The meaning of open
David Stuart investigates the concept of open research, and what it means to the scholarly community
The discussion in scholarly publishing has moved on from open access to open research, from opening up access to publications to opening up the whole research process and as wide a variety of research outcomes as possible.
However, as this article explores, the breadth of open research makes it a far more difficult idea to conceptualise than open access, with few concrete objectives. There are many challenges to be overcome and opportunities to be embraced, but more metrics may be needed to nudge us towards an increasingly open environment.
These topics (and more) will be discussed at our event, Challenges in the Scholarly Publishing Cycle 2019: Delivering the Open Research Agenda, which takes place on Wednesday 20 November.
What is open research?
At one level open research is easily understood, it is the opening up of the research process and its outcomes as widely as possible to make research more transparent, efficient and collaborative.
But dig more deeply and the concept becomes increasingly hard to pin down. What are its limits, its objectives, its goals? Similar questions could have been raised by the earlier and related issue of open access, which at first seems such a simple premise but quickly becomes increasingly complex.
Part of the problem is that, like the open of open access before it, the open of open research is a continuum. Publishing a research paper in a 17th century academic journal was undoubtedly more open than the personal letters or private notes that preceded it, but it is nonetheless a far cry from the open access articles of today.
Today’s article in an e-print archive is not only accessible to those with a personal or institutional subscription, but can be freely accessed by anyone in the world with internet access.
But even now, if we place a paper in an e-print archive, we have not reached the limits of openness: has it been published with as open a license as possible? Are we willing to waive our rights so that the work can be built upon for commercial purposes or without attribution? How open is the peer review process? Are the images, diagrams, and tables machine-readable? Has the text been semantically enriched to make it as meaningful as possible? Does placing an article on a server make it discoverable? And if we pay to publish in a high impact gold open access journal, which restricts who can afford to publish, is the journal really open access?
If it is difficult to define the open of open access, how much more of a challenge is it to define the open of open research? The processes and outcomes to be opened are much more diverse and potentially ephemeral. Without a definition, however, open research is both difficult to both promote and enforce.
Grace Baynes (above), VP of research data and new product development at Springer Nature, sees open research as ‘opening up every stage of the research process and sharing information as much as is practicable’. She said: ‘Not only the final published article but also the work that underpins those research findings. The data, software, code, protocols, and methodology, are all really important and open research is about opening those up as much as possible in the process of doing research, rather than at the very end. Open research is also about creating more transparency, in opening up the peer-review process and citizen science.’
Open research can potentially touch on every step of the research process, and in most cases we have barely begun to explore what openness can mean or how far it can be pushed. Open access may have turned out to be a more slippery concept than originally envisioned, but it nonetheless looks far more concrete than open research, a concept that can be too big to get our heads around and difficult to recognise if achieved.
A useful shorthand
The difficulty in creating an exact definition of what open research looks like means that it is tempting to settle for a vague concept to which we aspire, but it is necessary to unpack the concept a bit more, rather than leave it in the distance.
As Baynes put it: ‘Open research isn’t an end in itself, it’s a means to an end. Why we care about open research is really important, and that’s about helping researchers communicate and collaborate, and move faster, so that we can advance both the pace of discovery, our ability to make quality discoveries faster and more accurately, and also to make sure we’re using these incredibly intelligent, valuable committed minds on new questions, and new problems, rather than people repeating work others have already done, and wasting their valuable time and effort.’
This is the key point to open research, it is not valuable of itself but for what it allows: open data is valuable not because the data is open, but because when the data is open its maximum value can be extracted, and the transparency can help with the verification of the results. Similarly, open code is not about box-ticking for the sake of it, but rather an acknowledgement of both the importance of building on others’ work and to help build more robust code (as Linus’s Law puts it, ‘given enough eyeballs, all bugs are shallow’).
Importantly, as a whole, open research enables the engagement of the wider community in science and research. As Baynes pointed out: ‘If you think about the potential for the internet and sensor data, and what that has the potential to unlock in everything from understanding the human body and how to improve human health, to monitoring and protecting the environment, there’s so much that’s possible. But, in order to do that and do it well, it’s going to need real engagement with the public and the wider community to ensure trust and collaboration.’
It is only when we move beyond the box-ticking of individual aspects of open research, and consider why resources and outcomes are being made openly available, that open research will start fulfilling its potential. ‘Open research’ is a useful shorthand for the sort of open research practices that are thought to help to speed the pace of discovery – but it is far from a concrete concept, and must be reified anew depending on the particularities of the research and the changing affordances of the wider technological, scholarly, and cultural environment. What is considered ‘open’ today, may not be considered ‘open’ tomorrow.
Sustainable and agile open research
So what is open research today, and what are the near horizons of tomorrow?
