Scholarly journals are increasingly requiring the sharing of associated research data as a condition of article publication. Many publishers require a Data Availability Statement to be included with submissions, and some either recommend or require that data be deposited in a publicly-accessible data repository.

You often can find data sharing policies in the “Instruction for Authors” or “Author Guidelines” sections of the journal.

Examples of author guidelines with a data sharing requirement:

GigaScience

"GigaScience requires authors to deposit the data set(s) supporting the results reported in submitted manuscripts in a publicly-accessible data repository. ... This section should be included when supporting data are available and must include the name of the repository and the permanent identifier or accession number and persistent hyperlinks for the data sets (if appropriate). The following format is recommended:

"The data set(s) supporting the results of this article is(are) available in the [repository name] repository, [cite unique persistent identifier]."" (from GigaScience Instruction for Authors).

 Canadian Journal of Fisheries and Aquatic Sciences  

"Supply a data availability statement that says whether any, all, or portions of the data underpinning the work are available to others.

• If data are available, specify how data can be accessed and under what conditions data can be reused. Supply repository name, persistent unique identifier (PID: DOI/compact identifier/accession number), and web link.
• If data are not available, explain why (e.g., describe the ethical, legal, or commercial restrictions)."

(From Canadian Science Publishing Author Guidelines)

"For primary biodiversity data authors are strongly encouraged to place all species distribution records in a publicly accessible database such as the national Global Biodiversity Information Facility (GBIF) nodes (www.gbif.org) or data centres endorsed by GBIF, including BioFresh (www.freshwaterbiodiversity.eu) for freshwater data and the Ocean Biogeographic Information System (OBIS, http://www.obis.org/) for marine biodiversity data, which also holds supporting measurements taken alongside the species occurrence data." (from Canadian Journal of Fisheries and Aquatic Sciences Scope of the Journal and Guidelines for Papers).

Major Publisher Data Policies

• Black & Wiley
• Elsevier
• SAGE
• Springer Nature
• Taylor & Francis

Publisher Recommended Repositories

Some scholarly publishers provide resources to help submitters choose a data repository for deposit, including recommending specific repositories that meet their data sharing requirements. For example:

• SpringerNature repository guidance

• PLoS recommended repositories

When sharing research that involves sensitive data, protecting confidentiality is critical. Sensitive data includes personally identifiable information from human participants, as well as confidential industry data

• Research that includes the collection of sensitive data, such as some health research, must ensure that privacy is protected, and that confidentiality extends to data deposit. Creating a Data Management Plan (DMP) at the start of the project is highly recommended.
• Researchers who plan to deposit research data collected from human participants must ensure that those plans are included in their research ethics application.
• Informed consent documents must include a separate provision for data sharing. Consent to collection of data for a specific study does not imply consent to sharing of that data for reuse.
• Prior to undertaking the research, determine if the data will need to be de-identified or anonymized. This task can be time consuming, may affect the research project timelines, and may affect the budget.
• When choosing a repository for depositing sensitive data, the security measures, confidentiality review, and access controls offered will be a key consideration.

Data citation is a standardized method for secondary users of data to provide credit to the data's creator, just like citing books or journal articles referenced in a research paper. Researchers are encouraged both to cite the data of other researchers, and create data citations for their own datasets to increase the likelihood that their data will be cited.

The benefits of data citation include:

• the acknowledgement of research data as a valuable contribution to the scientific record
• the ability to reference data as standalone research output
• providing appropriate credit to data creators
• supporting reproducibility and verification of study results
• enabling reuse of data for future research
• allowing researchers to track the usage and impact of their data

The Joint Declaration of Data Citation Principles has established eight principles to guide the development of data citations. These principles are: importance, credit and attribution, evidence, unique identification, access, persistence, specificity and verifiability, and interoperability and flexibility.

There are various ways to format a data citation, but the most important components are: creator, publication date, title, publisher, and identifier (e.g. DOI or Handle, preferably formatted as a URL). A user should be able to find the exact same dataset using the information given in the citation.

A basic, widely-accepted citation format:

• Creator (PublicationYear): Title. Publisher. Identifier

Example:

A more comprehensive citation may include additional elements, most commonly: version, ResourceType (in this case usually [dataset]), date accessed (if no version is given but the dataset is likely to change over time), subset identifier/query PID.

• Creator (PublicationYear): Title. Version. ResourceType. Publisher. Identifier. Date Accessed

Example:

Data repositories (e.g. Dataverse, Dryad, ICPSR) often provide a default data citation for each deposited dataset, in a variety of different styles. Further, traditional citation styles are starting to prescribe standards for citing data. For example, APA has a template for citing published data, and one for citing unpublished raw data.

Metadata is used to describe data so that other researchers can find it and use it appropriately. There are different metadata standards to choose from depending on your area of research. Some examples of metadata used for a research dataset include:

• title of the dataset
• creator(s)
• date (created or published)
• method used to generate the data
• source of the data
• terms to describe the content
• technical descriptions including file names, formats, versions, etc.
• access information such as license and download links or data request forms

Support services for your chosen repository should be able to assist with determining what metadata to record. The most common metadata standards used for data management are Dublin Core and the DDI (Data Documentation Initiative).

Some disciplines have their own metadata schemas. Each schema has its own specified elements and structure.The Metadata Standards Catalog is a collaborative, open directory of metadata standards applicable to research data.

 Here are some examples from Curtin University Library.