Investigating and reducing non-sampling errors on EU-SILC
The European Union Statistics on Income and Living Conditions (EU-SILC) programme involves the collection of standardised data co-ordinated by Eurostat, the statistical office of the EU. Participation is compulsory for all EU member states. In most countries the data are collected entirely through a bespoke panel survey, while in a minority of countries some of the data are collected from registers. Previous research commissioned by Eurostat has studied in detail the sampling error associated with estimates based on EU-SILC data, but little attention has been paid to other sources of survey error.
This programme of research, which constitutes part of the Third Network for the Analysis of EU-SILC (NetSILC3), will investigate the extent and nature of non-sampling errors and will explore ways of reducing such errors in future. Sub-projects will address under-coverage, unit nonresponse and attrition, item non-response, and several aspects of measurement error including external validation of income measures, health measures, and housing measures. The research programme is co-ordinated by ISER’s Professor Peter Lynn and will involve researchers from 12 organisations in nine countries.
The objectives of the project are designed to identify the main sources of non-sampling errors in EU-SILC, to describe the nature and impact of each type of error, and to produce guidance on reducing them, through a combination of improved data collection practices and post-survey adjustment methods. Recommendations from the project should inform the revision of EU-SILC, as support to the elements to be included in the revised legal basis.
EU-SILC Longitudinal and Cross-sectional files