LES NOTES – SURVEY RESEARCH METHODS
LECTURE 1 – INTRODUCTION
Surveys
- Select the right items
o Past research
- Social desirability
- Large (sufficient) amount of respondents
o Dependent of the populaMon, very tricky oPen you don’t know the size of your
populaMon.
- RepresentaMve sample
- Not too long (5-10 minutes)
- Pretest with part of the sample and experts
§ Define the things you want to measure up front, ‘what is happiness’ make sure that
everybody has the same understanding of the construct you are measuring.
§ Depending on what you measure;
o Do you offer a middle opMon (neutral opMon)
o How many opMons 5/7/9
Survey vs Experimental research
§ Different groups
§ Same situaMons
§ ManipulaMons
§ Causality
§ Random assignment
Survey Research: Generates data, uses real life data.
Experimental Research: Generates data, uses manipulated data.
Modeling (real life data) or Meta Analyses (manipulated data) use exisMng data (not
generaMng new).
è But Survey Research can include other methods.
o Think of a survey with manipulated data (survey experiments) or a survey on
exisMng data.
Examples of Academic Survey Research
§ Scale development and validaMon
§ SegmentaMon and typology
§ RelaMonships between variables
§ Secondary data and experiments
What is survey research?
A survey is a systemaMc method for gathering informaMon with standardized quesMonnaires
from (a sample of) enMMes for the purpose of construcMng quanMtaMve descriptors of the
aeributes of the larger populaMon of which the enMMes are members. (Groves et al. 2009)
, § SystemaMc and Standardized
o DisMnguishes itself from other techniques such as qualitaMve research.
§ (A Sample of) EnMMes
o Mostly a selecMon of the populaMon is included in the survey.
o All enMMes can be included as well (e.g., employee survey).
§ QuanMtaMve Descriptors
o StaMsMcs needed to analyze the responses.
• Today, surveys provide managers with deeper insights into their customers and
employees.
• Surveys can contribute to generalizing experimental findings to different persons and
sehngs.
Survey PresidenMal ElecMons
§ Social desirability – people didn’t indicate their real choice.
§ Wrong sample.
Sampling Procedures of Survey Research
§ Probability sampling
o SelecMon of sampling techniques in which the chance of each case (e.g.,
individual) being selected from the populaMon is known and not zero.
§ RandomizaMon is used instead of deliberate choice.
• European Social Survey with a sampling frame and exact data on
background of all individuals in the sample
§ Nonprobability sampling
o SelecMon of sampling techniques in which the chance of each case (e.g.,
individual) being selected is NOT known.
§ The researcher deliberately picks the sample.
• Fellow students selected to parMcipate in a survey.
Time Horizon of Survey Research
§ Cross-SecMonal Design
o Opinion polls (e.g., to predict electoral results)
o Typical thesis survey (e.g., to understand consumer preferences or ahtudes)
§ Longitudinal Design
o Repeated cross-secMonal (different sample each Mme)
o Fixed-sample panel design (the same sample each Mme)
o Cohort study (follow the same people over a longer period of Mme, usually
aPer an event; e.g., people graduaMng MSc MarkeMng in 2010 and follow their
lives over a 10-year Mme period)
AdministraMon Methods of Survey Research
,Why so Popular? And What are the Challenges?
Main Pros
§ Ease: QuesMonnaires are relaMvely easy to administer to large number of respondents.
§ Simplicity: Coding, analysis, and interpretaMon of data is relaMvely simple.
§ Reliability: StandardizaMon reduces variability in the answers.
Challenges
§ Asking the right quesMons is not easy.
§ Properly wording quesMons is not easy.
§ Gehng the right sample (sampling) is not easy.
§ MoMvaMng respondents to respond (candidly) is not easy.
What is Survey Data good for?
The Four Phases of the Survey Research Process
1. Planning a Survey
o Define Research ObjecMves
o SelecMon of Research Variables
2. Designing a Survey
o ConceptualizaMon and OperaMonalizaMon
o Construct and Pretest QuesMonnaire
3. ImplemenMng a Survey
o Choose Relevant PopulaMon
o Sampling
o Data CollecMon
4. Gaining Insights from a Survey
o Data PreparaMon and Measurement EvaluaMon
o Data Analysis
o ReporMng of Results
è ConceptualizaMon and OperaMonalizaMon, Construct and Pretest QuesMonnaire,
Choosing relevant PopulaMon and sampling oPen at the same Mme.
Following pages gives an overview of the upcoming topics in the course:
, Analyzing a Survey (Topic Lecture 3)
To use regression analysis to infer causal relaMonships or for predicMon outside the data
sample, one must carefully jusMfy why there is a causal relaMonship between two variables or
why exisMng relaMonships have predicMve power for a new context!
