Research data management: from policy to practice with DMP Online

Editor's Notes

• #2: Good afternoon. We are Martin Donnelly and Sarah Jones of the Digital Curation Centre, at the Universities of Edinburgh and Glasgow respectively. We’ll be talking today about the journey from research data management policy to good practice, and how the DCC’s resources, notably the DMP Online tool, can support this journey.
• #3: Sarah will give an introduction to the DCC and our interest in research data management, before giving an overview of the policy situation in the UK and how we got involved in data management planning.Martin will then take over, talking in more detail about the DMP Online tool, the various collaborations we’ve formed through this work, and an overview of the major job of work that we’re both currently involved in, namely the DCC’s set of institutional engagements.
• #4: The UK Digital Curation Centre was established in 2004. We’re based across three universities, and have a remit to support UK Higher Education as a whole.Our mission has changed over time from a focus on digital curation and preservation, working largely with archives and repositories, to research data management in universities.
• #5: The DCC has four main strands of activity:We develop tools to help organisations assess their infrastructure &capabilities or to undertake specific tasks e.g. writing DMPs with DMP OnlineWe run a helpdesk, which is open to all, and provide guidance. How To guides are a new range of pragmatic, practical advice.We run training and community building events. The roadshows help institutions develop research data management strategiesWe support JISC by co-ordinating events, working with projects and synthesising/disseminating findings.
• #6: The DCC developed the curation lifecycle model to explain the range of activities involved in creating, preserving and sharing digital content.In RDM terms ‘curation’ is simply managing & sharing data. The DCC argues that this is just part of good research practice.
• #7: How datasets are created and managed in the short-term affects how much work it is to ingest and preserve them. The transition isn’t always easy, which is why it’s useful to work with researchers early on to support them to make informed decisions about how to create and manage their data.The KRDS costs and benefits studies found that ingest is by far and away the most resource intensive activity.
• #9: Last year the 7 UK Research Councils released common principles to harmonise their data policies.These push for open data, acknowledge the importance of policies and planning, and cover various aspects on curating data (including meeting costs).
• #10: Basic expectations across the board are that:Data are released as soon as possibleData are shared openly wherever possibleData are preserved for 10+ yearsDMPs are submitted that outline plans for data management and sharing
• #11: The DCC has responded to these requirements by providing lots of support on data management planningLiz Lyon first called for plans in 2007 in a recommendation in the Dealing with Data ReportWe have since analysed funders requirements and put together a checklist for a Data Management PlanThe Checklist is the underlying intellectual framework in DMP Online, the flagship of the DCC’s tools and resourcesWe also provide guidance documents and have custom guidance (disciplinary & institutional) built into DMP Online
• #12: A DMP is a basic statement of how you will create, manage, share and preserve your dataFunders expect the decisions to be justified, particularly where it’s not in line with their policy (e.g. limits on data sharing)
• #13: The main questions across the board cover:Data creationMetadata and documentationEthical and legal issuesData sharing Preservation
• #14: You see the common questions come through in the main sections of the DCC ChecklistWe also include administrative sections (intro, review, ratification) so you can ensure co-ordination and commitment across all of the stakeholders involved in managing data.
• #15: So in summary, these are some of the key DMP-related resources.
• #16: The main things to remember about DMPs is that all research projects are different- the DMP will vary with context.Apart from a few very specialised areas like backup - there are no universal rights and wrongs.Research data management by nature involves multiple stakeholders, so planning is important as a communication mechanism.The process of producing a plan (i.e. engaging with others and deciding on the best way forward) is as important as the plan itself.
• #17: SJ > MDThese expectations and trends are not a UK phenomenon. Martin and I have contextualised the UK experience of data policies and planning, by reflecting on international initiatives in Managing Research DataThis is why DMP Online is relevant to international audiences, so I’ll let Martin tell you all about it.
• #18: Thanks SarahWe started developing DMP Online in 2009, and launched the first version in 2010. We’re now on to v3.0, which includes some great new features that we’re really excited about.
• #19: The DCC Checklist is by nature very long, and its length was felt to be off-putting to researchers. Most of them don’t want to deal with this stuff even at a basic level, and a long Checklist with over 100 questions was not going to enjoy a large takeup.No matter how many times we said “you don’t need to fill it all in, just the bits that are relevant to you at this time” the message wasn’t going to sink in, so we developed a fairly basic wizard style tool which asked a few questions about what stage your research was at, who your funder was, etc, and then pulled out only the most relevant questions from the Checklist to help you meet the pertinent requirements. So instead of seeing 115 questions, you might be presented with only 15 or 20. Much better.We then added functionalities like export and customisation, and some generic guidance to help with some of the more esoteric sections such as file format selection and metadata.
