Open Science takes people willing to disrupt the system to move it forward

Note: This blog is derived from remarks that I gave earlier this year at the UN Open Science Conference (fast forward to minute 31) and have been lightly edited for length and format. This is the second of four blogs on how we can collectively change scholarly publishing for the benefit of all of us. In part one, I outlined the challenges we need to overcome to affect change.

One of the main reasons why I joined PLOS more than five years ago was its founding principle:  An organization established by practicing scientists and dedicated to transforming science communication. We believe it’s necessary to fundamentally rethink how we share and consume research based on two key principles:

• First, we must continue to move beyond the constraints of the physical format and take advantage of the opportunities provided by the digital world.
• Beyond technology, we must upend archaic systems of funding, reward, and recognition, to honor values like collaboration, inclusion, and transparency.

I’m proud to say that PLOS is one of the earliest pioneers in the Open Science publishing space. When we started 23 years ago, as a non-profit, with the mission of accelerating progress in science and medicine by leading a transformation in research communication, many in our industry thought we wouldn’t survive. As you can tell, we are alive and well, and thriving.

• Over the 20 years since our first journal launch we’ve helped prove the viability of Open Access. Despite the evolving conversation about how best to achieve it, it’s now mainstream.
• We changed the publishing landscape with PLOS ONE, enabling all rigorous, peer-reviewed research to be published and discoverable.
• We helped focus the conversation about peer review on rigor and transparency. More people now understand how impact-seeking should never be at the expense of these notions.
• Via a wide range of solutions, we’ve worked towards Open Science at a scale and in ways that increase the transparency and rigor of the entire research communication process.

More recently, we’ve expanded our portfolio with titles focused on critical challenges outlined in the UN Sustainable Development Goals, and signed the UN SDG Publisher’s Compact. Our intention is to create new and diverse communities of practice, who will, in turn, shape how we address the most pressing health and environmental issues facing our society.

My optimism is also fueled by the range of more radical start-ups entering this space, like Arcadia Science, a new research organization building open science principles into their foundations, including reimagining the publishing process from the ground up.

What PLOS Is Doing

Like Arcadia, we have never been driven by tradition, but by a willingness to question the status quo and an eagerness to explore how we can improve the system. First, we’ve delivered change by embracing progressive improvements to science communication while maintaining its core principles such as transparency, inclusion, and collaboration.

Some of this is about technology. Sadly, the publishing industry has largely digitized existing processes without attempting to imagine alternative solutions. Put simply, most of the industry is still print-first even when there is no print output.

Capturing the potential of digital technologies was fundamental to the vision of PLOS’ founders. And we’ve done that. Fast forward 20 years into the era of big data. Science can now address complex research questions in unprecedented ways. But this requires access to and publication of all supporting data as a norm.

We were the first publisher to introduce a data sharing requirement back in 2014 – and yes, it had its detractorsin the early days. But the policy has been tremendously successful and hasn’t encountered the negative consequences that many predicted. It also demonstrated the value of sharing data and gave researchers another way of demonstrating their impact. The sheer scale of introducing a change like this in a journal the size of PLOS ONE has also helped to normalize the sharing of data. We’ve built on this with other progressive, values-driven solutions.

PLOS journals have always welcomed submissions with associated preprints, but it wasn’t until 2016 that we formalized our support in journal policy and began accepting incoming direct transfers from bioRxiv. In 2018, we expanded that partnership and made it a 2-way street. Authors could now have their manuscripts forwarded to bioRxiv as part of the PLOS submission process, instead of two separate submissions. Since then, we’ve launched similar partnerships with other preprint servers, like medRxiv and earthRxiv. By simplifying the process, we’ve seen consistent growth in preprint posting over time.

Code and Methods Sharing

A second example sought to address one of the key barriers to reproducibility: the poor reporting and sharing of methods. PLOS partnered with protocols.io in 2017 to encourage sharing of reusable step-by-step protocols on the platform. But fewer than 1% of authors used this option. So, we sought to better understand why through interviews with researchers about their needs. Most importantly, we learned that expert feedback on protocols is important, as is wide accessibility. But it was also important to be able to receive credit for methods development, and to be able to version and modify protocols.

And so we went back to the drawing board to develop a new product, peer reviewed Lab Protocols. These combine a dynamic protocol in an easy-to-follow format with a peer-reviewed and published article.

One of the key overarching lessons from our work to date is that we’re often dealing with challenges that are not, at heart, technological in nature.

One concrete example. Our internal research shows that scientists are happy with the options available to them to share their data. Yet most of them don’t share data– the solution here is clearly not simply to build another system. The process of sharing is messy and takes time. Sometimes scientists don’t know how. And we’ve already learned that if there’s no reward, busy scientists aren’t going to invest the time.

Over the past 5 years we’ve focused heavily on how to invite behavioral changes through our publishing process. Our thinking at PLOS has been informed by the classic diffusion theory of innovation, developed by E.M. Rogers in 1962. Essentially, this theory states that adoption of new ideas, behaviors and products doesn’t happen simultaneously. Some individuals are more open to adaptation than others.

We’re using this to inform a portfolio approach, focusing on innovators and early adopters for newer ideas and opportunities. But if change is truly to happen, we also need to be thinking about how to bring along those who are more skeptical of change.

For these groups, offering new features in our journals – a format they’re comfortable with – is likely to be more successful. One fairly recent example here is our experiment with code sharing in PLOS Computational Biology. As Jason Papin, co-editor in chief, put it:

Code is the life-blood of research in our field. In order to truly build on the discoveries and advances in computational biology, we must be able to build on the work of others

When we spoke to researchers in the field, we heard that sharing code signals confidence and integrity on the part of the authors, which in turn supports trust in the community. And that they are ready and willing to open doors for fellow researchersby enabling them to more easily reuse or adapt scripts for their own studies, helping the whole field to advance more quickly.

Our team saw an opportunity to shift behavior by piloting a new journal policy. It would require authors to make public any code directly related to the results of their article upon publication. The proof of the experiment’s success? In 2019 – before the policy was introduced – the rate of code sharing for articles in the journal was 53%. Not too bad, you might think – but looking at articles published after the policy was introduced, that number jumped to 87%. That was enough for us to make it a permanent policy. We’re now focusing on code quality and reproducibility.

Behavioral Change

Another model we’re using at PLOS that influences how we approach behavioral change is one you may be familiar with. It’s a comprehensive strategy developed by Brian Nosek, Executive Director of the Center for Open Science, to help us understand the process of culture change and how we can accelerate it. The model helps us think about what’s necessary at each level. For example, the Center for Open Science has done a lot to tackle the level of infrastructure through development of their Open Science Framework – a free, open-source application to support research workflows.

A wide variety of Open Access and data-sharing mandates from funders are driving a great deal of top-down change at the policy level. We’re making notable headway at PLOS with the next level, “making it easy”, with much of the work I’ve just described.

Other like-minded publishers are pushing on different parts of the problem. eLife recently announced that it will become “a service that reviews preprints”, getting rid of the standard “accept or reject” decision after review. This not only has the potential to begin decoupling peer review and publication, but also to speed up sharing of critical research. And it puts authors in charge of their journey.

Focusing on key decision points in the existing reward system such as this can have meaningful impacts. Over time, by nudging different elements of the publications system, we’re moving closer to a system based on transparency and reproducibility.

Creating easier paths to Open Science practices is only one part of the equation. Next week, I’ll talk about how institutional business models can also affect change at scale.