This is an edited transcript of a keynote given at the International Conclave on eScience and Digital Libraries on 28 January 2021 on the topic of Open Science and Scholarly Communication. The slides are available here.
I’ve been fortunate in my career to participate in discussions in numerous countries and at all levels of the research and scholarly communication system, often with a focus on making the transition to open science. I want to outline three key lessons I’ve learned from these experiences, and explore how these relate to the scholarly communications landscape in India – particularly, the emerging Science Technology and Innovation Policy, and the Open Science Framework it proposes.
1. Systems thinking
As Hans Rosling explains in his book, Factfulness, we as humans like simple stories, simple problems and simple solutions. In short, we look for villains and heroes. However, as we well know, reality is not so simple, so instead, we need to look for causes. Most pertinently for me, we need to look for systems.
Scholarly communication is itself a complex system. Complex systems can be found in both nature and society, and are systemic in nature, meaning you cannot take one part of the system and understand it in isolation, without looking at the other parts. We cannot look at scholarly communication in India in isolation from the rest of the world because the rest of the world impacts on scholarly communication in India and what happens in India influences what happens in the rest of the world. Nor can we look at entities such as a journal, a researcher or a discipline in isolation from the rest, because the scholarly communication system is comprised of myriad one-to-one, one-to-many and many-to-many relationships.
Complex in context
Complex systems are also sensitive to context, something we see most significantly in scholarly communication when we look at academic disciplines. What works in one – physics, for example, where for decades, we’ve seen a culture of sharing preprints – does not necessarily work in another, such as history, where even the journal article is of minimal importance, and the monograph is the dominant form of communication. Disciplinary context is therefore incredibly important when we think about scholarly communication.
A further characteristic of complex systems and scholarly communication is path dependency, whereby past events influence what happens in the present and future. I see this most clearly in Western subscription models, where historic pricing based on the print model, i.e. the number of copies of a journal that an institution received when it was in print, continues to inform subscription prices today.
Where do we come from and… where are we going?
A transition happened at the point of the transition from print to electronic resources, but neither publishers nor libraries found it in their interest to move far off the beaten path when it came to pricing. Let us keep this in mind when we look at the transition to Open Access. Whilst on the one hand we might want to rip things up and start again, on the other, libraries and countries are on a particular path which can make it difficult to move to an entirely different one. This is arguably easier to do in India and other parts of the world which are still developing their domestic publishing ecosystems, but we are all locked into historic models of publication to do some degree. Therefore, where we have come from constrains where we can go, and this is often to our detriment.
Complex systems are also ’emergent’, meaning that changes occur but it’s hard to pinpoint exactly why they have materialised. For instance, is COVID-19 the reason we have seen such a rapid growth in the posting of preprints, or did it merely accelerate a trend that was already occurring? As this example shows, we cannot predict and control complex systems, they are episodic, going through periods of relative stability then entering a new phase of development which can involve rapid change. Perhaps COVID-19 will now allowed us to enter a new episode after being ‘stuck’ in the old one for many years? In this, we have neither easy answers, nor heroes or villains… the valiant librarian versus the rapacious commercial publisher? It’s appealing, but just not that simple!
Lifting the lid on scholarly communication
Some aspects of scholarly communication are also more complex than others. Open Access seems pretty complex and challenging to implement, but, simply put, it’s a bit like tinned goods on a supermarket shelf. The articles and research outputs are all in a common format, they might come in slightly different packaging, maybe some are from more prestigious brands or more expensive than others, but actually they have a common form that everyone can recognise and understand. When we ask how do we open these outputs?, we have two main choices: the gold ‘can opener’, or the green ‘can opener’. We have a fixed set of objects – leaving aside monographs and book chapters for the moment – and a finite number of ways to make them open.
Looking at other aspects of scholarly communication and the implications of open science, it’s much more complex. To my mind, open data looks rather like this: a sea of contrasting data in different places; buried on researchers’ laptops, on hard drives or out there in physical form. Sifting through it is an enormous effort, as is stopping people simply adding to the pile with yet more data to sift through. Somethings really are more complex than others!
I’ve worked on several projects around open data, from supporting the UK’s Open Research Data Task Force – a body looking to come up with a direction of travel for the UK, and to implement a pre-existing mandate around making research data open – to developing a research data management roadmap for the University of Oxford. It’s very different from Open Access: there are no publishers you can ask to open up the data, but there are multitudes of repositories, institutions and researchers. It’s a big collective action problem. And so, open data is highly complex and highly costly as I’m sure many of you are aware, which leads me on to my next point…
2. The reality of trade-offs
As Greg McKeown has observed, we can try to avoid the reality of trade-offs, but we can’t escape them. We can try to pursue a goal of equitable access for everyone whilst reducing costs and building a sustainable system of scholarly communication… but wait, that’s actually three goals which are not necessarily compatible. There are finite resources available to solve any given problem, and this requires us to set priorities and make compromises.
