Data management plans and the Recherche Data Gouv repository
Data management plans
Data management plans (DMPs) are operational documents that provide important information about the data produced, ranging from the conditions of their collection or production to their later publication. Funding agencies like the French National Research Agency (ANR) ask for DMPs to be created after the start of the project and then regularly updated. Beyond such contractual aspects, having a DMP is a 'best practice' for setting up or running a research project. It also means all data management and sharing issues can be explored.
The publication of data management plans should be encouraged in the framework of an open science approach because it clearly makes the re-use of datasets easier. It is advisable to follow the following instructions as regards the Recherche Data Gouv repository:
Recommended practice
- DMPs should be deposited on HAL and cross-referencing (reciprocal links between the dataset(s) and the data management plan) should be implemented.
Strongly advise against
- Depositing a DMP on Recherche Data Gouv as the only element in a dataset.
This recommendation is for all users but each institutional space has its own policy.
Details
Currently on HAL, there is a 'Data Management Plan' (a subset of 'Reports') document sub-type which means DMPs can be published as duly labelled digital objects and also cited while links can be generated from and towards other scientific productions.
Given this development, the repository-registry resource centre strongly advises you not to deposit a dataset consisting solely of the DMP (without the data attached to it) in the Recherche Data Gouv repository. This practice is equivalent to falsely counting a DMP as a dataset and undermines the repository's primary purpose, namely to host structured data.
It is not inherently bad practice to deposit a DMP as part of and accompanying a dataset but it is rather limiting. It does not enable the same DMP to be linked to several datasets or also means the DMP cannot be described using other metadata than those of the dataset which it is a part of. It implies that the DMP is only published if an associated dataset is also published. For all these reasons, this is not a recommended practice.
Find out more
- https://anr.fr/fr/lanr/engagements/faq-pgd/
- https://dmp.opidor.fr/static/glossary
- https://doranum.fr/plan-gestion-donnees-dmp/le-plan-de-gestion-de-donnees-pas-a-pas_10_13143_t94g-9j96/
- https://science-ouverte.inrae.fr/fr/les-donnees-et-le-numerique-scientifiques/gerer-des-donnees-et-des-codes/comment-rediger-un-plan-2
- https://ec.europa.eu/research/participants/data/ref/h2020/grants_manual/hi/oa_pilot/h2020-hi-oa-data-mgt_en.pdf
- https://coop-ist.cirad.fr/gerer-des-donnees/rediger-un-pgd/1-qu-est-ce-qu-un-pgd