Because good research needs good data

JISCMRD: Developing and implementing RDM policy

Jonathan Rans | 31 October 2012

The development of research data management (RDM) policy has been a hot topic over the previous year with the EPSRC’s deadline of May 1st 2012 for the production of a ‘roadmap’ strategy document driving activity in this area. Now the focus is shifting, with institutions looking to the future and asking how these aspirational policies can be turned into practical reality.

The JISC Managing Research Data (JISCMRD) Components of Institutional Research Data Services workshop policy session explored these issues with presentations from four institutions outlining the progress that they have made to date and the challenges that they have had to grapple with along the way.

Harmonising policies

Tom Parsons discussed the University of Nottingham’s ADMIRe project, which aims to develop policy and infrastructure supporting RDM across all disciplines using a top-down, bottom-up approach to implementation.

One problem the project is dealing with at the moment that is no doubt common amongst HE institutions is that of policy harmonisation. Nottingham now has an institutional RDM policy but also has a Research Ethics and Conduct policy that covers elements of RDM and an EPSRC roadmap strategy. These all serve slightly different purposes and conflict with each other and with external policies produced by the funding councils. When looking at ways of resolving these conflicts an interesting point was raised: is it always better to produce yet another new policy or could existing policies that enjoy traction and authority be used as vehicles for RDM policy?

The consensus was RDM is important enough to warrant its own policy rather than being buried somewhere that may not be completely appropriate and gives the opportunity to overwrite older, obsolete policies.

Investing time in robust strategy

Cathy Pink, from the University of Bath’s Research 360 project, discussed the process of adapting Monash University’s RDM strategy plan to produce a roadmap document that satisfied the EPSRC’s requirements. This highlighted one problem that many people are experiencing, the sheer time that it takes to produce a good, sustainable strategy. Quite apart from the considerable effort required to adapt an existing policy to an institution whose context and aspirations are different, the level of consultation and advocacy required to engage with both senior management and academics is quite high and can really slow down progress. However, if the aim is to produce a robust policy that can be properly applied and adhered to, that consultation is vital. At Bath, having the close involvement of the PVC research at every stage of the development process was seen as key to producing a plan with realistic goals that could be backed by senior management.

Cathy felt that there are fundamental questions about funder’s expectations for institutions that need to be answered before serious, fine-grained policy work can continue, most notably:

  • Will funders move the goalposts and produce more requirements that will need to be accommodated?
  • How realistic are the risks of audit to an institution?
  • How will RDM costs be funded, what can be built into grant applications and how should that be done?

This was an area where the DCC would be well-placed to play a central role, using its position to help clarify policy decisions and relaying information to funders about how policy is affecting RDM at a grassroots level.

Understanding liability

Birgit Plietzsch, from the University of St. Andrews, gave us a run-down of their institutional engagement, which included the development of a grant portfolio picture, the enhancement of internal staff skills and benchmarking exercises using DCC DAF and CARDIO tools.

One goal that St Andrews had hoped to achieve during the engagement was the development of an RDM policy. This was hampered by the difficulty in accurately predicting the costs of the RDM services that would need to be developed in support of a ratified policy. Vague liabilities are undoubtedly a widespread problem, even amongst those institutions with relatively mature polices. Policy developers know, for example, that some managed storage for research data will need to be provided, some even stipulate amounts, but there remains only vague understanding of the associated cost of provision.

Practical implementation

Stuart Lewis, head of the University of Edinburgh’s digital library services, discussed the work that has gone into driving the University’s landmark policy towards implementation. The Edinburgh policy has been in existence for long enough that academic staff are beginning to look for progress updates and asking when they can expect infrastructure updates to be complete. He felt that this is where the production of a roadmap, ostensibly to meet EPSRC requirements, has been useful, providing a framework for development linked to policy that can be made available to staff.

The importance of senior academic involvement at all stages of policy development was highlighted again; Edinburgh has representation from all three of its colleges on the policy steering committee and their involvement is seen as crucial to ensure maximum engagement from research staff.

Implementation of the Edinburgh policy is being progressed by the Implementation Committee's Action Group, whose membership is made up exclusively of people who ‘do stuff’ and can concentrate on the practical aspects of policy realisation. This is being progressed through a series of pilot studies of which there have been four to date; using the DCC’s DMPonline tool, working with the Datashare repository, examining active infrastructure and how storage is accessed, and developing training for liaison librarians.

One very hearting report to come back from the Edinburgh experience is that there is a strong appetite amongst Information Services staff for RDM training, evidenced by the speed with which the University's MANTRA course became oversubscribed.

Discussions

It became clear during the following discussion session that one of the main issues exercising participants is how to accurately bring costs into policy and strategy documents. Lindsay Wood, from the University of Newcastle’s Iridium project, asked whether there was any institution that had bucked the trend, starting with a business case and using that as the basis for policy development. Wendy White, of the Datapool project, told us about the University of Southampton, which has taken that approach and is fairly unique in having adopted a ten-year business plan for RDM. Wendy said that, while it is very useful to have figures on the table when discussing RDM with senior management, they are now being asked to put the figures into context and provide a narrative to demonstrate why proposed services are necessary for research.

Another talking point was whether to explicitly write unfunded research into RDM policy. It seems that a general lack of information about the value of such research to institutions is causing it to be dealt with fairly vaguely with only one institution explicitly referring to unfunded research in policy.