@@src2@@
La représentativité en statistique
Collection : Méthodes et savoirs
n° 8, 2013, 136 pages
Introduction de François Beck
Conclusion générale de François Beck
- Chapitre 1. Histoire de la représentativité statistique : quand le politique refait toujours surface. Emmanuel Didier
- Chapitre 2. La représentativité à l'Insee. Olivier Sautory
- Chapitre 3. La représentativité dans les instituts de sondage. Alain Tripier
- Chapitre 4. La représentativité et les populations difficiles à joindre : le cas des sans-domicile. Maryse Marpsat et Nicolas Razafindratsima
- Chapitre 5. Représentativité et comparaisons internationales : évolution et méthodes d'analyse. Arnaud Bringé et Patrick Festy
- Chapitre 6. Le point de vue de l'épidémiologiste : être représentatif ou universel ? Marcel Goldberg, Alice Guéguen, Rémi Sitta et Marie Zins
Conclusion générale de François Beck
Sample surveys are a common component of public life today, but their widespread use raises the question of representativeness, a notion not always accepted by statisticians, partly because its very definition is a source of disagreement. In statistics, the term representativeness is linked to the possibility of inferring a whole (reference population) from a part (sample). What criteria should be used to select the sample? How can a sample be constructed from very heterogeneous or hard-to-reach populations? Or from cohorts which evolve over time?
In the universe of quantitative surveys, the notion of representativeness now gives legitimacy to numbers in public debate. Not only in scientific research but also in civil society, survey development has been accompanied by the dissemination of survey results on a whole variety of questions, ranging from marketing studies to political, economic and social issues, with the argument of sample representativeness being given as proof of their validity.
This series of texts, based on contributions to the study seminars organized by the Centre Maurice Halbwachs and the Société française de Statistique, reflects on the history of the concept and on the difficulty of producing a definition of representativeness which, if not theoretical is at least coherent, as illustrated by its different uses and applications in wide-ranging survey contexts.
In the universe of quantitative surveys, the notion of representativeness now gives legitimacy to numbers in public debate. Not only in scientific research but also in civil society, survey development has been accompanied by the dissemination of survey results on a whole variety of questions, ranging from marketing studies to political, economic and social issues, with the argument of sample representativeness being given as proof of their validity.
This series of texts, based on contributions to the study seminars organized by the Centre Maurice Halbwachs and the Société française de Statistique, reflects on the history of the concept and on the difficulty of producing a definition of representativeness which, if not theoretical is at least coherent, as illustrated by its different uses and applications in wide-ranging survey contexts.