EUA Council for Doctoral Education

Back

Enabling FAIR research data management through doctoral education

12 October 2020 | Lennart Stoy, European University Association

Research data management following the findable, accessible, interoperable and reusable (FAIR) principles has quickly evolved into a major priority of European research policy. Drawing lessons from research in the FAIRsFAIR project, EUA project manager Lennart Stoy reflects on what doctoral education can contribute to this endeavour.

Covid-19 has shown just how crucial international research data sharing is. As part of the ongoing global response, major efforts have been made between groups in Europe and elsewhere to increase the availability of research data, as well as their interoperability and reusability. But these qualities of data do not come out of nowhere, and the example of Covid-19 highlights why we need to change how we treat research data in general. Future researchers need to be acquainted with and trained in research data management, and research stakeholders need to build the required human and technical infrastructure. 

Doctoral education is at the core of this. It is where researchers learn the "tools of the trade". It shapes how academics perceive the scope of their disciplines and how research is carried out, which tools and methods are applied, and which outputs are important. The challenge of creating more interoperable, and therefore more reusable, research data, exemplified by Covid-19 research, must be addressed at this level. In essence, advancing data management skills within doctoral education has important, structuring effects on the research enterprise as a whole. 

However, while our perception of the importance of data management is growing, at least of those working at the policy level, the practical implementation is not moving as swiftly. The research that we conducted at EUA with partners in the FAIRsFAIR project, which aims to "Foster a FAIR Research culture", shows that research data management training comes mostly in the shape of generalist courses. This is a good start as it introduces aspiring researchers to the main concepts and principles of research data management, but it only brings us so far. 

Research data management needs to be normalised in disciplinary and interdisciplinary practice and as part of the usual research process. Only where it is understood as a fundamental element of scientific research, will we avoid the feeling that research data management is an additional tack-on requirement of administrative red tape or a component of overburdening managerial culture within research organisations. Deep integration with disciplinary practices and research ethics and integrity is crucial. After all, RDM aids the reproducibility of science and makes each step in the research lifecycle more traceable and more transparent. Addressing this at the level of doctoral education would be the foundation for a long-term change of research culture. 

One of our previous studies highlighted that universities are keenly aware of this challenge. Even though research data-related competences are being addressed with increasing frequency from the Bachelor to the doctoral level, universities do want to expand data education, in particular for doctoral candidates and other early-career researchers, ranging from data analytics to data management. Fifty-six out of sixty-three universities (89%) that we surveyed in 2019 emphasised that there is a "high need" to strengthen the teaching of data management competencies at the doctoral level.  

Even better, data training should be started earlier, at the Bachelor or Master level with appropriate levels of awareness, skills and competencies. Hands-on experience in data management should be built over time and according to the need of each discipline. This would also support graduates leaving university for work in the private sector where, for instance, data management or quality management skills very akin to research data management are increasingly sought after. Doctoral education and changes in research culture, again, play an important role here, as they can be expected to lead, over time, to an integration of research data management into undergraduate and graduate programmes. This is of course not a short-term shift. 

There is no need to start fully from scratch. The analysis we did in FAIRsFAIR shows, for instance, that some scientific communities, rallying around and facilitated by transnational research infrastructures, are actively advancing research data skills, including data management. Many offer openly available training resources and courses on the use of their tools and platforms. Community initiatives are systematising data science and data stewardship skills and competences. The life sciences are another prominent example where frameworks defining data competencies and skills are already available. Other communities can use and adapt these frameworks and training contents in their contexts.  

As with many other targets in the transition to Open Science, crafting lasting cultural change is a multi-stakeholder effort. Universities need to train doctoral candidates in research data management, creating the minimum framework conditions for "FAIR" disciplinary research cultures to emerge, including changes to rewards and incentives frameworks. Researchers and research communities must in turn be in the driver's seat of developing their own standards and methods in research data management and make use of the expertise that is already out there. External funding should augment institutional investments, especially when funding comes with FAIR strings attached. Policy makers need to understand that the formation of the corresponding research practices can take years. And doctoral education, for all the reasons outlined above, is central to this endeavour.