Statistics
Week 1
Types of data analysis
Quantitative methods
- Testing theories using numbers – describing and analysing data using statistics
Qualitative methods
- Testing theories using language
- magazines/articles/interviews
- conversations
- newspapers
- media broadcast
We do research on an sample taken from the population. These results do not necessarily
apply to everyone. Statistics is part of the research process. Statistics will not help if the study
is badly designed.
We use statistics so that we can generalise our findings.
- How big is the chance that our sample is representative for the population?
5 steps for setting up a research
1. You find something that needs explaining
2. Hypotheses based on expectations (from literature, a theory or an observation)
3. Collect data
4. Analyse data
5. Draw conclusions
Falsification: The act of disproving a theory or hypotheses. One falsification has more impact
than numerous confirmations. It is easier to prove that a hypothesis can be rejected.
We need to define the variables in order to generate a hypothesis and analyse data. E.g.
- Type of ad; with and without certain words
- Product colour: products with one, two or three colours
- Country: people living in one or two countries
1
,Variable: anything that can be measured and can differ across entities or time. You are
interested in the relation between theoretical constructs.
- Effect of popularity on success
Popularity is based on number of friends
Success is based on total amount of money
An independent variable (IV)
- Proposed cause
- A predictor variable, influences the results
- A manipulated variable by the researches (I’m
manipulating the colour of the product)
A dependent variable (DV)
- The variable that shows the result or outcome, depending
on the intervention
- The proposed effect
- Measured not manipulated (does the colour of the product (IV) have an effect on
buying intention (DV))
Example:
A study using Dutch-accented and German-accented speakers of English to study the effect
of type of English accent on speaker appreciation.
IV: Type of accent (German or Dutch)
DV: Speaker appreciation
Control variable: I want to keep the age constant so all the participants are 20 years old.
Examples IV’s:
- Level of education
- Groups (with or without training)
- Age
- Type of ad (with or without gestures)
Levels of measurement
NOIR
- Nominal
- Ordinal
- Interval
- Ratio
2
, Categorical (entities are divided into distinct categories)
- Binary variable: only two categories (dead or alive)
- Nominal variable: more than two categories (omnivore, vegetarian, vegan or fruitarian)
- Ordinal variable: the same as nominal, but we can order the categories (low income,
middle income, high income)
Continuous (entities get a distinct score)
- Interval variable: equal intervals on the variable represent equal differences (the difference
between 6 and 8 is the same as the difference between 13 and 15) e.g. likert scale
- Ratio variable: The must be an absolute zero point (someone earning 16 euro earns twice
as much as someone who earns 8, but someone who earns 0 euro earns nothing) e.g.
percentage of questions right after an exam
Measurement error
The discrepancy between the actual value we’re trying to measure and the number we use
to represent that value. We try to keep measurement error low by using valid and reliable
measures.
Validity: whether an instrument measure what is set out to measure.
No need to know this by
heart.
3
Week 1
Types of data analysis
Quantitative methods
- Testing theories using numbers – describing and analysing data using statistics
Qualitative methods
- Testing theories using language
- magazines/articles/interviews
- conversations
- newspapers
- media broadcast
We do research on an sample taken from the population. These results do not necessarily
apply to everyone. Statistics is part of the research process. Statistics will not help if the study
is badly designed.
We use statistics so that we can generalise our findings.
- How big is the chance that our sample is representative for the population?
5 steps for setting up a research
1. You find something that needs explaining
2. Hypotheses based on expectations (from literature, a theory or an observation)
3. Collect data
4. Analyse data
5. Draw conclusions
Falsification: The act of disproving a theory or hypotheses. One falsification has more impact
than numerous confirmations. It is easier to prove that a hypothesis can be rejected.
We need to define the variables in order to generate a hypothesis and analyse data. E.g.
- Type of ad; with and without certain words
- Product colour: products with one, two or three colours
- Country: people living in one or two countries
1
,Variable: anything that can be measured and can differ across entities or time. You are
interested in the relation between theoretical constructs.
- Effect of popularity on success
Popularity is based on number of friends
Success is based on total amount of money
An independent variable (IV)
- Proposed cause
- A predictor variable, influences the results
- A manipulated variable by the researches (I’m
manipulating the colour of the product)
A dependent variable (DV)
- The variable that shows the result or outcome, depending
on the intervention
- The proposed effect
- Measured not manipulated (does the colour of the product (IV) have an effect on
buying intention (DV))
Example:
A study using Dutch-accented and German-accented speakers of English to study the effect
of type of English accent on speaker appreciation.
IV: Type of accent (German or Dutch)
DV: Speaker appreciation
Control variable: I want to keep the age constant so all the participants are 20 years old.
Examples IV’s:
- Level of education
- Groups (with or without training)
- Age
- Type of ad (with or without gestures)
Levels of measurement
NOIR
- Nominal
- Ordinal
- Interval
- Ratio
2
, Categorical (entities are divided into distinct categories)
- Binary variable: only two categories (dead or alive)
- Nominal variable: more than two categories (omnivore, vegetarian, vegan or fruitarian)
- Ordinal variable: the same as nominal, but we can order the categories (low income,
middle income, high income)
Continuous (entities get a distinct score)
- Interval variable: equal intervals on the variable represent equal differences (the difference
between 6 and 8 is the same as the difference between 13 and 15) e.g. likert scale
- Ratio variable: The must be an absolute zero point (someone earning 16 euro earns twice
as much as someone who earns 8, but someone who earns 0 euro earns nothing) e.g.
percentage of questions right after an exam
Measurement error
The discrepancy between the actual value we’re trying to measure and the number we use
to represent that value. We try to keep measurement error low by using valid and reliable
measures.
Validity: whether an instrument measure what is set out to measure.
No need to know this by
heart.
3
