Home>Comparing data from large-scale French electoral surveys
21.03.2023
Comparing data from large-scale French electoral surveys
As the structure that has disseminated a large proportion of the French electoral surveys produced since 1958, CDSP has developed a database of questions and harmonised variables from this collection in order to facilitate opportunities for comparative studies by the research community.
The CDSP’s collection of electoral surveys consists of 25 pre- and post-electoral surveys produced between 1958 and 2022. Heavily used by the academic community, their richness and heterogeneity have prompted an exercise of post-harmonisation. In all, more than 2500 variables drawn from 11 benchmark electoral surveys are grouped within a common database of metadata covering French elections during the fifth Republic.
Harmonising heterogeneous data
In order to carry out these exercises in curatorship, Malo JAN and Lucie MARIE drew on the standard and conceptual classification developed for the European True European Voter project together with the “variable cascade” model in the DDI-L standard for the production of metadata.
The variables in the 11 corresponding benchmark surveys were manually linked with a concept from the TEV nomenclature. In return, similar questions or variables from different sources can then be easily combined under a single concept and are no longer specific to a particular dataset.
By making it easy to find the information (variable name, question, category) on equivalent data, through a tree structure that brings together the main concepts of the survey on electoral behaviours, this smoothing process makes it possible to conduct historical comparisons of electoral opinions and behaviours in France.
Contributing to the resources of the European Question Bank
Entering CDSP data on the explore.cdsp.sciences-po platform will increase possibilities for comparing international data and using them at a larger scale..
Harvested by CESSDA’s European Question Bank, this platform, which emerged from the second phase of the UpMet project is dedicated to the exploration of longitudinal or comparative surveys.