RMDS Summary test 1
Unit 1: empirical research
Empirical research: reach with data
Empirical questions: systematically answering empirical
Excluding that others have a better answer
About things we can observe.
Wheel of science → important for test
Deduction = Theory and how to test the theory
Induction = Data and trying to arrive conclusions
Need a problem analysis: how big is the problem?
Important for test:
Ex ante evaluation (previous) evaluation of options → can we expect an option to work? Program
monitoring
Ex post evaluation (effect / impact research): did the selected option have the expected outcome?
(outcome evaluation)
Confirmation bias (myside bias = cherry picking)
Confirmation bias = search for analyses and recall information in a way that confirms pre-existing
beliefs, while giving disproportionately less consideration to alternative interpretations.
Affects what we think is true
Like crime: looking for evidence of a suspect and not looking further.
Consequences → mistake knowledge + bad decisions → avoid by procedure
Procedure: think, plan, observe, analyse, conclude
Wheel of science: is importat logic, it is not logistical / empirical cycle
,Type of research → important for test
1. Problem definition and -analysis: identify, describe and
analyze a problem/opportunity
2. Design: develop options/design options
3. Multi-criteria analyis/ex ante (van te voren) analysis:
compare options, using criteria
4. Decision-making rules: select a preferred alternative on
the basis of evaluation
5. Implementation/proces: act on the decision/implement
the decision
6. Ex post (achteraf)/outcome: monitor and evaluate the
consequences
Unit 2: clear research question
An empirical research question is only clear if it refers to meaningful
- Units of analysis: who or what is the question about
- Variables: possible characteristics (attributes) of these units?
- Settings: time and place
Three types of RQ
Normative questions Describing which it should be, is justifiable, cannot be answered by
observation only, not asking for legal facts
Conceptual What does it mean? Often starts with “what is…”, cannot be answerd using
questions observations
Empirical questions only and fully be answered using observations
- Descriptive questions: NOT about cause and effects
- Explanatory questions: cause and an affect, often 2 variables. Testing
causal relationships
How to’ questions can be ‘broken up’ into descriptive and explanatory (causal) research questions.
Research design
Unit 3 what are data
UNITS
Two aspects of observation
Research questions are about theoretical variables, describing units
- Theoretical variable(s) → Conceptualization, operationalization, measurement → data
- Unit(s) → sampling → data
Units of analysis: couples → unit mention in the RQ
Units of observation: two people → actual unit which is observed
Ecological fallacy: Drawing conclusions about lower level units solely on the basis of aggregate data.
Mixing up UOO and UOA
VARIABLES
Variable: is a complete and mutually exclusive set of attributes or values. Used to describe units.
Example: age (of a human being) has attributes or values between 0 and 120
Attributes or values
The words value and attributes refer to the same thing however
- Values is more often used to refer to numerical attributes (age or weight)
- Attributes is more often used to refer to ‘non-numerical’ attributes (colors, relegions)
Attributes have numbers in SPSS.
, 5 levels of measurement
SPSS Levels Explanation Example
Nominal Dichotomy If a variable has only two attributes 0-18 or >18
Nominal attributes with no numerical value, no Gender
order
Ordinal Ordinal If values can be ordered, distance is Age in categories 0-18,
unknown. 19-35,36-64, >65
Scale Ratio If the values can be ordered, the Age of a person, weight,
distance between the values is income,
known. Has a meaningful zero point.
(you can say, twice as much)
Interval Scale where the order and the IQ, Celsius temperature
difference between the variables is
known. Equally spaced values. No
zero point
Mutually exclusive: every unit should have only one value of a variable
• Are you 20-50 or 50-70? → 50 is in two categories → wrong
Exhaustive: every unit should have a value of the variable
• Do you love women or men? → both is not possible → wrong
SPSS
- Respondent didn’t want to answer: 998
- Question not asked – filter: 999
Consequences of missing values
- Filter questions: no loss of information
- Does not know: sometimes loss of information - bias
- Refusals (weigert), mistakes: often loss of information, bias = invalidity
→ can give different averages, graphs etc.
Labeling variables
A codebook describes: the meaning of variables and values used in the data matrix.
- Variable name (short): work or workYN
- Variable label (clear): meaning of variable (V1 = is respondent working or not?)
- Value label: meaning of the value (0= no work; 1 = work; 999 = no answer)
Unit 4
Create 1 syntax file for analyzing a data matrix in a project.
A syntax file shows all your actions in SPSS so is
transparent towards others.
Always click on PASTE to put it into the syntax file.
Want to change the variable? → create a new variable by
recoding the other one = transforming
Transforming/ recoding:
SPSS → recode into different variables
Unit 5
Four types of relationships between terms(concept) & facets (dimensions)
Unit 1: empirical research
Empirical research: reach with data
Empirical questions: systematically answering empirical
Excluding that others have a better answer
About things we can observe.