Planning a Survey (Topic Lecture 4)
Designing a Survey: Measurement (Topic Lecture 5)
Designing a Survey: QuesMonnaire Design (Topic Lecture 6)
LECTURE 1 – INTRODUCTION
Surveys
- Select the right items
o Past research
- Social desirability
- Large (sufficient) amount of respondents
o Dependent of the populaMon, very tricky oPen you don’t know the size of your
populaMon.
- RepresentaMve sample
- Not too long (5-10 minutes)
- Pretest with part of the sample and experts
§ Define the things you want to measure up front, ‘what is happiness’ make sure that
everybody has the same understanding of the construct you are measuring.
§ Depending on what you measure;
o Do you offer a middle opMon (neutral opMon)
o How many opMons 5/7/9
Survey vs Experimental research
§ Different groups
§ Same situaMons
§ ManipulaMons
§ Causality
§ Random assignment
Survey Research: Generates data, uses real life data.
Experimental Research: Generates data, uses manipulated data.
Modeling (real life data) or Meta Analyses (manipulated data) use exisMng data (not
generaMng new).
è But Survey Research can include other methods.
o Think of a survey with manipulated data (survey experiments) or a survey on
exisMng data.
Examples of Academic Survey Research
§ Scale development and validaMon
§ SegmentaMon and typology
§ RelaMonships between variables
§ Secondary data and experiments
What is survey research?
A survey is a systemaMc method for gathering informaMon with standardized quesMonnaires
from (a sample of) enMMes for the purpose of construcMng quanMtaMve descriptors of the
aeributes of the larger populaMon of which the enMMes are members. (Groves et al. 2009)
, § SystemaMc and Standardized
o DisMnguishes itself from other techniques such as qualitaMve research.
§ (A Sample of) EnMMes
o Mostly a selecMon of the populaMon is included in the survey.
o All enMMes can be included as well (e.g., employee survey).
§ QuanMtaMve Descriptors
o StaMsMcs needed to analyze the responses.
• Today, surveys provide managers with deeper insights into their customers and
employees.
• Surveys can contribute to generalizing experimental findings to different persons and
sehngs.
Survey PresidenMal ElecMons
§ Social desirability – people didn’t indicate their real choice.
§ Wrong sample.
Sampling Procedures of Survey Research
§ Probability sampling
o SelecMon of sampling techniques in which the chance of each case (e.g.,
individual) being selected from the populaMon is known and not zero.
§ RandomizaMon is used instead of deliberate choice.
• European Social Survey with a sampling frame and exact data on
background of all individuals in the sample
§ Nonprobability sampling
o SelecMon of sampling techniques in which the chance of each case (e.g.,
individual) being selected is NOT known.
§ The researcher deliberately picks the sample.
• Fellow students selected to parMcipate in a survey.
Time Horizon of Survey Research
§ Cross-SecMonal Design
o Opinion polls (e.g., to predict electoral results)
o Typical thesis survey (e.g., to understand consumer preferences or ahtudes)
§ Longitudinal Design
o Repeated cross-secMonal (different sample each Mme)
o Fixed-sample panel design (the same sample each Mme)
o Cohort study (follow the same people over a longer period of Mme, usually
aPer an event; e.g., people graduaMng MSc MarkeMng in 2010 and follow their
lives over a 10-year Mme period)
AdministraMon Methods of Survey Research
,Why so Popular? And What are the Challenges?
Main Pros
§ Ease: QuesMonnaires are relaMvely easy to administer to large number of respondents.
§ Simplicity: Coding, analysis, and interpretaMon of data is relaMvely simple.
§ Reliability: StandardizaMon reduces variability in the answers.
Challenges
§ Asking the right quesMons is not easy.
§ Properly wording quesMons is not easy.
§ Gehng the right sample (sampling) is not easy.
§ MoMvaMng respondents to respond (candidly) is not easy.
What is Survey Data good for?
The Four Phases of the Survey Research Process
1. Planning a Survey
o Define Research ObjecMves
o SelecMon of Research Variables
2. Designing a Survey
o ConceptualizaMon and OperaMonalizaMon
o Construct and Pretest QuesMonnaire
3. ImplemenMng a Survey
o Choose Relevant PopulaMon
o Sampling
o Data CollecMon
4. Gaining Insights from a Survey
o Data PreparaMon and Measurement EvaluaMon
o Data Analysis
o ReporMng of Results
è ConceptualizaMon and OperaMonalizaMon, Construct and Pretest QuesMonnaire,
Choosing relevant PopulaMon and sampling oPen at the same Mme.
Following pages gives an overview of the upcoming topics in the course:
, Analyzing a Survey (Topic Lecture 3)
To use regression analysis to infer causal relaMonships or for predicMon outside the data
sample, one must carefully jusMfy why there is a causal relaMonship between two variables or
why exisMng relaMonships have predicMve power for a new context!
Planning a Survey (Topic Lecture 4)
Designing a Survey: Measurement (Topic Lecture 5)
Designing a Survey: QuesMonnaire Design (Topic Lecture 6)
