Research & Statistics summary 2020 MIM Tielens
Exam preparations
• All questions are multiple choice questions (~ 50 questions)
• Bring your calculator
– Lap top / SPSS is not needed
– No complex calculations
• Types of questions
– As interim tests
– BUT also interpretation of output tables (as in assignments)
– AND includes Research Methods (lectures Weeks 1, 2 and 7)
• Contents – what you need to know
– Slides of lectures define the contents
,Research & Statistics summary 2020 MIM Tielens
Lecture 1 + 2 – research methods (11/11/2020)
1.Problem formulation
Research objective
- ‘why’ is the research needed
- for ‘what’ purpose will the research be used
- how to act?
Research questions
- always about knowledge
- always a question
- accompanied with ≥ 2 sub questions
Types of research questions;
1. Frequency questions (descriptive analysis)
- How many
- How often
pie diagram / bar diagram / frequency tape / Histogram
2. Difference questions (two variables)
- is there a difference between…
boxplots / distributions
3. Relation questions
- is there a relation between…
- two types of relations;
1. Positive correlations
high scores go together with high scores
2. Negative correlations
high scores go together with low scores
with the use of scatterplots the type of correlation can be determined
2.Conceptual model
The conceptual model answers;
- what ; which variables must be considered
- who ; are variables related
- function ; helps to formulate hypotheses
Requirements for making a conceptual model
1. All variables should have the same unit of analysis (the blocks)
2. All links (arrows) in the model represent causal relationships
~ X happens earlier in time and has influence on Y
Correspondence between research question & the conceptual model
- dependent variables correspond to the subject of the main questions
- every relationship corresponds to a sub question
, Research & Statistics summary 2020 MIM Tielens
3.Research design
Causality
- if X changes Y changes too
X = cause, independent variable, used to predict
Y= effect, dependent variable, used to explain or predict
- X should be earlier in time than Y
- spurious relationship
= a correlation between X and Y can be explained by another variable: Z
Conditions for causality
1. Statistical correlation
- can be measured
2. Cause before effect
- time dependence can be a problem
3. No spurious relationships
- measure all potential alternative causes and test if correlation still exists
4. Theory; causal mechanism
- how does the cause generate the effect?
4. Types of Research design
1. Experiment
- typically 50% gets treatment and 50% does not
- high control rate by the researcher
= useful to rule out alternative explanations
- used to test a hypothesis measuring the effect of the treatment
Main characteristics of an experiment
- control on independent variables
- randomization
- before and after measurement
- control group
2. Survey
- typically a large group of correspondents complete a survey
- less control by the researcher
~ ruling out alternatives needs to be done with advanced statistics
- used for explorative and descriptive research
Main characteristics of a survey
- systematic interviewing or observation
- one moment measurement
- large numbers ; makes it reliable but also difficult due to many variables
When to choose for a survey
- with predictors that cannot be controlled (age, gender)
- with non-observable variables (motives, attributes, opinions, needs, wishes etc.)
Exam preparations
• All questions are multiple choice questions (~ 50 questions)
• Bring your calculator
– Lap top / SPSS is not needed
– No complex calculations
• Types of questions
– As interim tests
– BUT also interpretation of output tables (as in assignments)
– AND includes Research Methods (lectures Weeks 1, 2 and 7)
• Contents – what you need to know
– Slides of lectures define the contents
,Research & Statistics summary 2020 MIM Tielens
Lecture 1 + 2 – research methods (11/11/2020)
1.Problem formulation
Research objective
- ‘why’ is the research needed
- for ‘what’ purpose will the research be used
- how to act?
Research questions
- always about knowledge
- always a question
- accompanied with ≥ 2 sub questions
Types of research questions;
1. Frequency questions (descriptive analysis)
- How many
- How often
pie diagram / bar diagram / frequency tape / Histogram
2. Difference questions (two variables)
- is there a difference between…
boxplots / distributions
3. Relation questions
- is there a relation between…
- two types of relations;
1. Positive correlations
high scores go together with high scores
2. Negative correlations
high scores go together with low scores
with the use of scatterplots the type of correlation can be determined
2.Conceptual model
The conceptual model answers;
- what ; which variables must be considered
- who ; are variables related
- function ; helps to formulate hypotheses
Requirements for making a conceptual model
1. All variables should have the same unit of analysis (the blocks)
2. All links (arrows) in the model represent causal relationships
~ X happens earlier in time and has influence on Y
Correspondence between research question & the conceptual model
- dependent variables correspond to the subject of the main questions
- every relationship corresponds to a sub question
, Research & Statistics summary 2020 MIM Tielens
3.Research design
Causality
- if X changes Y changes too
X = cause, independent variable, used to predict
Y= effect, dependent variable, used to explain or predict
- X should be earlier in time than Y
- spurious relationship
= a correlation between X and Y can be explained by another variable: Z
Conditions for causality
1. Statistical correlation
- can be measured
2. Cause before effect
- time dependence can be a problem
3. No spurious relationships
- measure all potential alternative causes and test if correlation still exists
4. Theory; causal mechanism
- how does the cause generate the effect?
4. Types of Research design
1. Experiment
- typically 50% gets treatment and 50% does not
- high control rate by the researcher
= useful to rule out alternative explanations
- used to test a hypothesis measuring the effect of the treatment
Main characteristics of an experiment
- control on independent variables
- randomization
- before and after measurement
- control group
2. Survey
- typically a large group of correspondents complete a survey
- less control by the researcher
~ ruling out alternatives needs to be done with advanced statistics
- used for explorative and descriptive research
Main characteristics of a survey
- systematic interviewing or observation
- one moment measurement
- large numbers ; makes it reliable but also difficult due to many variables
When to choose for a survey
- with predictors that cannot be controlled (age, gender)
- with non-observable variables (motives, attributes, opinions, needs, wishes etc.)
