Abstract
The two Keynotes were introduced by
She posed the question: How can the science system be made more resilient in times of crises and how can science contribute to the resilience of society? These questions are connected and a starting point to think about the future role of science in society.
She said that it was encouraging to see that trust in science still existed during 2020 although there were differences around the globe. The level of trust seemed to be correlated with the level of infections. It seemed lack of trust was not about the fear that scientists were incompetent but the fear that scientists were not acting independently. She said trust in science is connected with scientific publishing: the collective knowledge production system. Publications are the primary medium through which science influences society (e.g. through journalists). Therefore, the published record of science must always be reliable and trustworthy.
Trust does not only depend on access to publications but also on the transparency of the publishing process. It should be an effective system whereby transparency enables quality assurance. However, politicians and the general public should know how science works, transparency alone is not sufficient. Trust depends not only on understanding the system. It also depends on the motives of the people working in the system and the interpersonal relationships. She added: Trust is the expectation that the other will do you no harm.
We should also look at the motives within the scientific publishing system. E.g. the danger of predatory publishers, publishing as an end in itself, that it’s just about mere quantity, profit maximizing behavior, taking advantage of scientists. The question of trust with regard to plagiarism. We should not tolerate this, it will undermine the system of what we consider good scientific practice. She said that the transformation to Open Science is an opportunity to change the discourse about science: to create new knowledge and make this knowledge available for the public.
Kane continued with a general concern that 2020 would be a lost year, but the community refused to consider this a lost year. Many societies and organizations used this period to push things forward. According to Kane, this was somewhat surprising for the publishing industry as it had usually been an industry of conservatism and slow change. But she said that 2020 has shown swift actions from various organizations to ensure that the show would go on and collaboration would continue. In 2020, outputs exploded across STM, and the community also used this as a time to experiment, e.g. with virtual events and innovations in the Peer Review process. Kane added that perhaps the biggest impact was that the pandemic put science in the spotlight. And as a result, there is a better understanding of the critical role that societies, funders and publishers play in supporting this research output.
Kane did not think this time will bring us backwards, there are challenges, but she believed it will be a roadmap to create opportunities and collaborations for a more inclusive research process. Broadening of the research workflow will speed up scientific advancement. Kane concluded that she still felt very optimistic that we will end up in a healthy, stronger and better industry.
The pandemic has shown that the transition can be faster, preprints of COVID 19-research are published three times faster as other preprints. Springer Nature makes relevant research on COVID 19 freely available, and collaborated on this effort with other stakeholders. In addition, many teaching resources were made freely available to help a quick transition to remote teaching and learning. He said this demonstrated the very responsive nature of the publishing sector and that speed showed what can be achieved if we work together in partnerships to move to OS; we need to be open and experiment with researchers and new players in the market. He added that Open Access is the key building block to OS.
Vrancken Peeters is convinced that Gold OA is the preferred way forward (above green OA) even though full transition may be difficult. The alternative green route does take away incentives for gold. Publishers should collaborate with researchers and funders to publish gold OA. Researchers should be able to continue to publish in journals that disseminate content in the best way in known and trusted journals. Humanities Social Sciences’ OA articles more than tripled in transformative agreements. He said that transformative agreements are great examples of partnerships, while there are no blueprints of such agreements and sometimes it is necessary to take risks. You have to build trust with the funders, as well as the research community, first. The key benefit of transformative agreements is that it enables ALL researchers to publish OA irrespective of background. Content that is published OA immediately benefits non academic readers as well.
Publishing is only a small component of total spending but the benefits of OS are huge in comparison to costs spent on publishing. Vrancken Peeters concluded that there are still a lot of challenges to address but that it’s an exciting opportunity for publishers to play a bigger role in the research process and increase the benefits for science.
The panel discussion
MacLeod continued on compliance with the ARRIVE guidelines as there are issues hampering full compliance. Therefore, new guidelines, (NEW ARRIVE), have been implemented last year to make the process easier and simpler. He said that restoring trust in science has traditionally focused on researchers’ integrity and getting rid of plagiarism. But since researchers are very different, restoring trust in science should be system-focused instead of individual-focused. He shared a list in which universities are valued by their level of improvement instead of actual performance. So the focus is on the university’s research improvement strategy. MacLeod believed that way biomedical laboratory research can continue to improve what is considered best scientific practice.
