Understanding nonresponse on Understanding Society
Principal Investigator: Peter Lynn, Institute for Social & Economic Research, University of Essex
Co-Investigators: Jon Burton, Gundi Knies and Birgitta Rabe (ISER)
Project duration: 1 April 2013 - 30 September 2014
The project aims to find out more about which kinds of people take part in Understanding Society: the UK Household Longitudinal Study (UKHLS) and which kinds of participants agree to having administrative data linked to their survey responses for research purposes. This will help researchers to improve the accuracy of analysis using the survey data and will also help survey designers to find better ways of persuading people to take part in surveys and to give their agreement to data linkage.
It involves two sub-projects. The first sub-project, in co-operation with the Office for National Statistics, will link UKHLS sample addresses to responses to the 2011 Population Census. This will enable comparisons to be made between responding and non-responding households in terms of the micro-level Census data for those specific households. This unique opportunity to link the survey sample to the decennial Census returns will provide researchers with the clearest picture yet of the differences between survey participants and non-participants and will provide information that will help analysts to make appropriate adjustments for non-response.
The second sub-project will investigate the role of interviewers in obtaining respondent consent to have personal data from administrative records linked to their survey responses for research purposes. The project researchers will conduct a survey of Understanding Society interviewers to obtain information about their attitudes, personality, interviewing practices and experience and will link this to the survey data. This will allow the project researchers to explore the extent of variation between interviewers in success at obtaining respondent consent, and whether this variation can be explained by characteristics or attitudes of the interviewers. The project researchers will also describe the differences between respondents who give consent for data linkage and those who do not, in terms of a wide range of socio-economic variables and for two different types of linked data (health records and education records).