Survey measurement

This programme of research aims to implement randomised experiments within standard survey procedures to investigate the cognitive and affective processes underlying responses to different types of commonly used survey questions. Key objectives of the research are to improve the validity of the survey questions examined and to promote the use of experimental designs embedded within surveys. The research is carried in collaboration with Patten Smith (Ipsos-MORI) and Nick Allum (University of Essex)

To date the topics we have investigated include:

• How should we measure factual (e.g. political) knowledge in surveys?
• How do respondents interpret two different commonly used questions about social trust – do these questions measure the same thing?
• Order effects in measuring party affiliation and economic performance
• Context effects in the measurement of social trust
• Should we use ‘middle-alternatives’ in Likert-type attitude questions?
• Responses to questions about non-existent government policies (and what does this tell us about responses to real government policies?)

For further information on the project please contact Patrick Sturgis