What makes researchers willing to share their data?

Share this on social media:

In principle, the benefits of sharing data – by which I mean the digital or analogue sources that underpin research findings – are undeniable, writes Verena Weigert, senior technology manager at Jisc.

Making data openly available enables other researchers to validate and to attempt the replication of research results. Just as importantly, completely new work might spring from the same data when it is seen through different eyes. But in practice researchers are often reluctant to make their data available for use by others. They feel that there could still be untapped potential for them to exploit, or they worry that they might be out-competed by a larger, more generously funded research team. More off-putting still, getting data ready for other people to examine could prove to be a time-consuming process that brings little direct benefit, as data sharing is often not rewarded in formal research assessment.

Nonetheless, many researchers do share their data, for a variety of reasons. While a number of studies have assessed data sharing practices, barriers and enablers, until recently there has been little work to explore specifically what motivates those who do it. However, insight into what drives them may help policy makers and research funders to devise strategies that make data sharing a more appealing prospect in the future.

Knowledge Exchange (KE), a partnership between organisations in Denmark (DEFF), Finland (CSC), Germany (DFG), the Netherlands (SRUF) and the UK (Jisc) designed to support the use and development of ICT in higher education and research, has been working on uncovering the motivations of researchers to share their data.

The result is a new report, 'Sowing the seed: Incentives and motivations for sharing research data'. Interviews with researchers or groups of researchers in each KE country reveal what made it worthwhile for them to share data in their own particular case, and highlight the obstacles that had to be overcome to do so.

Of course, the report is based on a small and deliberately non-representative sample. In the work we chose research projects at a range of sizes and maturity, alongside a subject, methodological and geographic split. But, by design, they are all projects with a commitment to data sharing already.

The incentives that motivate researchers to share data fell into four main areas:

  • For some, data sharing is an essential part of the research process, for example when the project involves collaborations between disciplines or institutions, or with commercial enterprises;
  • In some cases, researchers can expect direct career benefits from sharing, such as greater visibility for a body of work that may lead to additional citations or to the development of new collaborations, reciprocal data exchanges that enhance their own work, or peer approval;
  • Sometimes, sharing is simply the way things are done within a research group or a discipline; and
  • In a few cases, funders or publishers require data sharing and (ideally) provide infrastructure and data services that support it.

The Finnish example focused on a small-scale piece of sociological work carried out by an individual researcher preparing a Masters thesis examining how elderly men gather together in cities. His methodology included ethnographic observations, interviews with elderly men and a field diary to help him study behaviour and social organisation. This is an interesting case because there is marked difference of opinion among researchers about whether field diaries like this one are suitable for sharing because of the sometimes personal and sensitive nature of the information they can contain. This researcher saw value in sharing it because he recognised that the field diaries of other, more experienced researchers might well be of value to him, but actually did it only when he won an award and was given support to do so by the awarding body. For an individual researcher with limited resources, sharing became practical only when help with cleaning, anonymisation and preparation of supporting documentation became available.

The case study from the UK explores practices in the chemistry department at the University of Southampton, where there is a long history of data sharing. Here, a multi-disciplinary collaboration with private businesses brings challenges relating to intellectual property, patents and other commercial sensitivities, but interviews reveal that there has been significant formal sharing aided by a variety of systems that were implemented by the university itself, which needs to ensure that research funder requirements are met. The department received funding from the Engineering and Physical Science Research Council (EPSRC), which is unique among UK research councils in handing primary responsibility for promoting and enabling research data management to research organisations themselves. Systems that supported data sharing include the university’s own data repository and a variety of subject specific systems such the National Crystallography Service’s eCrystals repository.

The Knowledge Exchange Report includes some detailed policy recommendations for research funders, learned societies, academic institutions, publishers and data centres to think about in order to encourage and support data sharing.  

What became clear is that funders have a pivotal role to play in developing formal data sharing policies that create equal and fair conditions under which all researchers can share data, without fearing that others might gain an advantage.

Funders, for example, should explore ways to reward researchers when they share quality data, possibly by taking this into account when assessing future funding bids. They could also promote the re-use of data via specific funding streams for secondary analysis. And interestingly, the idea that systems to monitor data reuse might be attractive to researchers found little traction in the KE study.

While metrics on data reuse will be of interest for those charged with developing research data management services, it seems that evidence of reuse such as citation, personal feedback, networking opportunities and new collaborations are what is valued by the researchers in the five case studies.