Lecture 1 : Research Methodology
Qualitative = about the experience of people
Quantitative = finding hard data, define it in numbers
First steps in a research project:
1) Problem
2) Research question
3) Literature study
4) Conceptual model
5) Propositions/ hypothesis
2) The research question should clearly state what the concepts are that you want to study
and what the basis is for your project.
4) Conceptual definition theoretical (like in a dictionary)
Operational definition practical/ concrete (makes the concept measurable)
Indicators descriptions of different aspects that are relevant to the operational definition
Potential issues:
- Validity (do I measure what I want to measure)
- Reliability
- Respondent error
Internal validity = The ability of a research instrument to measure what you want to
measure.
External validity = The extent to which findings can meaningfully be generalized to other
settings.
Correlations indicate the strength of a relationship between concepts
NOT an indicator of causality!
Lecture 2 : Data types
Data types = the description of what a particular number means
Data types determine which statistical technique you can choose to analyze your data!
Nominal Data
Numbers indicate categories
Order is not meaningful
Example: what is your religion? (Christianity (1) or Islam (2))
No order, so Christianity is not better than Islam.
Example: Dichotomous choices such as Yes/No or Male/Female
, Ordinal Data
There is a certain order in the labeling of numbers
Intervals between values are not equal
Example: Ordering a group of people by length and numbering them from 1 to 10
A 6 tells me that he/she is longer than a 4
The increase in length does not equal for each step on your scale
Interval Data
Intervals are equal (basic calculations possible)
Do not have a fixed 0 value
Example: Temperature
20 degrees is one degree warmer than 19 degree etc.
Depends on the scale you use, Celsius versus Fahrenheit
Ratio Data
Intervals are equal (basic calculations possible)
A fixed 0 value (0 means nothing)
Example: measuring length in centimeters
30 centimeters is one centimeter longer than 29 centimeter
Zero always means that you do not have any length at all
From interval/ratio to lower data type: no problem (HOWEVER: you lose information)
Ages from 1-100 reduces to 3 categories
From a lower level to higher data type: not possible!
THEREFORE: try to use a “high” a data type as possible when creating measurement scales
Symmetrical scale (perfect antonyms) = Treat as interval (no reason to assume intervals are
not equal)
Asymmetric scale = Treat as ordinal (interval size cannot be estimated properly)
Lecture 2 : Survey Questions & Response Scales
Qualitative = about the experience of people
Quantitative = finding hard data, define it in numbers
First steps in a research project:
1) Problem
2) Research question
3) Literature study
4) Conceptual model
5) Propositions/ hypothesis
2) The research question should clearly state what the concepts are that you want to study
and what the basis is for your project.
4) Conceptual definition theoretical (like in a dictionary)
Operational definition practical/ concrete (makes the concept measurable)
Indicators descriptions of different aspects that are relevant to the operational definition
Potential issues:
- Validity (do I measure what I want to measure)
- Reliability
- Respondent error
Internal validity = The ability of a research instrument to measure what you want to
measure.
External validity = The extent to which findings can meaningfully be generalized to other
settings.
Correlations indicate the strength of a relationship between concepts
NOT an indicator of causality!
Lecture 2 : Data types
Data types = the description of what a particular number means
Data types determine which statistical technique you can choose to analyze your data!
Nominal Data
Numbers indicate categories
Order is not meaningful
Example: what is your religion? (Christianity (1) or Islam (2))
No order, so Christianity is not better than Islam.
Example: Dichotomous choices such as Yes/No or Male/Female
, Ordinal Data
There is a certain order in the labeling of numbers
Intervals between values are not equal
Example: Ordering a group of people by length and numbering them from 1 to 10
A 6 tells me that he/she is longer than a 4
The increase in length does not equal for each step on your scale
Interval Data
Intervals are equal (basic calculations possible)
Do not have a fixed 0 value
Example: Temperature
20 degrees is one degree warmer than 19 degree etc.
Depends on the scale you use, Celsius versus Fahrenheit
Ratio Data
Intervals are equal (basic calculations possible)
A fixed 0 value (0 means nothing)
Example: measuring length in centimeters
30 centimeters is one centimeter longer than 29 centimeter
Zero always means that you do not have any length at all
From interval/ratio to lower data type: no problem (HOWEVER: you lose information)
Ages from 1-100 reduces to 3 categories
From a lower level to higher data type: not possible!
THEREFORE: try to use a “high” a data type as possible when creating measurement scales
Symmetrical scale (perfect antonyms) = Treat as interval (no reason to assume intervals are
not equal)
Asymmetric scale = Treat as ordinal (interval size cannot be estimated properly)
Lecture 2 : Survey Questions & Response Scales
