Open access (OA) and open science can sometimes be used interchangeably, yet in reality the former is only one element of the latter. While OA is about making the research article free to access, open science is about making all of the elements behind the research itself openly accessible. Great strides are being made in OA (and Springer Nature itself has been committed to OA for more than 20 years), but we need to move faster towards opening up all elements of research given how fundamental this is to trust, and the reproducibility and progress of science. The ability to share, validate, re-use and discover research, theories and data has an essential role in increasing the credibility of research and speeding up the advancement of knowledge, which can bring benefits and global solutions for all. This was seen up close during COVID-19, where open science led to vaccine development and solutions to contain, tackle and protect us against the virus in record time.
However, to fully transition to an open science environment four key additional building blocks are needed: preprints, open data, open code, and open protocols.
Early sharing
Preprints are an important way to enable early sharing and open discussion of research. As we saw during the COVID-19 pandemic, the ability to share research quickly catalyzed peer collaboration and discussion, which was pivotal in being able to drive knowledge forwards. Some preprint servers, like our In Review, also provide transparency and visibility over the steps that the manuscript is undergoing during peer review and pre-publication, helping researchers and communities gain trust and knowledge about the quality control that research undergoes before its formal publication.
Opening up data
Open Data is central to reproducibility and has an essential role in increasing the credibility of research. Yet a recent study showed that while the majority of OA studies state that their data is “available upon request,” only 6.8% of those authors then actually supply their data when requested.
The 2021 State of Open Data report says that the reasons for this include a lack of understanding, difficulties in knowing how to deposit data, convoluted processes, or challenges around correctly linking data and articles.
Publishers can help address this by, for example, having policies in place to ensure specific details about data sharing are included in every manuscript. However, it is essential that we complement those policies with direct support to authors for sharing their data following FAIR standards as much as possible. Having the right infrastructure and partnerships in place is also important. BMC and Springer Open journals are doing this by automatically depositing the supplementary information into Figshare, while other journals are providing integrated solutions for depositing data through their submission platforms.
Promoting reproducibility
Code is a key research output and it is essential that it is openly shared so that the results are reproducible, enabling other researchers to build on the work. As with open data, more needs to be done to encourage and facilitate the sharing of code, and to ensure code sharing is recognised and valued as a research object in its own right. Whilst the research community is supportive of the benefits of open science, they are not necessarily taking enough actions to share their code. There are many reasons for this such as not having code sharing policies, the right technological capabilities, or support in place to share code. In 2018, we started a trial with Code Ocean, a cloud-based reproducibility platform that enables reviewers and readers to access and run code, and reproduce results. This is now permanent and has been expanded across other Springer Nature journals.
Open methods and research protocols
Another important building block to open science is having easy and open access to protocols and methods. According to a PWC study undertaken for the European Commission, duplicating work due to a lack of awareness of existing research or negative results could cost up to €26 billion in Europe alone. Having access to detailed laboratory methods and protocols, and effectively supporting their use, is therefore critical to ensuring research results can be successfully replicated, and research time and money not wasted. This is why we created an open and free repository for protocols (the Protocol Exchange) more than 10 years ago and are encouraging sharing of protocols associated with our publications. Such action from publishers is also important as it offers additional ways for the primary role of authors in the development of a research study to be recognised.
While there is no “silver bullet” to achieving open science, we believe that these four building blocks are areas that we can all act on now. We will also continue to have these conversations with all stakeholders—policy makers, funders, publishers, librarians, researchers and organisations/institutions—to help drive positive change through collaboration, awareness, and innovation. Coupling real world examples with clear policies around open sharing and open research, better support and credit for researchers, better tools, services and platforms to enable open research practices in a more streamlined way, are all essential factors in accelerating openness and sharing to an established norm.
Dr. Erika Pastrana is editorial director and open science lead at Springer Nature.
