RDM in higher education

Editor's Notes

• #3: I’ll start by giving a brief introduction to the DCC, then I’ll cover what the drivers are for research data management in the HE sector.The main part of the talk will focus on key initiatives and trends in our sector – data management planning, the development of policies and strategies by universities and the infrastructure to support RDM.Finally I’ll cover how this work is being supported by JISC, the DCC and data centres.
• #4: The DCC is a JISC-funded centre to support universities with Research Data Management (RDM). Our aim is to build capacity, capability and skills across the HE research community by working with libraries, IT services, research offices and researchers.
• #5: We build this capacity in a number of ways: by providing guidance, running training and events targeted at different audiences, developing tools such as DMP Online which I’ll explain later, and also by supporting the JISC, particularly the MRD programmes that it funds.
• #7: There are a number of policy requirements pushing for research data management and sharing. The OECD declaration released in 2007 is the main overarching one. This essentially states that data are a public good – if they’ve been funded by the public purse they should be openly available wherever possible. RCUK has adopted this principle and released codes of conduct for research and more recently a set of common principles on data policy. Universities are starting to develop their own data policies too, a trend I’ll touch on later.So there are various policy drivers, as well as legislation such as the DPA and FoI which have an impact, but there are also calls from the community to manage and share data.
• #8: There’s a real push for increased openness and sharing from certain parts of the research community. There was an issue of science in 2011 which noted the importance of RDM and the need to re-skill and develop infrastructure. Initiatives such as the Open Knowledge Foundation and the Panton Principles have also emerged to promote open sharing – researchers have recognised that it leads to better science.
• #9: There are various benefits or rewards to managing data. In the HE sector you’ll often hear people talking about the ‘data behind the graph’. Researchers publish results but you need to be able to access the underlying data to validate these findings. And various studies have shown that researchers get more citations when they share data, receiving the recognition and rewards that they need. If data are manage well it’s also much easier for researchers to do their work so there’s a self-interest argument too.
• #11: The need for DMPs is being driven by research funders. Most of RCUK and some key health funders expect researchers to submit a short Data Management and Sharing plan in their grant application
• #12: Typically researchers are asked to cover what data will be created and how to ensure it’s reusable. They should note any anticipated difficulties, for example if ethical concerns mean that some of the data can’t be shared or if they need an embargo period to file for a patent to make sure that these limits to sharing are justified. The longer term plans for sharing and preservation also need to be outlined.Essentially the DMP is to reassure the funder that researchers will comply with their policies and share data wherever possible.
• #13: The DCC has provided a variety of support in this area: we’ve developed a checklist of possible questions you could be asked in a DMP our how to guide offers practical guidance to walk researchers through the process of writing a plan, explaining what funders are looking for in each question and DMP Online is a tool to help researchers write plans
• #14: At the outset in DMP Online researchers pick templates to suit their needs e.g. a funder template if applying for funding, or an institutional template if their uni has requirements. They then answer the questions based on the guidance and any suggested answers provided. We’re increasingly developing institutional customisations to build in details of local support.
• #16: The development of policies and strategies is also being driven by funder requirements, particularly the EPSRC data policy. Rather than asking researchers to write DMPs, the EPSRC has placed the onus on institutions to make sure that the appropriate infrastructure is in place to support RDM. They’ve asked unis to develop a roadmap by May this year and then to ensure the support and infrastructure is developed in the next 3 years.This has caused a real stir, particularly among university senior management. Lots are now very engaged and investing in this area for fear of losing funding.
• #17: The DCC has been collating resources to support unis to meet this requirement. We developed a series of blog posts earlier in the year, providing guidance and profiling emerging approaches. We’ve also linked to roadmaps that have been made public and more recently conducted a survey to identify current practice and what support needs there will be over the next 3 years.
• #18: Allied to the development of strategies is RDM policy. Edinburgh released the first policy in May last year and this has been much copied. Lots of unis have released policies since and c.15 more are currently being ratified so the trend will continue.Many of these policies reference the university’s commitment to supporting RDM by developing infrastructure and support, so that’s the final trend I’ll touch on.
• #20: Lots of unis have done surveys of existing RDM practice and one of the most alarming results across the board was how researchers are storing data. Often they’re not using the centrally provided networked storage – either due to a lack of capacity or convenience to access and share their data. Practice is very ad hoc with researchers storing data on USBs, external HDDs or their laptops.Many unis are now investing heavily in research data storage. At Bristol they’re using their HPC facility, a petascale data store called Blue Peta. They’re offering the 1st 5TB for free per nominated ‘data steward’ - researchers have to sign up as data stewards to make sure that someone is responsible for the data, can say who can have access and how long it needs to be kept. There’s a clear, upfront charging model too so any additional capacity needed can be costed into grant applications.Having the storage is only part of the issue though. You need the right applications layered on top of this and that’s where the problems often lies.
• #21: One prevalent issue is the need for researchers to have easy access to their data, so you’ll hear talk of an academic dropbox. There’s a blog post from the University of Bath which outlines this issue very clearly.Researchers are mobile. They’re often not working in their office or on campus, they’re in the field, at conferences, at other unis with collaborators... and they may not have a good network connection so need a local copy of their data for fast, anytime access.Many have turned to dropbox because it’s so easy to use. It provides a local copy that will sync back and they can share data with collaborators easily - no need to contact sysadmins or encounter difficulties providing access to anyone outside the uni.There are security and legal issues for universities if researchers use dropbox though – where is the data being held, under what licence, is it secure?
• #22: Different solutions are being piloted. At Oxford they’ve developed their own software. DataFlow has two components – DataStage and DataBank. DataStage provides the dropbox-like facility. At Lincoln they’ve been piloting owncloud, an open source tool that provides the same features as Dropbox.
• #23: The holy grail that everyone is aiming for is to bring these different services together. You can see here how they’re proposing to do this at Oxford. Very few institutions are anywhere need this stage though, just Oxford, Edinburgh, Southampton... and in no case is it embedded and fully functioning – these are all pilots. Most unis are still at the stage of identifying their problems and development components of an overall service.
• #25: I’ve spoken about lots of work going on in the JISC MRD programmes. This is really the main source of seed funding in this area.Simon Hodson is the programme manager if you want to find out more, and there’ll be a closing conference next March where they’ll showcase the outputs
• #26: The DCC also has a few strands of activity to support this work. We’ve run a series of roadshows to profile good RDM practice and get key stakeholders from an institution together to start to plan their strategy.
• #27: The main strand of activity for DCC is our institutional engagements. We’re working intensively with 20 unis, providing 3 months of consultancy effort to each to support them in any areas they like – it may be developing strategies, piloting services, delivering training programmes...
• #28: All of the lessons from MRD and DCC work are currently being pulled together in a how to guide on developing RDM services.
• #29: And finally I just want to note that it’s a mixed model of support in HE. I’ve spoken about what unis and the DCC are doing but there are various other national services and particularly data centres, which play a key role. The UK Data Archive provides excellent guidance and support for social scientists, NERC funds a number of disciplinary data centres in the environmental sciences and there are many more.
• #30: Thank you!If you want to find out more about DCC we have lots of resources online and you can follow us on twitter.