blog

I recently attended the EU Presidency Conference on Open Science in Amsterdam. Refreshingly (and speedily), on the day after the conference, a draft Call for Action was issued. The challenge is for the various working groups being convened by the EC to turn the Call into action.

Open science is often talked about as a way to change what is wrong about how science is currently published, shared and incentivised. We talk about providing open access to articles to enable access to essentially print versions of articles that would previously have been visible behind journal pay-walls, and we talk about providing ‘alternative’ views of how an article is used, so using alt metrics to take us beyond a reliance on journal-level metrics and academic citations. Addressing these two issues should be among the outcomes of open science, but its goals are much more forward looking and ambitious.

Open science aims to accelerate scientific progress to turn what is discovered into benefits for all, underpinned by the assumption that technology now allows researchers to collaborate, connect and share knowledge more easily than ever. Much of the discussion on day 1 of the EU conference was about current publisher and journal open access models. However, journals developed in a time when the routes to sharing findings were very different than today. The internet and the vast digital data storage infrastructure, present the opportunity for researchers to share any research findings that they wish and at speed; open science and pre-print platforms already exist (such as F1000Research and BioRXiV) and technology savvy researchers are now building their own platforms to share aspects of their data and findings, and particularly in research fields focused on public health crises and disease outbreaks (see for example, Virological.org, and Nextflu.org).

Early adopters of open science approaches to early findings and data sharing include bioinformaticians, genomics researchers, and disciplines where new techniques are delivering large data streams (e.g. neuroimaging) and whose researchers are used to sharing their data openly, both within their discipline and beyond. In an open science world, the term open access is somewhat anachronistic.

The other thorny issue is how to incentivize scientists to share findings and collaborate as part of transition to open science. As someone who has worked in the research impact for field for some time, one of the simplest things we could do to incentivise science and scientists, is to start to expand the range of indicators of scientific progress that can be reliably considered in funding applications and tenure interviews. This needs to be accompanied by more research on the validity of science-related metrics (research on research) – but that is a topic for another blog. The reliance on journals and article-level metrics has come about largely because of an absence of alternative robust and systematically available streams of science-related indicators. If it is possible to recognise and reward other important assets, activities and contributions to science, then this opens up a basis for incentivisation – and ensuring data sharing and transparency through open science is part of this.

While an article or narrative interpretation of a research finding remains important, digital technology allows the component parts of an article (e.g. databases and datasets, figures, images, software and code) to be tagged, identifiable and citable (using digital object identifiers (DOIs)). Additionally, the contributions to specific aspects of research can be captured using simple taxonomies of roles, such as that adopted in the CRediT taxonomy, to help move beyond a reliance on authorship position as an indication of contribution. Peer review activity is also something that should be rewarded as a valid contribution to others’ work and science more broadly – but this can only be reliably and consistently rewarded if it is open and identifiable.

The other issue around incentives relates to what is being incentivised. The last decade has seen the rise of an ‘impact agenda’ and while impact is the ultimate goal of science, the indicators that measure societal engagement and impact, are not usually the same as those that indicate research excellence. The recent popularity of alt (ernative) metrics has come about because of the need to show the value of research beyond citation-based metrics; they also provide a valuable lens on the use and reach of academic work beyond academia. As currently framed, altmetrics do not, however, reliably indicate anything about the quality and excellence of the research itself. In an open science world where researchers provide access and usability to a wider range of their research works, (a) we can more comfortably separate indicators of reach, non-academic engagement and use around science, from those that more clearly tell us about scientific excellence, and (b) we won’t need the term ‘alternative’ as there will be a broader portfolio of established indicators (and incentives).

In recognition of the opportunities to do things differently, I hope that the actions contained in the EU Call are framed positively; that in itself should provide the incentive for policy-makers and research leaders to transition to open science. I am also an advocate of doing something rather than talking about the challenges and doing nothing; and incremental steps tend to be more effective than revolution. Given that the main barrier for researchers in adopting new ways of making research and findings available, is a fear of affecting career progression, there needs to be evidence of how working with different, open science approaches can benefit researchers and science. A range of pilot projects to test new approaches, with researchers working alongside policy-makers, funders, institutions and publishers would be timely.

To change culture and reliance on a science indicator (journal-based), that we all (more or less) acknowledge is neither accurate nor fit for purpose, then we have to be willing to try something new. Open science offers the opportunity to change the way researchers do science and share their work to accelerate knowledge generation and impact for science and society, but brings with it the added promise of reforming the basis upon which we ‘measure’ science and the incentive structure upon which it is built.