Why do I need to manage my research data?
In recent years there has been a significant drive by research funders – and in some cases researchers themselves – to encourage greater openness with research data. This has been partly in response to concerns about the non-reproducibility of research and the potential for malpractice, but also in part to facilitate data reuse and aggregation. Another driver is the Research Councils UK Common Principles on Data Policy, which states that ‘publicly funded research data are a public good, produced in the public interest, which should be made openly available with as few restrictions as possible in a timely and responsible manner that does not harm intellectual property’.
In order to enable greater openness, research data needs to be discoverable, accessible, and described in such a way that it is intelligible to others. Where there is value in doing so, it also needs to be preserved and curated for the long term.
Besides benefiting the broader research community, data that is well managed is also likely to have more immediate benefits to the research group that created it and their collaborators. By storing, documenting, and preserving data efficiently, it should be easier for researchers to find the data they need when they need it.
If you are interested in learning more about the motivating factors behind the drive towards research data management, take a look at:
How much detail is required when documenting data?
There is a certain minimum amount of information about your data (metadata) that is required to ensure it can be properly cited: the names of those responsible for creating the data; a title; a publisher; and the publication year. This will generally be required during the data deposit process, along with information about how and why the data was generated. Most data repositories have a form to fill out when depositing data, and it is a good idea to see what information they ask for before you get too far into a project.
It is always sensible to add some documentation to a dataset whilst you are still working on it – explain abbreviations, and add notes about data that seem odd or which may cause confusion to those not involved in generating the data. Not only will this assist researchers who may wish to look at the data in future, but it will also help you and other members of your team to understand it if you need to revisit it yourself in a year or two’s time.
What is metadata?
In the context of research data management, ‘metadata’ is the contextual information about the data that will help others to find and understand it. This will usually include information such as ‘who created the data’, ‘what is the data about’, ‘are there any restrictions regarding who can use the data and in what circumstances’, and so forth. Different disciplines will generally find different metadata fields useful. Library catalogues are essentially catalogues of metadata.
In some disciplines research publications are often the richest source of information about how a particular dataset was derived, so it is important to link articles to data. In others it may be necessary to develop additional documentation about how the data was collected, organised and used.
Is it OK simply to instruct people who want to look at my data to call me or send an email?
One of the principles behind funder expectations is that ‘funded research data is a public good produced in the public interest and should be made freely and openly available with as few restrictions as possible in a timely and responsible manner.’
A simple direction to interested parties to contact the author would not normally be considered sufficient. Decisions about data archiving, preservation and possible future sharing of data need to be made.
Will the University offer help and support for meeting data management requirements?
The University will support researchers in meeting funder (ESRC, EPSRC, MRC etc.) expectations or requirements via advice and guidance. It offers services and infrastructure to support various aspects of research data management.
Enquiries relating to research data management should be directed to firstname.lastname@example.org, where they will be reviewed and addressed by a cross-departmental team of staff from IT Services, the Bodleian Library, Research Services, and the Oxford eResearch Centre.
Information about software, services, and good practice is available on the rest of this website.
The data upon which I based my analysis was not generated by me or my team. What should I do?
You do not need to deposit existing data or that belonging to a third party unless you have materially altered or added to it. If you have substantially altered the data – and this includes restructuring it so as to enable analysis – then whether you can or should deposit it will depend largely on the intellectual property (IP) rights invested in the data and the licence it was published under (if applicable). Unless it is clear that you have the right to publish the data in its modified form, email email@example.com for further advice.
The data I’m producing needs to remain private. Does this mean requirements regarding archiving and sharing don’t apply to me?
Funders expect data be be archived whenever possible. This does not necessarily mean it will always be shared publicly or without restrictions. In such cases it is recommended you produce a publicly-accessible record of your research data in ORA-Data.The record for the data should indicate why access to the data is restricted or not possible. For example, Expectation vi from the EPSRC guidelines on archiving and sharing data states: “Where access to the data is restricted the published metadata should also give the reason and summarize the conditions which must be satisfied for access to be granted. For example ‘commercially confidential’ data, in which a business organization has a legitimate interest, might be made available to others subject to a suitable legally enforceable non-disclosure agreement.”
