Ageing and public statistics

Dialogue sur les besoins statistiques et les contraintes des producteurs et utilisateurs des données sur le vieillissement.


This series of DIALOGUE seminars will provide an opportunity to share experiences and discuss the needs and constraints of the various actors (data producers and users) in the spheres of public policy and research. The aim is to consider possible changes to data collection systems. In this unprecedented context of a pandemic that has hit the elderly, two themes will be examined:

• Session 1. Data on excess mortality in the context of a health crisis.
• Session 2. Data on autonomy in old age.


Public statistics is a tool for counting, describing and analysing. It is used to support political objectives, to help build systems, such as social protection, and to monitor their development in the population: civil status data, population censuses, registers, surveys on health, work, the family, housing, etc. Although the needs are not perfectly superimposable, research and public policy often share the same measurement and analysis objectives.

On the basis of common objectives, shared methods and skills, public data collection has evolved. The 'panellisation' of samples or matching between databases increases the relevance of sources for building specific research projects. Administrative sources (e.g. tax, medical, pensions, etc.) have long been inaccessible (in legal terms and in terms of content) or poorly suited to research questions. Today, there are potentially veritable mines of data to be explored, the boundaries of which need to be defined in order to make them operational.

Part from targeted operations, such as surveys on issues of old ages inclusion, poverty or dependency (which we are going to explore), there are still gaps in social statistics, and the oldest age often escape measurement: difficulties in surveying frail people; out-of-scope (e.g. EHPAD)... The number of old and very old people in the samples is often too low to represent the diversity of situations. These observations are now prompting those involved in this field of research to meet and share ideas to design and build data sources that are less 'imperfect'.