What does EPSRC want me to do with my research data?
The EPSRC (Engineering and Physical Sciences Research Council) requires that all publications with a publication date on/after 1st May 2015 include a statement describing how to access the underlying research data (or a statement explaining any restrictions).
The following resources will guide you through EPSRC expectations on research data management:
- Key points of EPSRC expectations
- What data do I need to include in my research publication?
- What happens if I am not compliant with EPSRC expectations?
- Frequently Asked Questions
For the official, full version of these released by the EPSRC, click here.
All publications with a publication date on/after 1st May 2015 should have a statement describing how to access underlying data (or a statement explaining why access to underlying data has been restricted).
- Data should be stored in repositories guaranteeing safe storage
- Data should be stored for at least 10 years or for 10 years from the last request for access to the data
- Researchers should ensure that data which is not generated in a digital format is stored in a manner to facilitate it being shared. Requests for access to research data should not be rejected merely because data is not digitised. Third parties requesting access to the data might be required to cover the cost of data sharing (travel to the place where the data is held, pay for the cost of data digitisation)
Claiming costs associated with data management:
- All elements of research data management are eligible for research grant funding: direct or indirect costs associated with data management must be stated in the grant proposal
- All grant applications are encouraged to have Data Management Plans
Making data freely and openly available:
- All data supporting research publications should be cited with the use of persistent links, for example DOIs (Digital Object Identifiers), and should be accessible online no later than the publication itself (also applies to online publications)
- All data for which there is no intention to publish should be made accessible online with permanent links, such as DOIs, and with a relevant description no later than 1 year after the date of generation (if this date is clear) or 1 year after the end of the EPSRC grant (if the date of generation is not clear)
- If there is an intention to publish, a publication plan must be in place
- If compelling legal or ethical reasons exist to protect access to the data, these should be stated in the research publication, or in the data description (if data is published separately, without the research publication)
- Research data should be accompanied by metadata. Metadata is the description of research data, and is used to allow the discovery of data. Metadata should include information on:
- What research data there is. For example, an Excel 2010 file with raw data supporting figure 1 in the publication xxxxx
- Why, when and how was the data generated?
- How can the data be accessed?
- If the data is restricted, metadata should provide the reason for the restriction, and explain what conditions have to be met for the access to be granted: e.g. someone who might want to access commercially confidential data might need to sign a non-disclosure agreement
Exemptions from the mandate of making data publicly available:
- Personal data should not be released, unless consent of the person is given
- Sensitive data should not be released
- EPSRC defines senstive information as "information the release of which would compromise unprotected intellectual property or which, in the judgement of the security services, would result in unacceptable risk to the citizens of the UK or its allies."
- Delays/restrictions to data publishing are acceptable if necessary to protect intellectual property or commercially confidential data
- If data preservation is not possible or cost-effective, it is acceptable not to publish the data, as long as the ability to validate published research findings is not compromised, and the access to the data might be granted in other ways
- Relevant statements explaining these restrictions need to be made in the research publication and/or in the data description (metadata)
- EPSRC acknowledges that not all research papers are supported by research data. If there is no additional data supporting a publication, a statement explaining how to access research data might be omitted. The EPSRC relies on researchers making informed judgements about when it is appropriate to include statements.
The EPSRC does not provide detailed guidance on what data should accompany a research publication. The EPSRC relies on researchers to decide what is the minimal necessary research data that needs to accompany the publication. Researchers themselves are considered by the EPSRC as the experts who know the most about their research data and are aware of their value to others.
In general, the minimal amount of data is what allows the experiment to be reproduced and scrutinised. The EPSRC suggests that there might be occasions where it is better to publish analysed, representative, or random sample examples, instead of all raw data (especially in cases of large volume datasets). If only selected data are shared, the researcher needs to explain in details the steps that led to generation of representative or random samples.
The EPSRC will check if researchers funded by EPSRC are compliant with their expectations on research data management. The EPSRC will do the following:
After summer break 2015: EPSRC will check the availability of data under-pinning published research, examining the following aspects:
- Do publications have a statement describing how to access underlying data?
- If there is no statement – where is the data?
- Is the right type of data available?
EPSRC will do random checks on papers published on/after 1st May 2015 - individual researchers might be contacted
EPSRC will investigate any complaints about research data not being managed in line with EPSRC expectations.
EPSRC aims to embed compliance checking as part of regular grant assessment by the Research Councils Audit and Assurance Services Group (AASG).
AASG might perform thorough checks on randomly selected grants to assess compliance with EPSRC expectations on data sharing.