The question:
The subsequent
The panel dove a bit deeper into the alternative models of Peer Review and ways to create more connections between prepublication review and ‘official’ Peer Review. Heesen said that a journal can be run in different models: with traditional Peer Review, with algorithms; in a broader ecosystem. The panel agreed that the community cannot stay static with the Peer Review system and that alternative models should be explored and issues like finding enough reviewers, the sheer amount of submitted papers should be tackled. Scheel added that the current system is overloaded, and with the hierarchy in journals, some papers are reviewed at multiple journals which is inefficient. The panel agreed that a duplication of efforts should be avoided.
The panel discussion
Key takeaways from the white paper: over the last decade there has been a steady increase in output from authors in the Global South (GS), the transition to OA is not happening rapidly in the GS and most of OA publications are gold OA. Researchers are not aware of the existence of OA waivers. Powell added that there are of course differences between the different LMICs, depending on funding mechanisms, national infrastructures and other reasons. Generally, there is still a preference to publish in internationally recognized journals. One of the reasons is that many national journals are not indexed. The waiver policies should be communicated better as they cause confusion.
Recommendations are to create best practice guidelines, to have more inclusive Editorial Boards and peer reviewer networks, and to create more capacity building resources. Powell concluded by stating that it is evident that a business model built around subscriptions is unsustainable in the long run. Diversifying the publishing system needs further dialogue with funders and ongoing commitment of all R4L partners.
Dr. Okomo continued that it is not just an issue of high APCs or limited funding, it is about the entire research process and communication. Reducing the APCs will not be a solution. It should be more about collaboration and instead of just the “Global North’s prestigious journals” providing a high-quality context, journals from the GS should be indexed too. It is about working together and ensuring that all journals provide a high-quality context to cancel such a strong division. This division has given predatory journals the opportunity to aggressively market their services in the GS. Researchers do not have time to check these journals carefully. Dr. Okomo concluded by saying that the current publishing system is a vicious circle: when researchers from the GS aren’t published in international journals, they are consecutively not invited to sit on international Editorial Boards.
The
Powel added that R4L is not just about access but also more and more focused on authorship. With the SDG publishers compact, more awareness has been raised and the publishing community is heading in the good direction.
The session:
She said that many researchers have limited knowledge of OA, or the implications of publishing preprints. They haven’t heard of Plan S for example and do not have a thorough understanding of the different licenses. A Taylor and Francis Researcher Survey 2019 showed that the least favorite license was CC-BY. CC-BY NC ND was preferred.
Kamerlin and colleagues responded to Plan S via an open letter. Researchers should have the freedom to choose a publication venue, and ideally a diverse publication landscape is preferred. Publish-and-Read (PAR) agreements seem helpful for researchers, however, they do create inequity between researchers. Also, between researchers who have access to big publishing budgets and researchers without these budgets. The costs should not become a barrier to publication: Paywalls to read should not become paywalls to publish. She emphasized the role preprints can have in a diversified publishing landscape. Publishing preprints is great for researchers especially when the journal allows preprint updates, however, this is financially challenging for journals: currently there is no sustainable financial model. When looking forward, the need for a diverse publishing landscape with sustainable funding models that respect both researcher and journal needs, is evident.
In his talk
Dr. Pulverer continued on the different steps in the editorial process and why selectivity is important. Because it provides the certification of science, it highlights important scientific discoveries, and inspires across disciplines. EMBO feels it is an obligation of selective journals to optimize efficacy of selective publishing. They try to address serial submissions by inter-journal/inter-manuscript transfers and refereed preprints (this allows authors to post the referee reports with the preprint version). This concept was extended to pre-journal peer review (Review Commons). Reproducibility is a big issue in biomedical research and a lot of money is wasted because research cannot be reproduced.
Because of the selectivity and added value, the costs per article in Biomedical sciences range between 1.000–10.000 Euro. The costs are for editorial processes, peer review, quality control, curation, editing services, publicity, but also for investing in publishing and OS infrastructures. Pulverer said there is a need to find separate payment mechanisms for journals and specific services, e.g. peer review services. Full OA is only possible when there is fair access to journals for authors and readers irrespective of field, geography, seniority and funding support. Certain papers in the journals are not equably supported by OA, e.g. commentary and journalistic section in the journal vs. the research papers and reviews.
He closed by saying that short reviews and commentaries should be included in PAR deals and submission fees should be considered in funder mandates so the industry could move to an OA system that spreads costs fairly.
