“A blueprint for what you’re going to do”

This series of videos was arranged before I joined the Research Data Service team, otherwise I’d no doubt have had plenty to say myself on a range of data-related topics! But the release today of this video – “How making a Data Management Plan can help you” – provides an opportunity to offer a few thoughts and reflections on the purpose and benefits of data management planning (DMP), along with the support that we offer here at Edinburgh.

“Win that funding”

We have started to hear anecdotal tales of projects being denied funding due – in part at least – to inadequate or inappropriate data management plans. While these stories remain relatively rare, the direction of travel is clear: we are moving towards greater expectations, more scrutiny, and ultimately into the risk of incurring sanctions for failure to manage and share data in line with funder policies and community standards: as Niamh Moore puts it, various stakeholders are paying “much more attention to data management”. From the researcher’s point of view this ‘new normal’ is a significant change, requiring a transition that we should not underestimate. The Research Data Service exists to support researchers in normalising research data management (RDM) and embedding it as a core scholarly norm and competency, developing skills and awareness and building broader comfort zones, helping them adjust to these new expectations.

“Put the time in…”

My colleague Robin Rice mentions the various types of data management planning support available to Edinburgh’s research community, citing the online self-directed MANTRA training module, our tailored version of the DCC’s DMPonline tool, and bespoke support from experienced staff. Each of these requires an investment of time. MANTRA requires the researcher to take time to work through it, and took the team a considerable amount of time to produce in order to provide the researcher with a concise and yet wide-ranging grounding in the major constituent strands of RDM.  DMPonline took hundreds and probably thousands of hours of developer time and input from a broad range of stakeholders to reach its current levels of stability and maturity and esteem. This investment has resulted in a tool that makes the process of creating a data management plan much more straightforward for researchers. PhD student Lis is quick to note the direct support that she was able to draw upon from the Research Data Service staff at the University, citing quick response times, fluent communication, and ongoing support as the plan evolves and responds to change. Each of these are examples of spending time to save time, not quite Dusty Springfield’s “taking time to make time”, but not a million miles away.

There is a cost to all of this, of course, and we should be under no illusions that we are fortunate at the University of Edinburgh to be in a position to provide and make use of this level of tailored service, and we are working towards a goal of RDM related costs being stably funded to the greatest degree possible, through a combination of project funding and sustained core budget.

“You may not have thought of everything”

Plans are not set in stone. They can, and indeed should, be kept updated in order to reflect reality, and the Horizon 2020 guidelines state that DMPs should be updated “as the implementation of the project progresses and when significant changes occur”, e.g. new data; changes in consortium policies (e.g. new innovation potential, decision to file for a patent); changes in consortium composition and external factors (such as new consortium members joining or old members leaving).

Essentially, data management planning provides a framework for thinking things through (Niamh uses the term “a series of prompts”, and Lis “a structure”. As Robin says, you won’t necessarily think of everything beforehand – a plan is a living document which will change over time – but the important things is to document and explain the decisions that are taken in order for others (and your future self is among these others!) to understand your work. A good approach that I’ve seen first-hand while reviewing DMPs for the European Commission is to leave place markers to identify deferred decisions, so that these details are not forgotten about (This is also a good reason for using a template – a empty heading means an issue that has not yet been addressed, whereas it’s deceptively easy to read free text DMPs and get the sense that everything is in good shape, only to find on more rigorous inspection that important information is missing, or that some responses are ambiguous.)

“Cutting and pasting”

It has often been said that plans are less important than the process of planning, and I’ve been historically resistant to sharing plans for “benchmarking” which is often just another word for copying. However Robin is right to point out that there are some circumstances where copying and pasting boilerplate text makes sense, for example when referring to standard processes or services, where it makes no sense – and indeed can in some cases be unnecessarily risky – to duplicate effort or reinvent the wheel. That said, I would still generally urge researchers to resist the temptation to do too much benchmarking. By all means use standards and cite norms, but also think things through for yourself (and in conjunction with your colleagues, project partners, support staff and other stakeholders etc) – and take time to communicate with your contemporaries and the future via your data management plan… or record?

“The structure and everything”

Because data management plans are increasingly seen as part of the broader scholarly record, it’s worth concluding with some thoughts on how all of this hangs together. Just as Open Science depends on a variety of Open Things, including publications, data and code, the documentation that enables us to understand it also has multiple strands. Robin talks about the relationship between data management and consent, and as a reviewer it is certainly reassuring to see sample consent agreement forms when assessing data management plans, but other plans and records are also relevant, such as Data Protection Impact Assessments, Software Management Plans and other outputs management processes and products. Ultimately the ideal (and perhaps idealistic) picture is of an interlinked, robust, holistic and transparent record documenting and evidencing all aspects of the research process, explaining rights and supporting re-use, all in the overall service of long-lasting, demonstrably rigorous, highest-quality scholarship.

—

Martin Donnelly
Research Data Support Manager
Library and University Collections
University of Edinburgh