Open Access to Research Data

As in the case of research literature, one argument for making research data available in open access is that their production was financed with public funds. At an early stage in the history of open access, the Berlin Declaration on Open Access to Knowledge in the Sciences and Humanities recognised data as objects that should be made openly available. Besides the intrinsic motivation to be able to work more efficiently in increasingly data-driven research with the help of good data management and to benefit from open data oneself, the main drivers of the publication of data are the research funders.

When publishing research data, a suitable repository should be chosen – where possible one that provides open access to the data. A disciplinary repository that is well established in the community in question should always be the preferred choice, as one’s own data are thus in a good, specialised context, and findability is easier. The Registry of Research Data Repositories, re3data, can be used to select a suitable data repository. If no suitable disciplinary repository can be found, general or institutional repositories can be used. Well-known examples of general repositories are Zenodo and Dataverse.

To guarantee the long-term provision and findability of the data, a permanent address must be assigned. This persistent identifier also ensures the citability of the datasets. Preferred identifiers are the Digital Object Identifiers (DOIs) provided by the DataCite consortium.

As the reusability of research data is greatly limited if they lack adequate descriptions and metadata, it is imperative that they be curated in accordance with the FAIR principles before publication. Metadata should be assigned at the earliest possible point in time during the research process. They comprise both technical metadata (e.g.: When and by whom was the dataset collected?) and substantive metadata (e.g.: What is the content of the individual variables?). There are specific metadata standards for numerous disciplines. These standards are promoted, maintained, and published in particular by the NFDI consortia (see the RDA Metadata Standards Catalog). During curation, the data are, above all, technically checked. This includes checking the data format, the basic access, and the formal accuracy. Regarding data format, long-term accessible and open data formats should be used, for example. The checking of the data content must be carried out mainly by the researchers themselves. The curation ends with the choice of a suitable licence. The Creative Commons (CC) licences have proved their worth (more information can be found on the English-language pages of the research data portal forschungsdaten.info).  However, CC licences are not suitable for licensing code or software. There are special open source licences for this purpose. 

One criticism frequently voiced by researchers stems from their concern that others will benefit excessively from their wealth of data, and that they will not achieve the reputation they need for their scientific careers. It should be made clear in this connection that publication should be sought as early and as comprehensively as possible. However, the data may still be published at a later point in time – after the analysis or subject to an embargo period. This reservation is also being addressed through proposed reforms in the assessment of research. Sovereignty over the data can remain with the data producer. The additional costs associated with the required handling of the data are also frequently mentioned. It is already possible to also request funding for these costs when applying for research funding. Data management must be regarded as a key element of scientific research and be adequately staffed and funded.