Hierarchical analysis of unit–nonresponse in multistage surveys
S3RI Members: Gabi Durrant, John Micklewright, Chris Skinner
Collaborator: Fiona Steele (University of Bristol)
Research Staff: Julia D'Arrigo, Sylke Schnepf
The project has two principal aims: to better understand nonresponse as a social phenomenon, and to advance methodology for adjusting for nonresponse. Our results will guide survey practitioners in how to increase response rates and reduce nonresponse bias.
This three-year ESRC-funded project is directed by Gabriele Durrant (University of Southampton) and is being carried out in conjunction with Chris Skinner (University of Southampton), John Micklewright (Institute of Education), Fiona Steele (University of Bristol), Julia D'Arrigo (University of Southampton) and Sylke Schnepf (University of Southampton).
The specific aims of the project are:
- to develop models for nonresponse in multistage surveys, allowing for correlation in response behaviour among survey units (e.g. sample members allocated to the same interviewer).
- to analyse interviewer effects on the probabilities of noncontact and refusal.
- to develop response propensity models on process data to predict the likelihood of contact at each call.
- to develop post-survey adjustment methods for nonresponse with hierarchical survey data.
- to provide guidance for social scientists and survey practitioners on the use of methods to reduce nonresponse bias and to increase response rates.
- to disseminate the results of this work through publications, conference presentations, a short course, and an international research symposium.
The project is part of the ESRC Survey Design and Measurement Initiative.
For further information on the project please contact Dr Gabi Durrant
As part of this project, we are pleased to announce an international research symposium on Recent Advances in the Use of Paradata (Process Data) in Social Survey Research to be held at the Royal Statistical Society (RSS) in London on 9th December 2010. The event will be preceded by a 2-day short course and workshop on the same topic. For more information or to register, please go to the event's webpage, or contact: paradata-symposium@lists.soton.ac.uk.
Recent outputs include:
International Research Symposium
Special issue of the JRSSA on paradata