R0.4. Level of Curation Performed

The Dataverse software includes core functionality, particularly its permissions, notifications, and file ingest functionality, that facilitates all four types of curation levels listed in the certification guidelines, but by itself satisfies only the requirements of the first level.

A. Choose "Content distributed as deposited" if:

  • Depositors can publish datasets without collection support staff reviewing those datasets

B. Choose "Basic curation" if:

  • Collection support staff review deposited datasets before publication, for example by using the Dataverse software's "submit for review" workflow, to ensure that deposits contain data (and not other types of research objects or spam)
  • Depositors deposit certain types of data files, e.g. tabular data and FITS files, that the Dataverse software is able to ingest to create additional metadata and create TSV copies of tabular files

C. Choose "Enhanced curation" if, in addition to the curation practices described in "Basic curation":

  • Collection support staff help streamline and standardize the creation of dataset metadata by:
    • Providing instructions to depositors for creating/adding metadata.
    • Customizing Dataverse collections to require that depositors add certain metadata
    • Creating metadata templates for depositors to use
    • Customizing metadata fields to ensure that data is described in ways that follow domain-specific best practices
  • Collection support staff review deposited datasets before and after publication and work with depositors to improve how datasets are described

D. Choose "Data-level curation" if, in addition to the curation practices described in "Enhanced curation":

  • Collection support staff review data files and suggest or make edits to data files. In addition to downloading and opening files on their own computers, collection support staff may use external tools enabled in the Dataverse repository to review the data without needing to download the files.
     

Answers from successful applicants

Tilburg University Dataverse collection:

A. Content distributed as deposited
B. Basic curation – e.g., brief checking, addition of basic metadata or documentation
 

QDR:

C. Enhanced curation – e.g. conversion to new formats; enhancement of documentation
 

DataverseNO:

A. Content distributed as deposited
B. Basic curation – e.g. brief checking; addition of basic metadata or documentation
C. Enhanced curation – e.g. conversion to new formats; enhancement of documentation
D. Data-level curation – as in C above; but with additional editing of deposited data for accuracy

Datasets deposited into DataverseNO are reviewed/curated by Research Data Service staff before they are published. Research Data Service staff are mainly library staff working at DataverseNO partner institutions and having post-graduate level expertise within the different subjects represented by the deposited data. In addition, responsible Research Data Service staff have in-depth expertise in FAIR research data management (RDM). Typically, Research Data Service staff are (Senior) Research Librarians / Subject Librarian, but also other research support staff specialized in RDM may review/curate research data deposited into DataverseNO. Research data deposited in the top-level collection of DataverseNO are reviewed/curated by Research Data Service staff at UiT The Arctic University of Norway. If necessary, Research Data Service staff at UiT The Arctic University of Norway also give advice to Research Data Service staff at other DataverseNO partner institutions. The level of expertise of Research Data Service staff at partner institutions is not regulated by DataverseNO partner agreements. However, DataverseNO partner agreements require DataverseNO partners to fulfill all DataverseNO policies and guidelines, including the DataverseNO Curator Guidelines (see below). Research data deposited into special collections within the DataverseNO repository are reviewed/curated by Research Data Service staff who are highly proficient within the subject or discipline at stake. In TROLLing, review/curation is carried out by Senior Research Librarians responsible for language and linguistics at the University Library at UiT The Arctic University of Norway. If necessary, a scientific advisory board may be established for special collections within the DataverseNO repository; cf. TROLLing [1].

During review/curation, DataverseNO does not attempt to judge the scholarly quality of deposited datasets. As described in the DataverseNO Deposit Agreement [2], determination of the research quality is at the discretion of, and the responsibility of, the Long-Term Contact Person, as named in the metadata about the deposited dataset at stake.

Research Data Service staff review deposited datasets for alignment with criteria [3] [4] for depositing and/or to extend the metadata as needed to facilitate greater accuracy and discoverability. Both metadata and data files of deposited datasets are curated according to best practice. There are four areas to be checked: the uploaded files (both data and documentation), the registered metadata, the chosen license, and versioning, according to the checklist in the DataverseNO curator guidelines. Lack of compliance with the DataverseNO Deposit Agreement is communicated to the depositor and the dataset is returned for amendment. After finishing this review/curation process, the curator publishes the dataset.

Any changes in a dataset after its initial publication results in a new version of the dataset. Older published versions always remain openly accessible in DataverseNO. Published data can thus not be unpublished – with the only exception being cases where access to the file(s) in a dataset or the entire dataset has to be removed. This process is regulated in the DataverseNO Preservation Policy [3].

During the review/curation process outlined above, the curator gives advice to the depositor about how to prepare and describe the dataset in order to obtain maximum re-usability of the data, as described in the DataverseNO Curator Guidelines [4]. The review/curation process may imply curation at all levels (A–D), including D-level with advice on formats for dates and numbers, or column headings. This review/curation process is carried out before the initial publication of datasets, and before any publication of a new version of a published dataset. For more information on the curatorial review process, please see the DataverseNO Curation Guidelines and the DataverseNO Accession Policy [5]. Datasets that are not compliant with the DataverseNO policies and guidelines are not published. If a curator identifies fundamental nonconformity with the DataverseNO policies and guidelines, and the depositor does not agree to make necessary changes, the curator addresses the problem by raising the issue within the curator community of DataverseNO to reach a conclusion. The conclusion is communicated to the depositor. If the reached conclusion is not accepted by the depositor, the issue is raised to the Board of DataverseNO. If applicable, the Board of DataverseNO may discuss the issue further with an advisory committee, before a final decision is made.

References:
[1] https://site.uit.no/trolling/people/
[2] DataverseNO Deposit Agreement (Data Deposit): https://site.uit.no/dataverseno/about/policy-framework/deposit-agreement/
[3] DataverseNO Preservation Policy: https://site.uit.no/dataverseno/about/policy-framework/preservation-policy/
[4] DataverseNO Curator Guidelines: https://site.uit.no/dataverseno/admin-en/curatorguide/
[5] DataverseNO Accession Policy (Quality Control): https://site.uit.no/dataverseno/about/policy-framework/accession-policy/