If you believe that even publishing a record about your data would be problematic (e.g. it might expose matters of national security), please contact firstname.lastname@example.org.
Software and computer code
If the ‘result’ in a paper is the demonstration of a novel piece of software, does the software count as data?
Probably not. If data has been generated as a result of running software code, then it may be helpful to provide a link to that code in the metadata, but the software itself would not constitute data in most situations.
If custom written software is used to process the data do instructions need to be provided on use of this software?
If the software is essential to validating the research findings then adequate information should be provided to enable its re-running by third parties. This may involve taking additional steps to preserve the software in addition to the data itself.
I produce computer code and/or simulations based on data supplied by other research groups. How do funder expectations affect me?
You may wish to preserve your code (and in some instances the environment in which it was run) in order to adhere to the general spirit of making research more widely available. In this instance consider asking the data creators to deposit their data in an appropriate repository that will assign a Digital Object Identifier (DOI) to the data, so that you can then reference it.
The Software Sustainability Institute may be of interest.
ORA-Data and other archives
I have data that I need to deposit in an ‘appropriate’ data archive. How do I find such an archive?
An extensive directory of data repositories is available from re3data.org. These range from very generic commercially-provided repositories such as Figshare, to narrowly-defined subject-specific repositories. Generally speaking, it’s better to use a subject-specific repository than a generic one, as they will have staff that understand the data and can help curate it properly as time passes.
Some repositories request that the data they receive is accompanied by metadata (contextual information about the data) in a particular format, so it’s worth getting in touch with appropriate repositories before you get too far in to the data gathering process. It’s much easier to document your data as you gather it rather than leaving it until the end of a project – and documenting data during a project can also help you and any collaborators find relevant information more quickly whilst you are still working on it. Feel free to discuss this with your Subject Librarian or arrange to meet with one of the RDO support team by emailing email@example.com.
Unfortunately, subject-specific data repositories do not exist for many fields. If there is no appropriate disciplinary repositories for your data, you can meet funder requirements by depositing it in Oxford’s institutional data repository: ORA-Data.
Even if you deposit your data to an external data repository, you should still create a record for it in ORA-Data, so that the University can keep track of research outputs. This is increasingly expected by research funders, and can help with the assessment of impact.
When should I deposit my data in ORA-Data rather than another data repository?
ORA-Data is the University of Oxford’s institutional data repository. This does not, however, mean that all research data you wish to preserve should go into ORA-Data. If there is a specialist data repository for your discipline, you should under normal circumstances deposit your research data there rather than in ORA-Data. You can find out more information about specialist data repositories from re3data.org.
You should ONLY deposit your data in ORA-Data if there is no more appropriate specialized data repository in your field. You should however create a record for your data in ORA-Data even if you deposit the data itself elsewhere. This helps the University know what and where it is in the event of an audit, but it also improves the visibility of your data, as ORA-Data records are indexed by search engines such as Google.
You may wish to deposit your data in a general data repository such as Dryad or Zenodo, or a commercially-provided alternative such as FigShare. These may be convenient, but there are reasons why these are not advised as alternatives to specialist repositories or ORA-Data for data underlying published research conclusions:
- ORA-Data, and to some degree specialist repositories, are likely to have better longevity than free generalist (and/or) commercial repositories.
- ORA-Data and specialist repositories are likely to be able to offer a better level of post-deposit curation in the future than free services (things like format migrations and integrity checking).
- ORA-Data and (most) specialist repositories include a metadata review to ensure that minimum standards are met.
- With ORA-Data the data will be held within Oxford, so there are no concerns about legal jurisdictions; specialist repositories may be based outside of the UK, but usually make their terms and conditions very clear.
- Commercial and free services may lack strong long-term business models. Check their terms and conditions to see what will happen to your data should things turn sour.
