‘Raising your ReDMan: Approaches to Research Data Management’

This work (Walking Classes, by Karen Blake) is free of known copyright restrictions.

The following report has been written by John Murtagh, KAPTUR Project Officer for University of the Arts London, who attended the ‘Raising your ReDMan: Approaches to Research Data Management’ event at the University of the West of England, Bristol, Wednesday 23 May 2012.

Presentations are available online from http://tinyurl.com/csshzjk

The focus of this JISC Managing Research Data (MRD) programme event was to present four strands in the differing approaches to research data management (RDM) development chosen by institutions;  integrating RDM within the culture and structure of an institution; approaches to institutional and stakeholder engagements; and how these methods might be used, or adapted, by similar institutions.

Research data in the visual arts is still an ambivalent concept and term to us in KAPTUR and to the creative arts community at large. So when attending research data conferences/symposiums I am usually wearing my ‘research data – but in the arts’ hat on which has two effects 1) glazing (ever so slightly) over the sciencey-raw-data-content and 2) listening intently for a gobbet or two on something that might directly inform the visual arts side of MRD implementation. This is my report on an informative conference.

Simon Hodson, JISC MRD Programme Manager, gave a thorough overview of the national context. He reminded us that Research Councils UK, in particular EPSRC, require that the support and facility for data provision and storage is with the institution. Of particular note was publishers of journals who require data to be published in order to avoid cases of academic fraud; surely this is a significant indicator of journals (and research councils) leading the management, re-use and availability of research data? For the visual arts the increased nature of interdisciplinary research will mean this is likely to have an effect too. Although the caveat remains: our Sector is funded less by funding councils and publishes less in peer reviewed journals than science, technology, engineering, and mathematics (STEM) fields of study.

Institutional Engagement programme

In our next presentation Sarah Jones from the Digital Curation Centre (DCC) outlined the Institutional Engagement programme to work ‘intensively’ in support of increasing RDM capacity at 20 Higher Education Institutions across the UK, from ‘ancient to modern’ institutions. The programme is oversubscribed and even has a waiting list so I was delighted to learn that University of the Arts London has now been included in the final 20.

Sarah revealed a number of interesting details from existing engagements such as:

• Out of the 20 HEIs, the Library is leading in over half the institutions and IT services is leading in two.
• 6 out of the 20 HEIs asked to do a customised DMP Online; others have outlined general support, developing a strategy/roadmaps, defining training and data flow.

Sarah’s case study was from the University of East London which she described as an example of ‘small is beautiful’ whereby the hierarchy was such that adoption of an RDM policy was quick. Other interesting details of the help DCC offered was a policy briefing at UEL, a data management workshop as well as help with implementation of the policy.

A second case study from Queen Mary, University of London described the use of the DCC’s CARDIO tool whereby IT engaged a ‘ heatmap’ approach using colour coding to identify where the biggest gaps were in data support. Something that will prove useful in our own investigations at University of the Arts London.

The third case study from the University of Edinburgh consisted of implementing and emdedding their aspirational RDM policy.  I was interested to learn of the awareness events which the DCC helped run which should remind all institutions of the real task ahead which is not simply obtaining a policy.

The DCC Institutional Engagement Programme, I believe, will considerably further the development of RDM infrastructure at UK HEIs not only because of the time and support they are dedicating to it but also because of the intention to ‘pull out the lessons for the wider community’. The beauty of the KAPTUR project is that we already gain a lot from sharing our experiences across our consortium of four HEIs, and whatever comes out of our engagement with DCC will benefit our institutional partners as well (The Glasgow School of Art; Goldsmiths College, University of London; and University for the Creative Arts).

Drivers and Barriers

The breakout session discussed the ‘drivers’ and ‘barriers’ to research data management at institutions. Detailed notes are available online.

My own notes highlighted several exchanges such as the need to ‘sell’ the concept of RDM to the researchers themselves. It was emphasised that specific ‘engagement from below’ needs to happen – especially with Principle Investigators (PIs) for funded research. The PIs, it was observed by some, were very important in deciding what HEIs do with RDM. Disciplinary drivers are also important as well as the potential to submit datasets to the forthcoming REF2014 (datasets can be submitted as a research output for the forthcoming REF 2014: “…sets, archives, software, film and other non-print media, web content such as interactive tools…”).

A question was also posed: “Are colleges supporting data storage or the University as a whole?” This seemed relevant to the University of the Arts London which has six distinctive and distinguished Colleges; more generally it was observed ‘departmental culture’ is perhaps an important factor for cultural change and adoption rather than a ‘top-down’ approach. Connected to this was the opinion that researchers are generally unaware of the help and assistance that is available to them which was definitely a driver to change.