• #20: For those interested in such things, these are the technologies used in v3.0.
• #21: As I mentioned earlier, version 3 launched very recently, and has a number of great new features.The user interface has been tweaked to allow easier (one-click) access to most of the screens, and we’re investigating customised institutional versions with, among others, the University of Oxford.The tool now enables the application of multiple templates, so you can create a single DMP that satisfies your institution, your funder and your publisher at the same time. These templates can be phased more elegantly, so that you can ask (for example) a few questions at the application stage, more during the project’s lifetime, and then add even more detail when you’re close to completion.Users now have the option to make their plans more widely available. Authentication can be managed via the UK Federated Access Shibboleth mechanism, and we have coded the new system to enable easy translation into other languages, and to handle boilerplate text where this is thought to be beneficial.We have also been working behind the scenes to gain more official endorsement from some of the big funding councils, and this is starting to bear fruit.
• #22: So, in addition to the liaison with the funders, we’ve developed relationships with a variety of others. Our closest working relationship has probably been with the UK Data Archive, which is the designated place of long term deposit for the Economic and Social Research Council. Working with UKDA we have developed a data management planning template and guidance for ESRC applicants, and we also point to some UKDA guidance in the generic Checklist. We have also liaised with Wellcome Trust, the Medical Research Council and various other funders to develop dedicated DMP templates for them. Continuing in this vein, we’ve worked with disciplinary specialists and key institutional contacts to develop further DMP templates, and through the JISC Managing Research Data programmes we’ve contributed to a number of projects creating training materials around this area.Last but not least, we’ve shared experiences with a consortium of US universities – including the Universities of California, Virginia, and Illinois, and the Smithsonian Institution – which has helped them to shape their own DMP Tool.
• #23: These tables show the templates we’ve developed or are in the process of developing. I won’t go through them all now, but the slides will be available for later perusal.
• #24: And more templates are being developed all the time. If you’d like to talk about creating one for your institution or organisation, either catch me afterwards or drop me an email.
• #25: So that’s a pretty good high-level summary of what we’ve done in the data management planning area over the past four years or so.We’d like to end with a quick outline of the DCC’s institutional engagement programme, the major job of work that Sarah and I (and about a dozen other colleagues) are currently involved in. From last Autumn until next Spring – UK seasons, so the other way around for colleagues in New Zealand! – the DCC has been funded by the Higher Education Funding Council for England (HEFCE) to support eighteen HEIs in increasing their institutional data management capabilities. We’re working with a range of institutional types and sizes, from research intensive ancient universities, to new universities and small specialist institutions (e.g. art colleges). The way this works is we first of all make contact with someone already interested in this area, often in the Library, and through them we approach a senior academic, usually at Vice Principal level or equivalent, to make the case for working more concertedly in this area. Once an agreement is reached, the institution selects from a ‘menu’ of tools and services, e.g. (next slide)
• #26: Developing a Data Management StrategyDCC services to support aspects of the research and data management lifecycle, as given in Column 3, andThe tools to support different strands of this (some tools are simply utilised out of the box, others we can provide help and training with, and others – such as DMP Online – can be customised and tailored to match individual institutions’ requirements more closely.
• #27: Similarly, DMP Online can also be used in conjunction with other tools that support the data management/curation lifecycle, be these DCC tools or tools from other sources.
• #28: And at an information exchange level, here’s what we can do. Plans can be exported in a variety of formats, for human and/or machine readerships, and…
• #29: … the tool can link to these types of external systems using a variety of standards and protocols. Of course, this list is not exhaustive, and if you see an opportunity for linking DMP Online with other tools we might not have considered, let us know: the API will probably make it possible.
• #30: So in conclusion, we see the data management plan as a multi-purpose instrument – communication and context – and one that can bring together, if not level, the various stakeholder groups in the research data management endeavour.
• #31: Reversing the hydra metaphor somewhat, we hold that research is more than the sum of its parts, and when data management planning acts to facilitate communication for ensuring smooth and accurate interactions, it also serves as a way to bring it all together.
• #32: Thank you.