TOE – Correlational research
Lecture 1: Surveys
KOM recap
Experimental research
- Researcher manipulation
- Experimental and control group
o Often in lab
o Randomization of respondents into group
- Quantitative measures
- Suitable for Causal research
Qualitative research
- Study people in natural environment (not in de lab)
- Holistic approach
- Interviews, text analysis, focus groups
Correlational research
- Quantitative data
- Relations between variables
o Relations between education levels, age, income etc.
o We don’t control the variables (you can’t say these people receive
healthcare, and these people do not = not ethical)
- Causal study difficult
Correlational data is everywhere
- CBS, hoe blij klanten zijn met de service (blij poppetje of boos poppetje)
- Also worldwide
- Collecting data actively: surveys on your phone, how many steps you walk
everyday
- The use of digital
Quantitive data
Designed data: purpose designed by the researcher
- Costummade
Organic data: the data is less outthere, i have not designed it but i use it for my
research
- Aspiritional: facebook, twitter, instagram
- Transictional: bonuscart or mastercard
- It is not possible for everyone but you analyse the group you want to do
research on (the group which uses instagram)
- Readymade
,Correlational Data (designed)
We design a study & collect data to:
- describe the social reality
- study causal relationships
- generalize to the target population
Inferential goals:
- Description beschrijven
- Causation relatie / correlatie
- Prediction voorspellen
EU Summer Time Arrangements
- Have been in place since the 1980-s
- Uncoordinated timechanges detrimental for internal market
- Energy savings are marginal
- Positive for health: more outdoor activities but might affect biorhythm
- Road safety: inconclusive evidence
- Extended working hours for agriculture
- European Commission regularly receives feedback from citizens
- Are current summertime arrangements working?
Research cycle KOM:
Survey Lifecycle:
, Total survey error framework dia 32
Representation:
Coverage error: research the population in the Netherlands; but not everyone has
an emailadres.
- Being able to reach the people who are part of my researchgroup
- Overcoverd: people who have multiple telephone numbers
- Undercoverd: people who don’t have phone numbers
Sampling error:
Nonresponse error: people who not respont at surveys
Adjustment error:
Measurement:
Measurement error: alcohol glasses, the respondent who has aldready drunk will
not tell it to the interviewer
Processing error:
Coverage
Coverage error occurs...
- if not all members of the population have a known, nonzero chance of
being included in the sample, and
o Known, nonzero: if i have a list, that is nonzero chance
o People who don’t use internet chance is zero of being on the list
o If everyone is the same; no worry about covered and noncovered.
- if persons included in the sample differ from those excluded.
- Example dia 35
Coverage error
- Target population: the finite population we want to study
o Can members of the target population be reached? Are they
different from those who cannot be reached on statistics in
question?
o Frame population: all persons who have a chance to be included into
the sample
(e.g., a list, area sampling)
o Problem: target population ↔ available frame population
Sampling
Sampling error occurs...
- from surveying only some, rather than all, members of the covered
population.
- Uncertainty
- We need statistics to quantify this uncertainty
De groene groep willen we niet bestuderen, dus het groene deel dat overlapt is
door het blauwe gedeelte moet weg.
Lecture 1: Surveys
KOM recap
Experimental research
- Researcher manipulation
- Experimental and control group
o Often in lab
o Randomization of respondents into group
- Quantitative measures
- Suitable for Causal research
Qualitative research
- Study people in natural environment (not in de lab)
- Holistic approach
- Interviews, text analysis, focus groups
Correlational research
- Quantitative data
- Relations between variables
o Relations between education levels, age, income etc.
o We don’t control the variables (you can’t say these people receive
healthcare, and these people do not = not ethical)
- Causal study difficult
Correlational data is everywhere
- CBS, hoe blij klanten zijn met de service (blij poppetje of boos poppetje)
- Also worldwide
- Collecting data actively: surveys on your phone, how many steps you walk
everyday
- The use of digital
Quantitive data
Designed data: purpose designed by the researcher
- Costummade
Organic data: the data is less outthere, i have not designed it but i use it for my
research
- Aspiritional: facebook, twitter, instagram
- Transictional: bonuscart or mastercard
- It is not possible for everyone but you analyse the group you want to do
research on (the group which uses instagram)
- Readymade
,Correlational Data (designed)
We design a study & collect data to:
- describe the social reality
- study causal relationships
- generalize to the target population
Inferential goals:
- Description beschrijven
- Causation relatie / correlatie
- Prediction voorspellen
EU Summer Time Arrangements
- Have been in place since the 1980-s
- Uncoordinated timechanges detrimental for internal market
- Energy savings are marginal
- Positive for health: more outdoor activities but might affect biorhythm
- Road safety: inconclusive evidence
- Extended working hours for agriculture
- European Commission regularly receives feedback from citizens
- Are current summertime arrangements working?
Research cycle KOM:
Survey Lifecycle:
, Total survey error framework dia 32
Representation:
Coverage error: research the population in the Netherlands; but not everyone has
an emailadres.
- Being able to reach the people who are part of my researchgroup
- Overcoverd: people who have multiple telephone numbers
- Undercoverd: people who don’t have phone numbers
Sampling error:
Nonresponse error: people who not respont at surveys
Adjustment error:
Measurement:
Measurement error: alcohol glasses, the respondent who has aldready drunk will
not tell it to the interviewer
Processing error:
Coverage
Coverage error occurs...
- if not all members of the population have a known, nonzero chance of
being included in the sample, and
o Known, nonzero: if i have a list, that is nonzero chance
o People who don’t use internet chance is zero of being on the list
o If everyone is the same; no worry about covered and noncovered.
- if persons included in the sample differ from those excluded.
- Example dia 35
Coverage error
- Target population: the finite population we want to study
o Can members of the target population be reached? Are they
different from those who cannot be reached on statistics in
question?
o Frame population: all persons who have a chance to be included into
the sample
(e.g., a list, area sampling)
o Problem: target population ↔ available frame population
Sampling
Sampling error occurs...
- from surveying only some, rather than all, members of the covered
population.
- Uncertainty
- We need statistics to quantify this uncertainty
De groene groep willen we niet bestuderen, dus het groene deel dat overlapt is
door het blauwe gedeelte moet weg.
