Why standardising data policy matters

Author: Graham Smith

Open Data Programme Manager

Open data is key to the progress of research and knowledge - it is the evidence behind published results, and alongside outputs such as code and protocols, plays a key role in ensuring the research can be trusted and reproduced - essential components for open science. With governments, funders and research institutions increasingly strengthening their policies on data sharing, demands on researchers is also growing. Yet despite the support for open science and open data, we are still seeing barriers and hesitations around its use, adoption and development with authors, researchers and the wider community. 

For the past seven plus years Springer Nature has conducted an independent survey of researchers with Figshare and Digital Science to track trends in data sharing and get a deeper understanding from researchers about their motivators and influences. One of the key aspects that researchers tell us year on year is the practical challenge of knowing exactly what is required when sharing their data - evidenced by the fact that while they are supportive of data sharing less than 40% of authors actively make their data available. This is something that with our commitment to open science, we are seeking to better respond to.

How can we support researchers with open data sharing?

One such example has been the recent introduction of our new data policy which has simplified and strengthened requirements around data sharing across our journals and books.

Springer Nature has long been an advocate for open data sharing and use of data availability statements (DAS) - linking publications with their underlying data in original publication. In early 2016 we launched a tiered data policy framework - with different policy levels - that was adopted by many of our peers and the RDA. We are now in a position to move beyond this framework to a new policy that better meets researcher requirements and reinforces openness and transparency. The policy:

• makes it mandatory for all original articles to have a Data Availability Statement (DAS). The DAS should explain how data can be accessed and any restrictions on its use. This applies to both original and reused data, whether or not data can be shared publicly;
• strongly encourages data sharing and for certain community-endorsed data types makes it a requirement to share in specific repositories;
• peer reviewers are entitled to request access to underlying data;
• and while it doesn't require all data to be publicly shared, we still ask that the manuscript explains the availability of the data and any conditions for access. 

Through this evolution of our data policy, we aim to:

• create clearer guidance on data requirements for authors;
• provide greater transparency on data sharing for readers;
• enable consistent compliance checks from editorial support teams;
• align our open science ambitions with those of funders, governments and institutions;
• and improve the efficiency of policy implementation. 

We want to make it as easy as possible for authors to be able to understand and comply with data sharing best practice. In developing a single data policy across our journals portfolio we have had the opportunity to do this by standardising our guidance, offering authors clarity when publishing with us.

We are also moving away from heavily templated data availability statements and providing examples specific to different research areas. We want to encourage authors to make their DAS specific to their papers rather than using more generic statements. The idea is that this will then filter through to the readers who will get greater utility of published research.

What’s next?

The new policy will be rolled out progressively across Springer Nature journals. Over this implementation period certain legacy policies will continue to apply; the relevant journal submission guidance will outline the current policy.

Whilst this policy addresses the need for more accessible, actionable and measurable data policy across the research community. There is more that can be done to support researchers:

• we need to ensure that areas that traditionally may not have considered their outputs as ‘data’ such as Humanities and Social Sciences are supported correctly;
• policies should also consider and cover code created or used by authors. We are currently working on a policy that addresses software and code availability statements;
• and we can all make it easier for authors to comply and lower the barriers around engagement - two such example are the use of embedded data curators providing support and expertise for researchers, or integrated data options as part of an author's submission workflow (Figshare and CodeOcean).

We have many resources now at our fingertips to help drive forward open research methods and open science practices. Springer Nature is committed to collaborating on and advancing solutions to enable this - but as partners and publishers we not only need to focus on policies but also need to refine our efforts around support and integration, so that we can meet researchers needs with clear, easy solutions and processes to support the drive for open science.

More on the policies can be seen on our author information pages, pointing to our updated and centralised research data guidance and providing contacts for our research data help desk which will continue to support editors and authors through the new policy implementation.

Notes

What do we mean by ‘research data’?

When we are talking about ‘research data’ we include things like qualitative data, survey responses, and even text transcripts. We are purposely broad in our definition because it’s important to us that our policy implementation takes into account the needs and expectations of the community, as well as privacy and sensitivity considerations.