Research at Risk

Directions in Research Data Management

Here is a post from David Kernohan to report on a workshop we recently held with ARMA, RLUK, RUGIT, SCONUL and UCISA in early November, on 6th and 7th.

Over two days in the Moller Centre at Churchill College, Cambridge, delegates debated the key barriers standing in the way of our fully realising the benefits of research data. The workshop drew together research practitioners, library, IT , research office professionals, funders and senior managers and asked them to reflect on issues, priorities and solutions across 5 key areas.

• Policy
• Skills and Capabilities
• Incentives
• Infrastructure and interoperability
• Business models and sustainability

The workshop facilitated a joint discussion for exchange, thoughts and priorities on these issues and is also an input to a roadmap that Jisc is developing with the professional associations and representatives mentioned above. The roadmap is intended to be for stakeholders in the sector for the next 5 years of developments that are required in research data management. It is fair to say that the roadmap sees compliance with existing funder requirements as a waypoint on the journey rather than the end in itself – discussion at the workshop certainly pointed to a more ambitious vision of open scholarship where the sector can create, curate and use research data outputs for better research. The roadmap is now in early draft and will be published in 2015.

There has already been a great deal of work in this area, but gaps and opportunities remain and need to be addressed. Historically, there has been a “lack of recognition that a national rather than an institutional approach would save everyone time and resource” – this has been raised in recent surveys of the sector and was a point strongly put by Martin Lewis in his presentation in Cambridge that focused on collaboration, and indeed shared services. The message is that institutions have made great strides in the UK, and indeed the work with Jisc has contributed, but now more sharing and coordination is needed. The workshop covered a lot of ground, exploring sharing and collaboration across disciplines, research teams, institutions and regions. In discussion groups we examined how incentives and status could be used to drive data sharing, and how sharing resources can support this.

Amongst delegates there was a clear desire to change the language around RDM from compliance to a more positive discourse of improved research. Current compliance drivers, it was felt, will recede into the background over the next few years as practices and processes become normalised. This was acknowledged as difficult area to navigate on the one hand mandates and compliance are an important aspect to ensure necessary change happens, but on the other hand unless we can better define the benefits a wholesale shift in practice will not be achieved.

Clearly not all research data can or should be shared, but there was a lot of support for the idea of openness as the default for data. Benefits from this include bringing further transparency to the peer review process, and allowing further research to be performed building on existing data. These are arguments everyone is quite familiar with. But building the foundations to allow this to happen is not straightforward and there is more to be done.

There were many conversations about incentives, with a lot of support for the idea of academic champions actively promoting the benefits of open practice and data sharing. Another key theme was around consistency and clarity in funder-led research data policy – with simply-stated policies being similar across multiple funders, and the need for further analysis to understand differences and bring further commonality. There was an appetite to bring the various stakeholders into discussion with funders to support the further iteration of policy development, and to ensure that data sharing was both recognised and rewarded. There were some very clear directions emerging from discussions, for example national arrangements for key identifiers such as ORCID and approaches to data citation. See the Orcid pilot which is working on this.

Clearly collaboration and shared solutions can help institutions defray the cost of infrastructure, but institutions will need to make a parallel investment including those in staff time and capability. Here, sharing approaches to resourcing this area between institutions can be very powerful. There was a fair amount of focus on the numbers of staff required for an effective research data service, largely recognising that research data should become part of existing roles.

The findings from this workshop are being drawn together alongside a range of other inputs, and early in the 2015 we will publish a roadmap in partnership with the sector bodies. The roadmap will also help Jisc refine our ideas about what is needed in the RDM space, so it will inform the Research at Risk plan which will contribute to the much called for foundation for collaboration. As well as the overarching themes I have delineated here, there were specific suggestions that will feed in to the wider Jisc work on RDM that Rachel introduced in her earlier post.

John MacColl, Vice-Chair of the RLUK Board, closed the workshop by reflecting on the need for a new prestige for data in the wider scholarly environment, with a supporting infrastructure that is of benefit to all researchers and institutions. This should be a stewarded offer, drawing on the strengths of library, IT, research support and researcher mindsets.

For now here are the slides from the workshop, and these are I think a good set of resources that set out some of the issues and the perspectives in terms of universities, researchers and different disciplinary needs, and all of these contributions were excellent in terms of setting context, giving examples of work and evidence of issues that need to be addressed – so do look at the slides if you can.

Finally I’ll mention Stephen Pinfield’s presentation on a framework that helps to outline all of the elements that need to be considered in implementing research data management. I won’t outline the whole framework here, but to give you a flavour – so drivers are storage, security, preservation, compliance, quality etc. – and the components of a research data solution are strategies, guidelines, policies, processes, services etc. – and influencing factors are acceptance, demand, roles, skills, incentives, politics, governance etc. The stakeholders involved were also set out by Stephen, these include IT, library, academic departments, senior managers and research support services. Luckily these were the stakeholders that were represented at the workshop! Clearly research data management is an issue that truly cuts across the university and needs collaboration. We hope that the roadmap produced will help to set out a shared direction to support that.

By Rachel Bruce

Rachel Bruce, Deputy Chief Innovation Officer, Jisc