparations: before sessions
• Allowed to google answers
• Material in bachelor should be enough
• Make mistakes, but always attempt
To pass the course…
1. Participate in all tutorials and computer labs
2. Review assignment (no grade)
3. End paper (grade 5.5 or higher)
Statistical techniques from the bachelor program
• Z-test
• T-test
• Correlation & regression
• ANOVA
• Non-parametric tests
Multivariate analysis: the basics
The arrow -> = means causality
Bivariate analysis: relation of two variables.
• X ->Y
Multivariate analysis: relations between multiple variables at a time.
• X1 & X2...? -> Y – multiple predicters and one outcome
• X -> Y1 & Y2...? – one predicter and more outcomes
• Both
More variables create complexity. Makes it more realistic, like the real world.
Complexity of research questions
• Univariate descriptive: one variable in the research question. Describes
data/sample.
• Bivariate: two variables in the research question. Different forms:
- Symmetric: don’t difference between predicter and outcome. Not
interested in a direction. Just want to know if there is a relation.
- Asymmetric: one is the predicter and the other one is the outcome. But
not always the case. Can be directional but not causal. Not able to prove
causality -> need an experimental technique. Don’t use the word
INFLUENCE when you can’t prove causality!
Ø Non-causal
Ø Causal
• Multivariate: more than two variables in the research question.
- (more variables = more complexity)
- Indirect effects, interactions
2
