Tools for data management planning

12 December, 2011

The DCC ran a workshop at IDCC which brought together the teams that have developed the UK and US data management planning tools: DMPonline and DMPTool. The workshop initially looked at the data policies in each country that define requirements for data management and sharing plans. After coffee we had a demo of each of the tools and discussed future development plans.

It was gratifying to hear how the US work has built on the DCC experience: Sherry Lake of the University of Virginia has taken a selection of questions from the DCC Checklist to help US researchers understand the kind of things funders are looking for, and Martin Donnelly has shared practical lessons with the US tool developers. Despite its relative youth (it was only launched in November 2011), the DMPTool already provides many features which the DCC can learn from, such as the facility to share a plan, the institutional customisations and the page of at-a-glance funder requirements.

Lots of useful questions came out in discussion:

What is a DMP?

Sherry's presentation offered two definitions of a DMP:

  • a brief description of how a researcher will comply with a funder's data sharing policy, reviewed as part of a grant application
  • a comprehensive plan of how to manage data throughout the lifecycle

The DCC sees it as both. DMPonline assists researchers to develop plans based on funder requirements at the grant application stage, but is also intended to help researchers create full plans once funded. There was a general murmur of consent at the meeting that both types of plan are needed, but this premise raised important questions about exporting plans - see What output should the tools provide?

How do first time users know what to write?

The need to connect the tool to a whole set of research support was stressed. There are lots of institutional customisations of the DMPTool allowing more specific guidance notes, contact names and even suggested answers to be provided. The DCC has resisted including example text that can be copied/pasted for fear of the exercise becoming a meaningless, box-ticking exercise but Sarah Shreeves reassured us that the suggested answers are very specific and are only proposed in sections where researchers may be unaware of the details of services provided e.g. storage, backup and long-term preservation.

What output should the tools provide?

David Shotton, Professor of Bioinformatics at the University of Oxford, raised a question about the output of the tools, as funders tend to restrict the length of what can be submitted to a few pages. The US DMPTool only exports the researchers' input with the basic theme headings, whereas DMPonline also includes the funder requirements and associated DCC questions, making for a longer output. He suggested that the DCC Checklist questions could be given as points for consideration, as in the US tool, so the output meets what funders ask for.

Examples outputs for comparison based on a dummy plan made for UKOLN tool training

DMPTool output

DMPonline output

If researchers need to tweak their plan after export to fit in with funders' formatting requirements, they would ideally resynchronise the updated version back to the tool ready for when they create a full plan. Richard Jones of Cottage Labs suggested several options to do this: a template could be provided that allows you to upload reworked content, researchers could edit the plan within the tool instead of dowloading to edit, or export could be avoided altogether by providing funders with a unique identifier to view the plan in the tool (this is how the DMPTool 'share' function works). Nods of approval from Mark Thorley of NERC suggest that the latter is worth working towards.

It was great to share experiences and hear about each others' successes, especially as the ideas for the DMPTool came out of conversations at IDCC in 2010.

  • In the UK we're making increasing headway with funders to gain their endorsement, are working closely with two JISC-funded projects to enhance DMPonline and are close to releasing version 3.0 with improved functionality.
  • The US tool demonstrates just how much can be delivered in such a short time when a community comes together. The tool has been developed entirely on volunteer effort - no funding has been assigned - and the rate of development and uptake is impressive: it was developed in 8 months and attracted 548 users in its first month.

The workshop highlighted how much has been achieved by both groups due to the collaborative effort, and raised lots of areas where we can learn more from one another.

Workshop slides are available, as is Martin and Sherry's paper:

DMPonline and DMPTool: different strategies towards a shared goal