Finally, some journal publishers accept data deposits alongside the articles they publish. If depositing in a publisher’s data repository, check that the terms and conditions meet your funder’s minimum expectations, and create a record in ORA-Data.
For more detailed advice, see the Archives and Other Options page.
Can I use ORA-Data at any time, or only at the end of a project?
You can deposit data into ORA-Data or create records for data in ORA-Data at any point in a project. Some funders such as the EPSRC indicate that a record describing your research data should normally be made available within 12 months of the data being generated, even if access to the dataset itself is restricted.
Remember that if you need to pay to deposit the data your project has produced, you will need to complete the deposit process and payment before your grant expires.
Can I deposit my data in free archives such as Figshare, Zenodo, or Dryad?
Yes, but… Be careful that the repository can meet the terms and conditions of your funder, and remember to add a record in ORA-Data.
Generalist and commercial services such as Zenodo and Figshare are usually free, they offer the immediate assignment of a Digital Object Identifier on deposit, and may come with visualization tools.
They may also come with significant risks, however, particularly with regards to their longevity. Free services in particular should be handled with caution – what measures do they have in place to guarantee that they and your data will still be around in three years? In ten? Check the terms and conditions to see what will happen to your data if the service disappears.
How do I…?
I’m putting together a project bid and need to complete a data management plan. How do I go about doing this?
Most of the major funding bodies provide a data management plan (DMP) template as well as guidance for completing the plan. You can find summary information about funders’ DMP requirements on the Funder Requirements page and the Digital Curation Centre website. Before going any further, it’s also worth visiting your funder’s own website to ensure you are referring to the latest versions of the template and guidelines where available.
The Digital Curation Centre offers a useful web service for completing DMPs: DMPonline. This presents the technical plan template as an online form, alongside guidance and tips for completing each entry from both the funder and the University (where such advice is available).
If you would like more detailed advice as to what to include in a DMP, arrange to meet with one of the RDO support team by writing to firstname.lastname@example.org.
I need to complete an information security questionnaire in order to access and use a sensitive dataset. How should I go about completing the questionnaire?
Such questionnaires will usually seek reassurances about the security of the computing facilities and infrastructure that you will be using. There may be restrictions regarding where you can access the sensitive data, such as a requirement that you must be within the physical University department in which you are based (and therefore behind appropriate firewalls) or that you will only ever access the data from a specific computer terminal. If this is the case then your departmental IT support staff will probably need to answer many of the technical questions relating to infrastructure and storage, including the set-up of the computers, how they are maintained, and how they connect to the rest of the University and ultimately the outside world.
Information about data encryption, passwords, email, mobile devices, and other general security issues is available from the InfoSec team at University of Oxford IT Services. Take a look at the InfoSec website for a host of useful tips. Guidance on the University’s Information Security Policy is also available.
Further questions on this can still be sent to the RDO contact email, email@example.com – the team reviewing messages includes members of IT Services.
I would like to store my data securely within the University. What are my options?
Storage for data in current use (as opposed to archiving for data after a project concludes) is usually provided at the departmental level in Oxford. Ask your departmental IT Officer if they have any server space that you or your research group could use.
Some of the tools and services provided by the University for researchers include space for data and document storage. The SharePoint Online service, for instance, provides up to 1 TB, and OneDrive for Business (part of the Nexus365 suite of tools) offers 5 TB.
For larger projects, the Infrastructure Chargeable Services team at IT Services offer managed server space for a fee.
IT Services provides a University-wide, centrally-funded backup service for staff and postgraduate students – known as the HFS (Hierarchical File Server). It provides an automated backup service for personal computers and servers alike.
The University is currently undertaking a project to scope and provide dedicated storage for research data.
See the Backup, Storage, and Security page for more on this topic.
I need to include a Digital Object Identifier (DOI) for my data in my article submission. How and when do I get one?
Most data repositories will assign a unique identifier, most commonly a Digital Object Identifier, when you deposit you data with them.