Quality of demographic data in GGS Wave 1

Jorik  Vergauwen; Jonas Wood; David De Wachter; Karel Neels

doi:10.4054/DemRes.2015.32.24

Volume 32 - Article 24 | Pages 723–774

Quality of demographic data in GGS Wave 1

By Jorik Vergauwen, Jonas Wood, David De Wachter, Karel Neels

Date received:

16 Oct 2013

Date published:

13 Mar 2015

Word count:

13249

Keywords:

data quality, demographic indicators, fertility, Generations and Gender Survey (GGS), GGP, GGS, nuptiality, validation, vital registration

DOI:

10.4054/DemRes.2015.32.24

Replicable Material:

32-24 Appendix (pdf file, 4 MB)
demographic-research.32-24 (zip file, 32 kB)
readme.32-24 (text file, 8 kB)

Download PDF
Submit a Response Letter

Abstract

Background: A key feature of the Generations & Gender Programme (GGP) is that longitudinal micro-data from the Generations and Gender Surveys (GGS) can be combined with indicators from the Contextual Database (CDB) that provide information on the macro-level context in which people live. This allows researchers to consider the impact of socio-cultural, economic, and policy contexts on changing demographic behaviour since the 1970s. The validity of longitudinal analyses combining individual-level and contextual data depends, however, on whether the micro-data give a correct account of demographic trends after 1970.

Objective: This article provides information on the quality of retrospective longitudinal data on first marriage and fertility in the first wave of the GGS.

Methods: Using the union and fertility histories recorded in the GGS, we compare period indicators of women’s nuptiality and fertility behaviour for the period 1970-2005 and cohort indicators of nuptiality and fertility for women born after 1925 to population statistics.

Results: Results suggest that, in general, period indicators estimated retrospectively from the GGS are fairly accurate from the 1970s onwards, allowing exceptions for specific indicators in specific countries. Cohort indicators, however, were found to be less accurate for cohorts born before 1945, suggesting caution when using the GGS to study patterns of union and family formation in these older cohorts.

Conclusions: The assessment of the validity of demographic data in the GGS provides country-specific information on time periods and birth cohorts for which GGS estimates deviate from population statistics. Researchers may use this information to decide on the observation period or cohorts to include in their analysis, or use the results as a starting point for a more detailed analysis of item nonresponse in union and fertility histories, which may further improve the quality of GGS estimates, particularly for these earlier periods and older birth cohorts.

Comments: Detailed country-specific results are included in an appendix to this paper, available for download from the additional material section.