For Baynes the three key developments in recent years have been the growth in open access publications in journals, research data and open data, and a proliferation of tools, both from start-ups and from funders.
These are only the first steps, however, in an increasingly complex open information landscape which poses challenges to everyone working in the scholarly research lifecycle: funders must encourage open research without dictating researchers’ research practice; researchers must balance personal interest and public good with an increasingly wide range of publishing choices and funder requirements; publishers must provide both innovative services that meet researcher and funder needs without risking the value of the current system; and libraries must both help researchers navigate the complex information ecosystem and increasingly help them measure and demonstrate researchers’ contributions to it.
In this rapidly changing environment it is important for organisations to be agile. This, according to Baynes, is what Springer Nature is doing, developing sustainable and agile approaches that encourage open research: ‘We are one of the largest open access publishers, but also one of the most agile. It’s not the set business model, one-size-fits-all approach, it’s very much adapting and understanding what stakeholders want. For example, understanding the barriers, challenges and motivators for researchers to make data more openly available, well described, and fair.’
It is also important to be collaborative, as Baynes noted: ‘There seems to be a bigger shift towards awareness and understanding of the value and impact of open research, and much more collaboration between stakeholder groups (funders, publishers, researchers, libraries and institutions) about how we can do that together.
'We’ve all the same common goal, but at the moment it’s just working out how to get all those groups together.’
Of course, in such a fast moving heterogeneous environment, not everybody will be looking to the good of the system. Kieran Prince (above), international sales manager at OpenAthens, pointed to recent cases of open research finding its way into closed silos, and the increasing need for effective oversight and for funders to keep track of how research is being published: ‘It’s a large industry, so it’s always going to be ripe for the kind of practices where you’re getting open access and open research ending up in subscription models.’
Good practice in open research will not just be driven by the availability of tools and services, or the enforcement of funder policies, but by creating a system that incentivises and rewards increasingly open behaviour and disincentivises unwanted behaviour.
Incentivising open research
All stakeholders in the scholarly research cycle need to understand the impact of their contributions and be appropriately rewarded: funders need to see if their money is being spent in the most effective way, researchers need to get credit for their contributions, and publishers need to see the impact of their works. But in an increasingly varied and fast-paced ecosystem, the task is far more complex than the tracking of journal citations.
As Baynes noted, there is a need for ‘ensuring that researchers properly get credit for their contributions, both in terms of citations to those core elements, in terms of claiming credit for the work that they’ve done at the point that they make it openly available, and in being reflected in things that really matter to researchers, like the ability to progress their careers, and secure grant funding.’
At the same time, as Prince pointed out, publishers need a greater understanding of who is accessing their content: ‘As our customers are moving to hybrid models, where they’re producing a mix of open access and subscribed content, we’ve heard from some customers they’re getting a significant increase in page views, but a drop in revenues.
‘So the publisher being able to identify who is accessing the content, can enable them to market content at specific sets of users. If they can identify a sector where content isn’t necessarily being consumed, they can increase their marketing budget there. Being able to identify exactly who’s accessing, even if it is an open access, is quite powerful.’
This comes with additional challenges in an open environment, however, as Prince noted: ‘We need to make sure we’re not putting up barriers to content, as the whole concept is that it should be freely accessed, freely available to everybody. There are ways we can tie it to the individual without having those barriers, and then we can provide value. Things like blockchain can really help with that, protecting privacy and security of the individual, but also tying all of the services into one person. The move away from access and towards an identity-for-everyone is definitely a trend we think we might see in the next few years.’
The future of open research
That open research is in the collective interest is not really in question. There are areas where it may be argued that individual rights surpass collective interest (such as patient data) or that collective interest is outweighed by the accompanying risks (such as biological warfare), but generally it is acknowledged that open research is a public good.
The challenge, however, is the creation of an environment where it is also in the individual’s best interests. After all, opening up research can be resource intensive, and opens work up to greater criticism and potential misuse.
There is, therefore, a need for greater understanding of how open research is shared and reused, so that it can be rewarded appropriately. The open metrics that need to accompany open research will not be easy to establish, they are necessarily more ephemeral than traditional citations that are still not accepted, and far more varied than the altmetrics that have only recently begun to be considered seriously.
As William Bruce Cameron famously stated, ‘not everything that can be counted counts, and not everything that counts can be counted’. For open research to succeed we need to understand both what can be counted, and what counts. There are unlikely to be any simple solutions to the problem of incentivising open research.
To learn more about open research and how institutions are delivering it to the wider academic community, book a place at Challenges in the Scholarly Publishing Cycle 2019: Delivering the Open Research Agenda, which takes place on Wednesday 20 November, 2019.