Affiliation Information in DataCite Dataset Metadata: a Flemish Case Study. (29th March 2021)

• Record Type: Journal Article
• Authors: Van Wettere, Niek
• Is Part Of: Data science journal. Volume 20(2021)
• Publication Date: 2021-03-29
• Subjects: DataCite -- Scholix -- research data -- metadata -- affiliation; Science -- Data processing -- Periodicals; Database management -- Periodicals; 502.85
• DOI: 10.5334/dsj-2021-013
• Languages: English
• ISSNs: 1683-1470

Abstract

This article aims to evaluate how and to what extent metadata of datasets indexed in DataCite offer clear human- or machine-readable information that enables the research data to be linked to a particular research institution. Two main pathways are explored. First, researchers can encode their affiliation information at the moment of data submission. This can be done by means of free-text metadata fields or via the inclusion of identifiers such as GRID/ROR and ORCID. Second, affiliation information can be traced indirectly through linking between a dataset and associated publications, given that the metadata of publications is often more explicit about affiliation information than the metadata of datasets. Both pathways of affiliation information encoding are evaluated on the basis of metadata pertaining to datasets created at the five Flemish universities. It is shown that good practices such as encoding of affiliation information in a dedicated metadata field or inclusion of ORCID in the metadata are on the rise, but could be expanded further. Finally, the establishment of links between datasets and related publications is often lacking in dataset metadata, although there are important differences between data repositories, as is also demonstrated in a more data-intensive follow-up analysis based on random samples of metadata records. It is important that data repositories address this issue by providing a metadata field clearly dedicated to associated publications, prominently displayed on the landing page of the dataset.