Tips for institution-wide data management planning

A report on one of the break out sessions on institution-wide approaches to data management held during the JISC Managing Research Data programme meeting, 1-2 December 2011.

Monica Duke | 08 December 2011

On December 1st and 2nd I attended the JISC Managing Research Data launch meeting, together with other DCC colleagues.  We facilitated a workshop session on the first day offering an introduction to DCC tools, a really great summary of which was recorded by Paul Stainthorp.

Simon Hodson, the programme manager asked for volunteers to blog about each of the other breakout sessions.  Angus Whyte has written a blog post about a session on data management planning.  There was a flurry of discussion on Twitter about the 'context' issue that Angus mentions in his post (search for #jiscmrd tag), possibly sparked by an unintentionally controversial-sounding tweet from me as I seemed to be the (un)official twitter reporter during that session.

Here are my observations on a later session on institution-wide approaches to data management.  The attendees in this breakout group represented institutions at different stages of implementation, some of which were starting their planning journey and others who had already done previous work on data management, especially through JISC project funding.  There were also some 'fellow travellers' i.e. participants attending the event with a general interest in data management, without being funded on a project in this stream.

Questions that participants wanted to discuss in the session included:

  • What are the components of the human infrastructure for institutional data planning?
  • What lessons can we learn from those who have been there already?
  • How can you push RDM out to disciplines that do not see themselves as being data-driven?
  • How are projects planning to deliver what they produce?
  • Have there been issues with engagement?

Overall the session shaped itself up as an exchange of experiences between projects, with those who had already travelled the path offering tips to those starting out. The following hints and tips were collected:

1. Involve as many stakeholders as possible from across the institution.  

Ways of identifying stakeholders include thinking about the skills that are needed to deliver a service and looking for support services and influencers. Groups to include encompass officers with responsibility for support in data protection and law, research administration and support office, IT services, the library, particularly subject librarians, directors of research. Don't restrict yourself to senior managers and researchers.  PhD students often have good knowledge of data use. There may already be existing forums within the institution which involve a wider range of people, and there may be existing processes for policy making: find out about and tap into existing structures within the organisation. You may need to plan for skills development and training for some of the user groups e.g. the University of Southampton is looking to develop skills for subject librarians. Sector champions who promote data planning within their research group or service will be able to reach out and engage their colleagues, and can be very effective.

2. Establish who will own the service at the end of the project.

The concerns of different stakeholders need to be understood, and this can lead to identifying who may want to invest in future services.  Overall ownership may be difficult to achieve if different technology stacks have different owners.

3. Survey the users to avoid assumptions.

4. Parallel working groups may be needed so that technical discussions are separated out from service-level discussions (otherwise technical discussions can tend to dominate).

5. Data management services may be offered institutionally on the models of other service offerings; to end-users they will appear as another institutional service. Service costs to consider can be separated out into types: software maintenance, user support, sustaining the community, storage.

6. Need to plan and work out when to have a disciplinary strategy vs a broad one, which can be challenging.  When similarities can be identified across disciplines a broad outlook can pay off, but very specific requirements do exist and must be met.  A separate strategy may be needed.  Discipline-specific requirements can relate to size of data, confidentiality issues, methods used.

7. Tools like DMPonline can be used as a driver to engage researchers and other participants.  If groups identified as participants cannot commit due to lack of effort or other reasons, replacements must be found.  Involvement as early adopters or testers can be a way to obtain buy-in for a future service.

The breakout group also identified some common interests to work on together.  These were (1) gathering evidence, in particular there was a request for information channels to come from the evidence gathering team, and to help link people up and (2) common interest in the use of dspace as a platform.

If you are interested in following the progress of all the projects in this round of JISC MRD funding, an aggregation of the project blogs has been created for this purpose.