This point about trade-offs became particularly apparent to me when we were working for the European Commission, looking at how to create a more competitive and sustainable Open Access publishing market in Europe. It became clear there were multiple goals at play, one being the actively-pursued European goal of immediate Open Access with liberal licencing, the second being a sustainable and competitive market. These two aims remained in tension and without synergy. Making a rapid move to immediate Open Access was going to make it very difficult to either maintain or create a competitive and sustainable market so, consequently, trade-offs are necessary between those goals.
More recently I’ve been working with a group in the Netherlands looking at how to transition to 100% Open Access. The Netherlands has made fantastic progress and already 70-80% of new Dutch research is made Open Access. However, getting to 100% is extremely challenging, costly and difficult. The question is: at what point does a principle simply become unworkable? Again, trade-offs need to come into effect and maybe 90 or 95% Open Access publishing is actually good enough?
Ultimately, my experience has been that change takes time and there are no shortcuts. There’s a combination of factors that need to be in place for cultures to change and for the move to open science to take place, as shown in the graphic below. These things are incremental, they build on each other, and without the infrastructure to make it easy, or without some community norms, a top-down policy or incentive-led approach is guaranteed to fail. These things have to move in sync and it is complex to put all of those factors in place.
3. “Things can be bad and getting better”
Overall, though, I see many reasons to be optimistic about the future of open science. To quote Hans Rosling again, ‘things can be bad and getting better’.
I spent a number of years working on a project for the UK university sector called monitoring the transition to Open Access, in the UK and globally. We looked at five indicators of how the transition was going:
- Were Open Access options available to authors?
- Did authors actually use the options available to them?
- Does being Open Access make any difference to the usage of articles?
- Is Open Access helping to bring down or contain the costs of publishing?
- Is the transition of open access affecting the sustainability of learned societies?
This last one was a particular concern for the UK, as many learned societies use their publishing income to support the health of their disciplinary communities.
We found that, over a few years, four of those five measures were getting consistently better. There were more options to publish Open Access, more options were used by authors, there was more – and in some cases, very substantially more – usage of Open Access articles. Furthermore, there was no evidence of an adverse impact on learned societies. BUT the UK was spending more on making this happen – there was a trade-off involved. Rising expenditure for article publication charges, on top of subscriptions, was the trade-off for increased availability, take up, and usage of Open Access articles. The rise in costs was bad, but overall things were getting better.
Our work enabled an understanding of the extent of the cost increase and enabled UK stakeholders to start taking action to control those costs. You need visibility of progress and spend and then you can use that information – perhaps through collective action and bargaining power – to get better value for money.
An encouraging global landscape
The same picture is apparent at a global level. By one estimate, over 70% of article views will be to Open Access articles by 2025. We are also seeing a very rapidly shifting global landscape, with China and India now disrupting the traditional dominance of global scientific communication by North America and Europe. And, looking at the lowest income, least developed parts of the world, we can also take encouragement from the fact that, while global scholarly output is growing by about 3 – 4% a year, the growth in output from these countries averages over 10% per year, as shown in the graphic below (source: Powell et al, 2020). There’s obviously still an enormous gap in participation between the lowest and highest income countries, but that gap is starting to close and actually is closing at quite a rapid rate. So, you see – once again things are bad, but they are getting better.
India’s Open Science Framework
Turning then to India’s draft Science and Technology Policy, I find that the policy proposes some attractively simple answers – an ‘all-encompassing’ Open Science Framework, the ‘one nation, one subscription’ plan, and an archive that will aggregate all of the outputs from Indian research. These have real value as totemic initiatives that can galvanise the movement towards open science. However, experience tells me that putting these aspirations into practice will present challenging questions, as complex systems are resistant to simple solutions.
I am encouraged to see an acknowledgement of trade-offs within the policy. There is recognition that exceptions must be made to the requirement for data to be made FAIR (Findable, Accessible, Interoperable and Reusable), and the promotion of repository deposits (‘green’ open access) rather than article publication charges (‘gold’ open access) represents a sensible strategy on grounds of cost-effectiveness. Nevertheless, we know that many Indian researchers choose to publish in gold open access journals, and this mismatch between the policy imperative and practice on the ground will need careful monitoring.
Finally, the policy demonstrates awareness of much that remains bad, or undesirable, in the Indian scholarly communication landscape. Predatory journals are widespread, and levels of repository deposit remain low. The quality of Indian publications has not yet increased at the same rate as the quantity, and large swathes of the country continue to face challenges in gaining access to content. There is a de facto reliance on illicit sources of content such as Sci Hub which is the subject of an ongoing court case in India. We must not ignore these negative aspects of the landscape, and the trade-offs that will be required to address them. Many questions remain to be answered in the transition to open science in India. Yet I am confident that this draft STI policy is a sign of a country whose role in global scholarly communication is only going to grow in significance over the coming years.
Like to comment? Let me know your thoughts by emailing me at: email@example.com