Recently an Opinion Article ‘Fairness in scientific publishing’ by Philippa Matthews (PM), University of Oxford, passed peer review on F1000Research after publication of the version 2. Both versions were openly peer reviewed by three reviewers, one of whom was Gustav Nilsonne (GN), Karolinska Institute. In this blog, both discuss what the FAIR Principles – Findable, Accessible, Interoperable, and Re-usable – mean for authors, reviewers, readers and publishers.
How do you think the FAIR Principles could help move towards a more open publication process?
PM: One way to answer this is to ask “what do I need when looking for a particular resource?” First it
needs to be properly titled, tagged and labelled so that I can be sure that I can identify it when searching. It must be stored somewhere secure and permanent in a standardized format that I can download and use. It should be supported by links to references, resources and/or other datasets that adhere to the 
same rigorous standards. Supporting documentation must be clear, detailed and transparent so that I understand its origins and the content, allowing me to replicate the process myself, as well as using it appropriately to address my own questions. In order for me to do this, I also need explicit assurances that such re-use is permitted. If I share my own work in a format that meets these requirements, and insist that publishers support me in providing the necessary information, structure and resources to do so, then it is Findable, Accessible, Interoperable, and Re-usable, and I believe the process is genuinely open.
GN: The FAIR principles describe a set of desirable characteristics that allow published data and code to be as useful as possible to the scientific community. This helps both by highlighting that data and code 
are precious scientific products that should be published, and by providing guidance for how to make the most of their value. Key aspects of the FAIR principles are consistent with other similar guidelines. I have been involved in the development of criteria for Badges for Open Practices. Like the FAIR principles, the criteria for the Open Data Badge include that data should be published in a repository with a unique identifier, with sufficient metadata, and with a license permitting reuse. Both sets of principles seek to raise the bar, I think, rather than to formalize less stringent practices. This is important, not least in a time when policies for open data publishing are taking shape at many scientific journals and funding bodies.
Results from Philippa’s survey found that 70% of academic respondents were frustrated with how long it can take to get their work published. Why do you think speed of publication is valued by so many researchers and what can be done to address this frustration?
PM: First and foremost, timely publication is a route for sharing data, resources and results, and thus for driving collective scientific progress – we can all see the benefits of minimizing delay. But advancement of science also comes as an indirect function of timely publication, as a result of supporting the development of individual academics, research groups, or institutions through opportunities for education, careers, funding, collaborations, and interactions with media, industry, or pharma.
Creative ways to reduce delays are emerging. These include models that change the traditional timeline of publication (pre-print and pre-review publication) and improvements to the peer review process – open and collaborative reviewing make reviewers more accountable, and reviewers may feel incentivized by initiatives that offer them credit for their work. Relaxing stringent formatting requirements will also reduce the time demands on authors.
“It is not clear to me that the traditional journal model is still adding value to science today, largely because gains in quality following editorial and peer review are offset by the long delays.” Gustav Nilsonne
GN: As Philippa writes, it is obviously wasteful to slow down dissemination of new findings by the insertion of a publication delay of months or years between manuscript submission and final publication. It is not clear to me that the traditional journal model is still adding value to science today, largely because gains in quality following editorial and peer review are offset by the long delays. Traditional publishing models need at the very least to be supplemented by preprint publication to reduce these delays.
Publication delays also adversely affect my motivation. Luminous and paragonic colleagues like Santiago Ramón y Cajal, Ragnar Granit, Peter Medawar, and George Klein have all written about creative joy as a main reason for working in science. Publication delays decouple work from output, in a way, and hence require some substitution of discipline instead. I think the word frustration was not chosen by accident. In my experience, it helps to use publication models with short delays, for instance by publishing preprints.
The survey also found that less than a quarter of academics were comfortable or confident about contacting editors and publishers before submitting their work. What do you think needs to be done to promote more interactions between academics and publishers?
“It would be a positive development if the current approaches to open review could be used as a model for transparent communication between authors and publishers.” Philippa Matthews
PM: In biomedical sciences, real culture change is needed to underpin better communication, requiring commitment from both academia and publishers. My own personal experience has been that both parties are enthusiastic about sharing ideas and tackling collective problems, but we need to work hard to make sure the right channels of communication are in place. I have seen how dialogue can be advanced through social media, conferences and workshops, written publications, and individual conversation.
It would be a positive development if the current approaches to open review could be used as a model for transparent communication between authors and publishers; any exchange between the author and publisher could easily be made accessible as part of the metadata surrounding an article. This would break down the misapprehension that such conversation is in any way ‘cheating’ the system, and might encourage publishers to include opportunities for questions, discussion and feedback as part of the routine submission process.
GN: Perhaps it is natural to not feel comfortable or confident contacting an editor if they are not personally known to the author, and if the journal does not encourage pre-submission enquiries. I agree with Philippa that openness could help to allay fears about corruption by communication outside a formal system, although as far as I can see it is not yet an established fact that that fear is generally holding authors back.
In his review of version 1 of the article, Gustav asked in if Philippa could provide the underlying survey data, which she did in version 2. Why do you think it is important for researchers to have access to data like these and what are the benefits of open data?
GN: In general, open research data are valuable for three main reasons. They permit other researchers to make new findings; they allow scrutiny of the reported results; and perhaps most importantly, they allow for systematic reviews and meta-analyses with less bias and better potential for modelling effects that may not have been the original targets of investigation.
Philippa has published the survey data for her paper openly in a repository. This is good practice and allows any colleague to investigate the data and draw their own conclusions. The repository ensures that the data will continue to be available and also that they are citable.
“I had been inclined to underestimate the value of my own resource, but in allowing others to access it in a full form, it can be used for maximum future benefit.” Philippa Matthews
PM: It is interesting to reflect on the experience of this particular piece of work; I initially provided data that underpinned my opinion article in the form of slides showcasing all the results, but Gustav encouraged me to submit all the metadata in its raw form. With hindsight, I had been inclined to underestimate the value of my own resource, but in allowing others to access it in a full form, it can be used for maximum future benefit.
In more general terms, open data are a powerful resource in several ways. First, openness is a big step for quality assurance in making researchers accountable and ensuring that analyses can be replicated. Second, power is amplified in the true statistical sense of the word – enlarging datasets for analysis, or validating previous observations by replicating them in a new setting, allows us to draw more robust conclusions with more confidence. Third, a dataset that might have been conceived with one question in mind might be used to address something different – this is a really exciting possibility that might be extremely powerful in driving discovery.
F1000Research operates an open peer review model and all correspondence between authors and reviewers are publishing alongside articles. How did you find having peer review conducted in this way, both as an author and reviewer?
“My impression is that openness instead protects against irrelevant concerns, as the fairness of a reviewer can be judged by anyone when the signed report is published.” Gustav Nilsonne
GN: I have experienced this model from both sides. In these cases, it has seemed to me that the correspondence was more constructive compared to the traditional anonymous peer review model. I have speculated that reviewers build their arguments more carefully when they know that other colleagues can read their reports.
Some people have worried that open models leave researchers at risk of negative treatment if the reviewer knows for instance that they are young, or working at a less prestigious institution. But my impression is that openness instead protects against irrelevant concerns, as the fairness of a reviewer can be judged by anyone when the signed report is published.
PM: From both perspectives, this has been a refreshing and positive experience. I feel that those who review my work are held to public account, making them more likely to be measured, thoughtful and constructive, and that I am likely to be able to make genuine improvements as a consequence. As a reviewer, I am grateful that the time and effort I invest in the process can be shared and accessed, that my feedback is a valued contribution that can genuinely improve quality, and that I can retain a record of this work as part of my academic portfolio.