Archiving and publishing research data

At the end of your study, you must archive the data that underlie your publication for a minimum of 10 years. Proper archiving promotes scientific integrity and enables (internal) reuse. It adds to the findability and usability of your research, and as a result increases the impact of your scientific output. Radboud University's RDM policy also states that you must share your data as open as possible, and as closed as necessary. That means that in most cases you want to publish (a part of) your data. Both archiving in a closed environment for scientific integrity and publishing require similar preparations:

Decide what part of your data needs a closed archive and what part can be published
Find a suitable repository
Prepare your data for archiving
Archive your data

Decide (open and/or closed archiving)

The Radboud University RDM policy states that “Research data should be archived as open as possible, as closed as necessary. The starting point is that research data should be made publicly available.”. This means that you need to decide what part of your data can be openly archived (published) and what part belongs in a closed archive.

Not publishing data or restricting access to data

There are some situations in which you might not be allowed to publish your data, or must restrict access to the published data.

Human participant data which cannot be anonymised, i.e. pseudonymised or directly identifiable personal data. Data sharing must always be in compliance with the agreements you made with your research participants in the informed consent form. Always share data as de-identified as possible
Data involving controlled use of animals
Data subject to contractual constraints, such as licences, collaboration or data processing agreements, or funder agreements
Data with commercial potential
Data that may be patented
Data subject to export controls
Environmentally sensitive data
Politically sensitive data

If you are not sure if your data is suitable for publishing or if you must restrict access to the data, contact your institute’s data steward.

Documents are not useful to others

Some documents are not useful for other users, but need to be archived for scientific integrity or other reasons.

Administrative documents, such as your data management plan, ethical approval documents, lab journals, audit trails, etc.
Unprocessed or unpublished versions of the data. However, please consider if you really need to archive those versions and if the raw data and/or the final version of the analysed data might be enough
Specific milestone versions of your work-in-progress articles, code, data, etc. The final products are often suitable for publication, but important milestone versions might be of value to show scientific integrity
Unpublished documents (analyses, code, etc.) that were not part of a publication, but are valuable and need to be stored properly
Your personal notes

Find a repository

Once you have decided whether your data belongs in a closed and/or open archive, you need to find a suitable repository (or sometimes two if you need an open and a closed archive). First check your institute’s RDM policy or ask a (senior) colleague for advice. They might recommend specific repositories that are well known in your field of study. Also see if your research institute can make use of the Radboud Data Repository (RDR). If you use the RDR, you automatically comply with Radboud University’s standards and the requirements stated below.

Trustworthy and FAIR-enabling

The repository of your choice, whether it is closed or open, needs to be trustworthy and apply the FAIR principles (Findable, Accessible, Interoperable and Reusable). Using a trustworthy and FAIR-enabling repository can make your data more future-proof and it increases the impact of your scientific work. Additionally, the latter (FAIR-enabling) is required by Radboud University as defined in the RDM policy.

Judge the trustworthiness of a repository

To judge the trustworthiness of a repository, check the following:

Any certification it holds, such as the CoreTrustSeal and ISO or Nestor Seal DIN standards. If a repository is certified as trustworthy, there is less/no need to check the following points;
The repository’s mission, policy and guidelines. Try to find the answer to questions such as: does the repository have a long-term preservation goal? Does it commit to the FAIR principles? Does it offer guidelines for sustainable data formats and metadata standards? What is their policy regarding access management and privacy concerns?
The use of persistent identifiers such as DOI’s, which ensure the findability of the data;
The use of licenses. For example: Does the repository offer broadly acknowledged (open access) licenses for research data and code? Does it offer the option to add data use agreements for sharing research data under restricted access conditions?

Check if a repository is FAIR-enabling

To check if a repository is FAIR-enabling:

Again, first see if the repository has a certification, such as the CoreTrustSeal and ISO or Nestor Seal DIN standards? If so, you can probably skip the next points on this checklist;
Does your dataset get a globally unique and persistent identifier (PI)?
Does the repository have metadata standards?
Can you make use of your field’s vocabularies? For example, can you select keywords from a specific vocabulary such as SFN or MeSH keywords.

Prepare your data

Once you have found a proper repository, you need to prepare your data for long-term storage, whether this is closed archiving or publishing.

Add rich metadata to your dataset. Provide a proper title, add a comprehensive description of the research context and the content of the dataset, include keywords, and fill in as many of the other metadata fields that the repository of your choice provides. This step is key for the Findability and Reusability of your dataset. It is not only important for published datasets, but also for closed archived datasets, because those metadata might be the only visible and thus findable part of your dataset.

Add documentation. For Findability and Reusability, add rich documentation about the context, quality and condition, or characteristics of the data. Usually this is done in a readme file that is part of the metadata and openly accessible.

Additional FAIR-ification steps:

For Accessibility, make sure all access rights are clear and implemented correctly. For closed archiving, this means making sure that those who are allowed access to the archive, have access and no-one else. For publishing, this means choosing a fitting access level and license or Data Use Agreement. Remember to publish your data “As open as possible, as closed as necessary”
For Interoperability and Reusability: If your field has a vocabulary for specific terms or concepts, use those in your dataset and add information about the vocabulary. Otherwise add documentation to clarify what your terms and concepts are (e.g. by adding a codebook)

If you are publishing parts of your data, make sure that they do not contain any information that you are not allowed to share under the selected access level, especially when it comes to personal data.

More information on anonymisation and pseudonymisation

Contact

Do you have a question? Get in touch with Research Data Management support:

rdmsupport [at] ubn.ru.nl + 31 24 - 361 28 63