Research Data

The FAIR principles stand for Findable, Accessible, Interoperable, and Reusable, and focus on how data can be managed and described in a way that allows others to understand and reuse them in the future.

• Findable: Data are easy to locate. Data and supplementary materials have sufficient metadata and a persistent identifier.
• Accessible: Data are accessible. Metadata and data are understandable for both humans and machines. Data are deposited in a trusted repository.
• Interoperable: Data can be combined and reused across, for example, systems, institutions, and time. Existing standards, vocabularies, formats, and methods for metadata are used.
• Reusable: Data can be reused. Clear usage licenses provide information on how the data can be reused.

It is important to note that data will rarely be fully FAIR, and FAIR data are not the same as open data.

Open data are data that can be freely shared and used. The FAIR principles account for ethical, legal, and commercial restrictions and considerations in research when sharing data.

Following the FAIR principles throughout a research project promotes effective research data management, which remains essential even when data cannot be shared as open data.

The FAIR principles were first introduced in the article "The FAIR Guiding Principles for scientific data management and stewardship".

The Norwegian School of Sport Sciences has guidelines and procedures to ensure that research complies with regulations on privacy and data protection. See NIH's page on privacy for more information.

Data processed at the Norwegian School of Sport Sciences must be classified, as different types of information require different handling and storage. See NIH's Classification Guide for more information.

Data materials must be securely protected throughout the entire research project, and their storage must be carefully considered. See NIH's Storage Guide for more information.

The Sharing Guide explains how to share data internally and externally based on the classification of the data. See NIH's Sharing Guide for more information.

Documentation of research data involves descriptions of what you do and have done with the data throughout a project. This enhances the quality of the data and results by making findings verifiable. Documentation is usually noted continuously as an integral part of the research method.

Unlike metadata, documentation does not need to follow a specific standard, and the way you document the research process may vary across disciplines.

When data is to be archived, the documentation should be included in a ReadMe file. It is important that the documentation at the time of archiving contains general information about the project, licensing, reuse information, an overview of the files, methodological information, and data-specific information about the content of the various files.

Metadata is structured and standardized information about research data that enables datasets to be found, understood and reused. Metadata must be readable by both humans and machines.

Types of Metadata

• Descriptive – For example, author(s)/researcher(s) name(s), document title
• Administrative – For example, creation date, versions, and licenses
• Structural – For example, relationships between files, meaning of variables
• Technical – For example, information about the file format and any required software or hardware

When research data is archived, information about the dataset must be provided in a standard format (metadata), often accompanied by a ReadMe file that serves as a guide to the dataset. Publishing high-quality metadata and comprehensive documentation in ReadMe files helps ensure that the data can be cited and reused responsibly.

Metadata Standards

There are many different metadata standards, ranging from generic ones suitable for all disciplines to specific standards tailored to particular fields and disciplines. The choice of repository will often determine which metadata standard should be applied.

The Metadata Standards Catalog (The Research Data Alliance) allows you to browse metadata standards by discipline.

A ReadMe file is a document that explains the dataset, making it possible to understand and reuse the data. It typically includes information such as:

• Contact details
• Documentation of the dataset’s origin, completeness, and limitations.
• Dataset format, standards, licenses, codes, or required software.

Writing a ReadMe file is part of the documentation process and is done while working with the data. It is helpful to use a template when writing a ReadMe file. DataverseNO provides a general ReadMe file template for datasets and another template tailored for archiving software or code.

The file should be saved in plain text format (.txt) and named something like "00_ReadMe.txt" to ensure it appears at the top of a file structure. The file should be uploaded along with the data files.

File Names
Effective file naming ensures files are easy to locate and understand.

• Use short, descriptive names with general information first.
• Avoid spaces. Use Hyphens (First-Study), underscores (First_Study), or CamelCase (FirstStudy).
• Avoid special characters (e.g., %¨^&åøæ ...).
• Use the international data format (YYYYMMDD).
• Add a version number at the end (e.g., 01 or 001).

Effective file names convey a lot of information. It can be helpful to explain details such as dates, project acronyms, and researchers' initials in a ReadMe file.

Folder Structure
Set up a folder structure that is easy to adjust as the project evolves. Use descriptive names for folders and organize files into subfolders instead of placing everything in one large folder. Separate active work from completed part of the project for better organization.

When preparing to archive your research data, it is important to choose file formats that support long-term preservation and reuse. Open, standardized file formats are recommended because they:

• Facilitate easier sharing and reuse.
• Ensure compatibility with various software tools.
• Help data remain accessible and readable over time, even with technological changes.

Examples of suitable file formats include plain text (.txt) for documents or .cvs for tabular data.

Much research relies on existing data, and an important part of preparing a research project is mapping previous research.

You can find existing research data by searching various databases and data repositories. There are national archives and services for archived research data, international repositories and databases, as well as several metadata repositories that aggregate metadata about archived research data from various sources. These metadata repositories can be useful not only for locating archived datasets but also for identifying relevant data repositories within different disciplines.

If you use existing data, it is important to cite the dataset.

Metadata Repositories

• BASE (Bielefeld Academic Search Engine)
• DataCite
• Google Dataset Search
• Elsevier Data Search/Mendeley Data Search
• Re3data (Registry of Research Data Repositories)
• OpenAIRE

International Data Repositories

• CESSDA Data Catalogue
• Zenodo

National Resources
It is possible to apply for access to data from public databases and registers.

• Find data for research (sikt.no)
• The Survey Bank - Sikt's Research Data Catalogue (sikt.no)
• Microdata (microdata.no)
• Norwegian Health Data (helsedata.no)
• Data.norge (data.norge.no)

When reusing data, it is important to document the following:

• Where the data comes from
• Which version was used
• How the data was collected
• Who is responsible for the data
• Does the data need to be cleaned or processed before use in the project?
• Other information necessary to verify the results

Clarifying the License
Before reusing research data, it is crucial to identify the type of license or agreement that outlines the permitted uses of the dataset.

When reusing research data, the data must be cited.

Research data used in publications should be cited. Whether the data is self-produced or obtained from other sources, it is important to credit the original creator(s).

Citing datasets promotes recognition of research data as valuable research resources in their own right, while also giving credit to the creator(s) who make their data available for sharing and archiving. Read more about why citing data is important.

When citing a dataset, include information such as the author, title, publication date, repository or publisher, version number (if applicable), and DOI or another persistent identifier.

Datasets are cited according to the applicable citation style, with in-text references to the dataset and a full reference included in the reference list.

In addition, the reference can be included in a data availability statement, which may be a requirement from the journal.

Several of the useful software development tools and services used for version control and collaboration, can also be used to make code and software openly available. Some commonly used examples include GitHub, GitLab and Bitbucket.

In addition to sharing code or software through services like these, it is recommended to publish them in a suitable repository. This makes the code citable, reusable, and ensures long-term preservation.

• For projects using GitHub, the integration with Zenodo is a good option. Zenodo, operated by CERN/EU, is a general repository for research outputs.
• Many general research data repositories also support archiving and publishing source code and software.

Open licences for source code and software are distinct from the licences usually recommended for academic publications or data. It is recommended to use one of the dedicated software licences, as these are made specifically considering source code and how this can be altered and distributed. Some software license also include patent rights in addition to copyright. See also the Creative Commons FAQ about why CC-licences are not recommended for software.

• European Union Public Licence (EUPL)
• OSI Approved Licenses (openscource.org)