Share and share alike!

Share this on social media:

Kirsty Merrett offers some heartfelt advice on how to help researchers throughout the research lifecycle 

My time at Bristol University started as an undergraduate then continued with an MSc and PhD, and library work supported my postgraduate endeavours. I later took an MSc in Information and Library Management, as librarianship felt like the career for me. 

I work for the University’s Research Data Service, where we help researchers throughout the research lifecycle, advising on Data Management Planning, data storage and data sharing through the University’s data repository, data.bris. The most enjoyable part for me is dealing with sensitive data, particularly data concerning human participants, and how we can share it securely and safely. Talking to researchers about consent, ethics, and anonymisation is invigorating. I can really get my teeth into it and draw from my own research experience as a qualitative researcher - having a foot in both camps really helps when relating to researchers. 

UKRI’s first common principle on data policy is 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’ and we wholeheartedly agree. 

We published our first dataset in 2012 and have published 511 native datasets since, so you can tell we're passionate about maximising the reuse of publicly funded research data! A great deal of data created here is from publicly funded projects, and we have a strong infrastructure to store and share it. There are many reasons why data should be shared, the most obvious of which is that funders get more bang for their buck, because we can accelerate scientific discovery when other researchers can use shared data as a starting point, even if they are in different fields. 

We also want published papers supported by the data underpinning it wherever possible, and ideally, data examined in the peer review process, so others can test that the research is robust, and the claims are credible. There can be tensions between researchers’ sense of ownership of the data and the expectation of funders to share it, and this is more challenging in projects where there are risks of participant disclosure. As a lapsed social scientist, I completely understand researchers’ reservations about sharing such sensitive data and damaging researcher/participant trust. But it can be done safely, and we believe we’ve worked out how to do that. 

For the first few years of data.bris the datasets were ‘open’, but we started to ask questions about how to share sensitive data, particularly that involving human participants. In 2017 we turned our attention to this issue so we devised a plan for providing different levels of access – open, restricted and controlled, and designated a committee to oversee decision-making of the highest level of access, in line with recommendations from the Expert Advisory Group on Data Access

We share the most sensitive of data through controlled access at our repository, data.bris. But there are several issues we need to work through to achieve this position. How do you check applicants are 'bona fide' researchers? What governance do we need in place to share data? Who needs to sign agreements? Who can we share it with, and what countries? What about multi-institution applications? What about legacy data where consent to share wasn’t raised? When you set these questions in the context of those tensions which sometimes arise around funder requirements for sharing, researcher’s sense of ownership of the data and time constraints on researchers, another issue arises: how do you provide a robust procedure without making it too cumbersome for the applicant or the data steward to go through the process? We’ve discovered that whilst we have a pretty robust and transparent process, inevitably new questions arise with each application. It’s a steep learning curve. 

We've been working on this for three years, and we've got some best practice ideas on how to share sensitive data responsibly. It takes a lot of administration time, but it’s rewarding to finally release a dataset that may not have been shared, or shared safely, otherwise: 

  • Have a standard Data Access Agreement drawn up;
  • Devise a workflow, including how you will verify applicants are bona fide researchers and how you will obtain supporting documents, detail who needs to be involved, and at what stage;
  • Make the process transparent - clearly define what is required to apply for data, the anticipated time frame, and processing costs, if any apply; and
  • Set up an independent decision making committee drawn from across the university.

Our top tips are to get niggles or anomalies with the application sorted out before asking for a decision, and to be prepared for multi-threaded questions involving different parties: the data steward, the applicant, secretary’s offices, research ethics committees, IT services. We presented on these issues at the International Digital Curation Conference last year, and we've had a few visitors from overseas come to us to discuss how we do what we do. We love to share!

Kirsty Merrett is an assistant research support librarian at the University of Bristol.

 

Do Kirsty's experiences chime with you? Would you like to share your thoughts as a librarian with the readers of Research Information? Email tim.gillett@europascience.com