From standards and support to awarding ‘data Oscars’, we need to start encouraging researchers to share data, writes Sarah Porter
I recently spoke at a JISC event on research data management in universities, which grappled with a number of different issues including why we need to manage our research data, how we decide what to archive and what to discard, and, of course, how to share it.
Of all the topics, it was the last one – sharing - that proved the sticking point. A cacophony of voices argued for sharing complete datasets for a number of reasons - perhaps most strikingly because public research funds should lead to publicly-owned data. The argument is that it is a public good so the question of ownership should perhaps be irrelevant – although it is always easier to support openness when we are thinking about someone else’s data rather than our own.
There’s no doubt that data sharing has practical benefits for the individual: the practice can provide more exposure thereby improving employment prospects; and on the other hand, access to very large data sets made up of other people’s work, combined with enormous computing power, allows researchers to address research problems that would otherwise be intractable.
At an organisational level, an institution’s research reputation benefits when authors make their work freely available to readers over the internet via an open-access repository or journal. Studies have shown that these papers are more frequently cited than those solely available via subscription based journals. Moreover, if data sets have been reused it can help the university show the impact of its research as part of research assessment ; an open data set can improve the integrity of the research paper it supports.
If there are so many benefits to data sharing, then, surely, the argument should be over? We know from experience that the wider ethical arguments can be outweighed by time and financial pressures. Whether researchers are prepared to share their data inevitably depends on a number of factors including whether data is qualitative or quantitative, what stage of their careers they are at and where the requirement to share has come from. So the challenge of incentivising data sharing will be crucial if we are to drive policy changes nationally to support this thinking. As Kevin Schürer, pro vice chancellor of the University of Leicester, argued convincingly in his keynote speech, we have to stop shaking a stick at researchers and start showing them the carrot.
For me, this means supporting researchers from the start of their careers to help them develop good data management practices as part of their professional toolkit. We can’t expect them to pick it up as they go along, and we can’t always expect them to model their behaviour on that of their supervisors, who may not be as up-to-date with these ways of working. JISC has developed online learning materials called MANTRA to teach researchers these skills, so there’s no reason, as Jeff Haywood of Edinburgh University pointed out, why these materials shouldn’t become compulsory elements in research degrees.
Secondly, it’s important to start small and work up when considering culture change like this across a whole organisation. Key staff like librarians who are experts in how to describe and manage complex resources have a role to play in supporting researchers with the challenges – pragmatically, it is much easier if there is a known expert to contact for help and advice. Librarians recognise their own needs for development in this area when faced with the broad range of types of research data that are being created – from spreadsheets to three-dimensional simulations - and the speed at which research is evolving to include new data types. The Digital Curation Centre is an excellent source of advice here and provides examples of good practice in data management.
Thirdly, we need to start rewarding young researchers for opening up their data sets. One idea from the conference was to run the ‘Data Oscars’ – tongue in cheek perhaps but the establishment does need to stand up for its principles and look beyond the research paper for evidence of quality work going on. We want a thriving, active publishing industry in our sector but one that is underpinned by a spirit of public benefit from public investment – hence openness.
Finally, I do believe that we need to develop standards on this to help those institutions that are beginning the journey. Others will be further along and JISC is working hard to make it clear how research-intensive universities like Edinburgh have approached the problem and gained buy-in at a senior level to support their initiatives. I would also add a caveat – that it doesn’t have to be all data for everyone at all stages. We could look at a staged approach whereby different levels of disclosure were identified for different types of research, depending on the need for confidentiality in the case of health-related data for example.
Above all this, key external drivers for data sharing and reuse are emerging fast. This summer the UK’s House of Commons Science and Technology committee’s report on peer review advised that ‘those who expend time and effort adding value to their data to make it useable by others’ should have this ‘acknowledged as a valuable part of their role.’ Such comments will no doubt form the basis of future requirements from funders and by then, some of today’s post docs will be professors. So if we want to support research excellence in the future, we need to start thinking now about how we can effect a culture change among researchers and start involving data sharing and good data management practice as key components in maintaining excellence in research.
Sarah Porter is head of innovation at JISC
Watch Sarah Porter and Simon Hodson’s live webinar addressing the challenges of managing research data on 12 October 2011 between 12-1pm on www.jisc.ac.uk