Family structures in France: population census vs. survey on family and housing vs. labour force survey

Aude Lapinte - Guillemette Buisson

Documents de travail
No F1703
Paru le :Paru le15/05/2017
Aude Lapinte - Guillemette Buisson
Documents de travail No F1703- May 2017

The purpose of family surveys (EFL), baseline surveys to analyze family structures and their evolutions, is to complement the analyzes usually conducted from other more regular sources, such as the annual census survey (EAR) or the labour force surveys (EEC). They also make it possible to highlight the limits of these other sources, and their possibility of use on family themes.
This working paper first compares the family structures of minors in the EFL 2011 and the EAR 2011. Overall, the results from EFL and EAR are consistent. However, the proportion of children living in single-parent families and the proportion of single-parent families are slightly higher in the EAR than in the EFL. Thus, 17.9% of children live in single-parent families in the EFL, compared with 18.6% in the EAR. This over-representation was due to the conventions on the notion of couple in the census that were reviewed in 2015 following this work. This over-representation should therefore be significantly reduced.
The second part is devoted to the comparison of the family structures in which minors live in the EFL and the EEC. Unlike the EAR, the EEC underestimates the proportion of single-parent families relative to the EFL (-0.9 point). In addition, there are significantly fewer children in stepfamilies in the EEC than in the EFL. These findings make the use of the EEC difficult to estimate the distributions of the different types of families. Moreover, the changes in the distributions of the different types of families measured in the EEC fluctuate greatly and its use therefore seems again uncertain. The weighting system chosen partly explains the discrepancies between EFL and EEC.
Finally, the third part aims to evaluate the quality of certain results published from the EFL, comparing different ways of calculating them. The objective is to measure possible biases but also to estimate a number of innovative results. In particular, the EFL makes it possible to know the family situation of children when they go to their relatives with whom they do not live most of the time.