The process of creating an in-house interface for converting Blaise output into DDI metadata

Alexandre Mairot (1)

1 CDSP - Centre de données socio-politiques de Sciences Po (Sciences Po, CNRS)

ELIPSS (Longitudinal Internet Studies for Social Sciences) is a longitudinal probabilistic French panel. To release the 70 produced datasets, we had to deal with numerous issues, in a environment with limited resources and an important fragmentation of the tools: Blaise for the surveys, DDI-C for the documentation, SPSS for producing the datasets and SAS for the weights of the respondents. Also, heterogeneous survey programming and documentation practices had to be taken into account. Our primary goal was to favorise the reuse of data and metadata. To handle the challenges of this project, it was necessary to imagine and implement new processes and tools that are adapted to the context and have the smallest impact on the existing environment to increase the quality and the efficiency of the data management. The solution we found was to create an interface, a liaison between all the tools for fluidizing the production and dissemination cycle of the datasets. The solution we conceived is based on the output and the input files of the software linked between them by a tool developed with R. This presentation underlines the building process of this solution. We also give feedback on how it performed during 5 years in our environment.

Mots clés: ELIPSS; longitudinal panel; data management; DDI metadata

Domaines: Sociologie; Méthodes et statistiques

Alexandre Mairot. The process of creating an in-house interface for converting Blaise output into DDI metadata. EDDI19 - 11th Annual European DDI User Conference, Dec 2019, Tampere, Finland. ⟨10.5281/zenodo.3568059⟩. ⟨hal-02981752⟩