Because good research needs good data

RDMF11: notes from breakout group 1

DCC Staff | 22 July 2014

Report from RDMF Breakout Group 1: “How can we introduce RDM into researchers’ workflows and save them time/effort?”

Stephen Grace, University of East London (chair)

We began by recounting examples from our own experiences as (by and large) research support people, in order to understand the existing working practices that we are seeking to alter. Physical lab books may (collectively) contain “a life’s work” – which seems archaic to us, but it is hard to change working practices, especially when the researchers see little benefit to themselves and will generally do the bare minimum in order to minimise disruption and inconvenience.  So the question is: How do we incentivise and motivate researchers to use the tools we provide? One approach is to ‘sell’ RDM as a benefit to their own future selves. So the case studies and scenarios we produce need to be targeted at a readership of researchers, not just our RDM colleagues.

Experience suggests that anything requiring significant investment of time on the part of researchers will not succeed: if a tool disrupts their workflow, they won’t use it. A quick win would be to tackle the storage problem first. Lowering barrier to entry is critical, and service offerings have to be based on value. This led into a discussion of a pressing and critical need to stop researchers from using inappropriate storage media. There was a (possibly optimistic) view expressed that once researchers have a dedicated area/space for data storage, they will use it. But didn’t we think this about institutional repositories, years ago? And do researchers recognise their own bad practice for what it is?

Staying with the incentivisation theme, we discussed nascent frameworks for assessing / tracking the value of a dataset for research assessment purposes. Data citation frameworks are emerging via (e.g.) Force 11 and DataVerse, but there is still a long way to go.

We discussed the consultancy approach, whereby researchers seek assistance from dedicated data experts, such as the initiative outlined by Oxford’s James Wilson in his morning presentation, e.g. building capacity around previously received enquiries. This helps build understanding from both directions. Barriers to broader adoption of this model could include an aversion to centralisation and  “institutional” terminology - a one size fits all approach will not be well received by researchers – and a lack of solid costing and resource allocation models. There’s no straightforward way to make it scale, and ensure that the expertise is available when “the worst possible combination” of projects gets funded, i.e. those that rely upon the same scarce resource and the same point in time!

This led us to the final theme of the breakout session, namely costs and budgets. One delegate reported that only 1% of their institutional IT budget is dedicated to supporting research, despite it making up 40% of the university’s business. When the disconnect is expressed in these terms to senior managers, they tend to sit up and take notice, so we need to keep banging this drum. Similarly, costs for data management need to be written into senior management consultations with BIS – if they all say the same thing, perhaps the message will get across! And finally, it was pointed out that there is currently an appetite in government to pay for data! This won't last forever, so we need to make hay while the sun shines!