Total Survey Error in Practice

25Errors in Linking Survey and Administrative Data

Joseph W. Sakshaug^1,2 and Manfred Antoni³

¹ Cathie Marsh Institute for Social Research, University of Manchester, Manchester, UK

² Department of Statistical Methods, Institute for Employment Research (IAB), Nuremberg, Germany

³ Research Data Centre (FDZ), Institute for Employment Research (IAB), Nuremberg, Germany

25.1 Introduction

Linking surveys with administrative records is regarded as a useful and cost‐effective means of supplementing survey data collection with existing administrative sources. Evidence of the widespread practice of linking surveys with administrative data is reflected in Table 25.1, which shows a selection of large‐scale surveys and the types of administrative data they link to. The types of administrative data linked to these and other surveys vary considerably, but most often include social security records, tax and benefit records, healthcare enrollment and billing records, and education records.

Table 25.1 A selection of large‐scale surveys that link survey records to administrative records.

Name of survey	Country	Type of administrative data
45 and Up Study	Australia	Admitted patient; emergency department; cancer registry, mortality; medicare benefits schedule; pharmaceutical benefits scheme data
Australian Longitudinal Study on Women’s Health (ALSWH)	Australia	Admitted patients data; cancer registry; perinatal data
Canadian Community Health Survey (CCHS)	Canada	Data on deaths; hospital records; medical services; ...