Skip to main navigationSkip to main content
The University of Southampton
Southampton Statistical Sciences Research Institute

Survey methods

We are at the forefront of international research on methodology for the design and analysis of sample surveys, and the use of surveys and administrative data in official statistics.

Our survey methodology research group is considering the ways in which different methods of collecting survey data influence the quality of the statistics. This has become particularly important as data collection operations have been changed in response to COVID-19, and our group leads the GenPopWeb2 network funded by the Economic and Social Research Council setting the research agenda in this area. Other projects include the effect of mixed mode data collection.

We have a major research contract with the Methodology group of the Office for National Statistics which has been in place for over 25 years, involving a wide variety of projects, including recent research on methods for the population census and an administrative data census, investigation of the feasibility of producing regional price indices, and assessing the accuracy of population estimates. Our researchers in Official Statistics also have strong connections to National Statistical Offices round the world, and our research experience enhances our teaching in Official Statistics. The MAKSWELL project completed in 2020 and involving several European partners examined ways in which new forms of data could be used to improve sustainability and wellbeing indicators and promote their use in policymaking and evaluation.

Our research into uses of administrative data is wide-ranging, from quality frameworks and quality measurement for administrative datasets to innovative uses of administrative data in official statistics and use of non-probability samples. We also undertake research on capture-recapture for population size estimation, especially related to population censuses (traditional and administrative data-based) and on extensions to their methodology, such as trimmed dual system estimation. This links with the work on capture-recapture in the biosciences.

We also have a group specialising in small area estimation and robust estimation, using multilevel modelling and state space modelling approaches. These have applications in surveys and data with blurred location information. Research on ways to improve standard models applied in geography, based on satellite data and other alternative data sources, is underway.

Research projects

Privacy Settings