A further example was that of research project websites that were now orphaned once funding had ceased. Researchers at one particular university requested that their IT services ‘look after’ or host the website. Who gets to decide if it’s even worthwhile keeping? What curational skills are needed or should be passed onto researchers?

It was observed that ‘data rejection’ occurs at the UK Data Archive because of the quality of the metadata rather than its reuse value. This is where Librarians are of value to the field of research data – ensuring that data is described with high quality descriptors – but do they have it? (RLUK report on re-skilling) The training of researchers in RDM skills is important. Just as Open Access and institutional repositories arguably emanated from the field of Library’s concerns about the high cost of subscriptions to journals perhaps it is also they who will play a key role in RDM? A couple of interesting blogs have recently been written by Steve Hitchcock at the Data Pool project: What can research data repositories learn from open access? Part 1 and Part 2.

And now onto barriers. The conflict (a much nicer phrase to use is ‘at slight odds’) between a university’s research office and library or the lack of engagement between the library, research office and IT Services. It was argued that a Data Management Policy *should* drive collaboration between the three. Triumvirate collaboration across Central Services needs to happen. It was highlighted that User Needs Analysis and functional analysis (i.e. who in the university can do this?) is being used to show what researchers actually want within a University and this can help bridge any territorial competitions, perceived or not.

Approaches to RDM

Next up were a number of presentations on the approaches to institutional and stakeholder engagements from our hosts University of the West of England as well as the University of Lincoln, Northumbria University and University of Sunderland.

Stella Fowler the Project Manager of UWE’s Managing Research Data project outlined her ‘service approach’, focussing on ‘quick wins’ in a Seven Stage Roadmap.

Joss Winn from the Orbital project described looking at a ‘Minimum Viable Product’ for RDM which was developer led: “A Minimum Viable Product has just those features that allow the product to be deployed, and no more.” He described how in a survey of what researchers wanted (9 percent return rate/46 replies) researchers ended up wanting managed storage – to share papers and slides online and to collaborate; a Virtual Research Environment. I was impressed by how far Lincoln had come and as Joss reiterated: development is driving RDM and you can still get to the same destination.

Sue Childs from the DATUM for Health project at Northumbria University described a training programme based approach involving a pilot for PhD students; these students would then have to produce a DMP as part of their research project. DATUM’s emphasis was not on creating things from scratch but tweaking and changing existing systems. They developed ‘how-to’ guides for folders, files & version control as well as developing a shared drive for those files.

Other points:

• Need for a researcher-focused strategy and action plan.
• RDM should not be an ‘admin’ burden – researchers should only have to input their data into a system once.
• Researchers wanted clear practical how-to guides.
• Integrated systems.
• Giving researchers skills to appraise the data that they should keep as it’s impossible to keep data for ten years or more. Curational skills!

The University of the Sunderland’s presentation described the Cerif4Datasets project, in which various metadata schemas were investigated. Three different types of metadata schemas were described: descriptive, structural and administrative. Particularly relevant ones for the KAPTUR project were the Categories for the Description of Works of Art (CDWA), and the VRA Core Categories (which VADS own catalogue schema is based upon). Also described was the Data Dictionary – Technical Metadata for Digital Still Images (ANSI/NISO Z39.87-2006). The DataCite metadata schema has already been a reference point for the technical specification to ensure data could be mapped to their mandatory fields if institutions decide to use DataCite in the future. More information about this can be found here: Minting DOIs for research data in the UK. The Integrated Research Input and Output System (IRIOS 2 project) is also of interest.

Breakout sessions: Approaches to RDM

The second breakout session I attended was hosted by Joss Winn from University of Lincoln to discuss approaches to RDM (detailed notes are available from this session).

Notes from all the breakout sessions are available:

• University of the West of England breakout
• University of Lincoln breakout
• Northumbria University breakout
• University of Sunderland breakout

The question was posed whether ‘Minimum Viable Product’ (MVP) for RDM differed depending upon the type of University, for example former Polytechnics  or post-1992 universities to the so-called ‘red brick’ institutions? The University of Lincoln’s size certainly seemed to help them with their own attempt at MVP. A question was raised about existing problems for repository systems working alongside central systems such as Current Research Information Systems (CRIS) and Human Resources (HR) systems.

Also discussed was how to manage researchers who selectively publish their research data. At an early stage of a research project there was a desire not to exhaust a dataset nor to publish all of it at once. Also discussed was editing data once it has gone ‘live’ as well as licensing and sharing data for third party use, specifically the field of Medicine needing to edit datasets that may need to be corrected or changed.

Overall the discussions and presentations were fruitful in establishing common themes/issues/ideas for encouraging effective RDM at an institutional level. The overall impression I had was that RDM is moving steadily and irrevocably forward.