Wheel of science → important for test
Deduction = Theory and how to test the theory
Induction = Data and trying to arrive conclusions
Need a problem analysis: how big is the problem?
Important for test:
Ex ante evaluation (previous) evaluation of options → can we expect an option to work? Program
monitoring
Ex post evaluation (effect / impact research): did the selected option have the expected outcome?
(outcome evaluation)
Confirmation bias (myside bias = cherry picking)
Confirmation bias = search for analyses and recall information in a way that confirms pre-existing
beliefs, while giving disproportionately less consideration to alternative interpretations.
Affects what we think is true
Like crime: looking for evidence of a suspect and not looking further.
Consequences → mistake knowledge + bad decisions → avoid by procedure
Procedure: think, plan, observe, analyse, conclude
Wheel of science: is importat logic, it is not logistical / empirical cycle
,Type of research → important for test
1. Problem definition and -analysis: identify, describe and
analyze a problem/opportunity
2. Design: develop options/design options
3. Multi-criteria analyis/ex ante (van te voren) analysis:
compare options, using criteria
4. Decision-making rules: select a preferred alternative on
the basis of evaluation
5. Implementation/proces: act on the decision/implement
the decision
6. Ex post (achteraf)/outcome: monitor and evaluate the
consequences
Unit 2: clear research question
An empirical research question is only clear if it refers to meaningful
- Units of analysis: who or what is the question about
- Variables: possible characteristics (attributes) of these units?
- Settings: time and place
Three types of RQ
Normative questions Describing which it should be, is justifiable, cannot be answered by
observation only, not asking for legal facts
Conceptual What does it mean? Often starts with “what is…”, cannot be answerd using
questions observations
Empirical questions only and fully be answered using observations
- Descriptive questions: NOT about cause and effects
- Explanatory questions: cause and an affect, often 2 variables. Testing
causal relationships
How to’ questions can be ‘broken up’ into descriptive and explanatory (causal) research questions.
Research design
Unit 3 what are data
UNITS
Two aspects of observation
Research questions are about theoretical variables, describing units
- Theoretical variable(s) → Conceptualization, operationalization, measurement → data
- Unit(s) → sampling → data
Units of analysis: couples → unit mention in the RQ
Units of observation: two people → actual unit which is observed
Ecological fallacy: Drawing conclusions about lower level units solely on the basis of aggregate data.
Mixing up UOO and UOA
VARIABLES
Variable: is a complete and mutually exclusive set of attributes or values. Used to describe units.
Example: age (of a human being) has attributes or values between 0 and 120
Attributes or values
The words value and attributes refer to the same thing however
- Values is more often used to refer to numerical attributes (age or weight)
- Attributes is more often used to refer to ‘non-numerical’ attributes (colors, relegions)
Attributes have numbers in SPSS.
, 5 levels of measurement
SPSS Levels Explanation Example
Nominal Dichotomy If a variable has only two attributes 0-18 or >18
Nominal attributes with no numerical value, no Gender
order
Ordinal Ordinal If values can be ordered, distance is Age in categories 0-18,
unknown. 19-35,36-64, >65
Scale Ratio If the values can be ordered, the Age of a person, weight,
distance between the values is income,
known. Has a meaningful zero point.
(you can say, twice as much)
Interval Scale where the order and the IQ, Celsius temperature
difference between the variables is
known. Equally spaced values. No
zero point
Mutually exclusive: every unit should have only one value of a variable
• Are you 20-50 or 50-70? → 50 is in two categories → wrong
Exhaustive: every unit should have a value of the variable
• Do you love women or men? → both is not possible → wrong
SPSS
- Respondent didn’t want to answer: 998
- Question not asked – filter: 999
Consequences of missing values
- Filter questions: no loss of information
- Does not know: sometimes loss of information - bias
- Refusals (weigert), mistakes: often loss of information, bias = invalidity
→ can give different averages, graphs etc.
Labeling variables
A codebook describes: the meaning of variables and values used in the data matrix.
- Variable name (short): work or workYN
- Variable label (clear): meaning of variable (V1 = is respondent working or not?)
- Value label: meaning of the value (0= no work; 1 = work; 999 = no answer)
Unit 4
Create 1 syntax file for analyzing a data matrix in a project.
A syntax file shows all your actions in SPSS so is
transparent towards others.
Always click on PASTE to put it into the syntax file.
Want to change the variable? → create a new variable by
recoding the other one = transforming
Transforming/ recoding:
SPSS → recode into different variables
Unit 5
Four types of relationships between terms(concept) & facets (dimensions)