PLOS launched the Community Action Publishing program and price transparency is an essential part of this program. The actual costs are 50% higher than charged APCs but raising APCs is not an option as PLOS is committed to an equitable ecosystem. PLOS and libraries are partnering with this community-led model to make selectivity work without charging high APCs. The goal is to ensure that costs are balanced across stakeholders and communities. With CAP, the APCs shift to institutional flat fees to distribute the costs of publishing based on the publishing costs of these institutions, this will minimize the barriers to publishing. That way, the costs will be transparent.
Mudditt continued with the question if enough revenue could be generated through this partnership model. Are we able to move away from APCs? Over time, the model will evolve and PLOS can develop a new model for OA. Transparency is a fundamental requirement for this new model to focus on shared value. Mudditt concluded by saying that there is money in the system. The community should focus on how to make selectivity more affordable and ensure the maximum amount of money is spent on research. It is not about maximizing profit, it is about building an equitable sustainable OA system and transparency has to be the cornerstone for rebuilding trust.
Dr. Butcher continued on the functions of Nature journals. The filtering element: about 8% of submitted papers are accepted. Enhancement is important too: Through value added during the peer review process but also the value added by editors during the process. There are lot of different editors and vital support functions. There are not only support functions for specific journals but also more general support functions as legal, HR, Web developers etc. Nature journals also play an important role in amplification. Butcher mentioned the new journal Nature Ageing that will be launched shortly. It does not exist formally yet but online 1st papers have already been accessed 60,000 times because of the work done by journalists.
When looking at the future of publishing OA in Nature Journals: there are 3 routes: Firstly, Transformative agreements, Secondly, Author-chosen OA. The APC for both paths is 9,500 Euro and these costs can be explained by the high-quality standards offered by the number of editors. The third route is the guided OA pilot. The essence is that authors submit to the Nature journal portfolio and work with the editorial team to come up with the best possible publication scenario. The idea is that researchers can trust the Nature Portfolio editors to identify the best home for their paper. When the best possible publishing option is selected and the author takes up the editor’s offer, the APC is 4790 Euro, around 50% of the list price. Butcher added: This does not lower the Nature standards, papers will still be rejected when they are not scientifically sound.
The
Butcher added that for Springer Nature, when developing the guided OA model in collaboration with both researchers and funders, they got feedback on the development process and whether or not they were moving in the right direction. Funders embraced this concept of thinking differently and it was also a good process for the editorial team to think in different ways. Pulverer added that this way of looking from the ‘consumer’s’ point of view is an important new development, without losing sight on quality. The development of Review Commons was also an iterative process of feedback from the community. It is a positive development to have informal discussions and translate these into working groups to make standards for different publishers.
Kamerlin stressed that highly selective journals are not just the big, high impact journals; smaller society journals can be selective too. One of the challenges is that research becomes much more interdisciplinary. Reviewers cannot keep up with developments in all fields and step outside of their field so rigorous Peer Review is important as a trust metric. Pulverer added that publishers have to do a better job in reducing wastage and quality control with regard to reproducibility. Checking tools are essential, it should be a combination of ethics and quality control, peer review altogether. It’s an obligation for selective journals to capture high quality research papers and select on quality. Steps to increase efficacy should be taken. Butcher agreed that common standards are crucial, and Mudditt added that the role of selecting journals is also to look at possibilities to get Peer Review in at the front end, at the stage of preregistration. This is beneficial for researchers as well and it takes the pressure of the current Peer Review system.
In his talk
He continued on the need for protocol transparency. A publication can be a summary of work of many years but does not necessarily cover everything, like early drafts, protocols and experiments that did not work. With preregistration you can capture that early work and get it disseminated while still maintaining credibility and interpretability. A pre-registration is a time-stamped research plan before the study is conducted, so before the data are collected. It is submitted to a public and searchable registry. A registered report is an article type, a proposed study submitted for peer review. All registered reports should include a pre-registration but not all preregistrations will be published as registered reports. Both of them make a clear distinction between planned hypotheses and unplanned discoveries. This helps to address the publication bias against null results.
A 2 stage peer review process is included so feedback from reviewers is received before the study is conducted so protocols can be modified. Stage 1 peer reviewers discuss the statistical analysis. By having that discussion at that stage without looking at final results, they can demonstrate if methods are sound and this should be done in an unbiased way. Then the work is conducted, and authors can submit final results a few years later. Method and protocols should not have changed; new parts: results and discussion. At Peer Review Stage 2, reviewers check if conclusions are justified by the data.
The advantages of the preregistration format: it is reproducible, there is added detail to the protocols, so it is more transparent with Open Data and open materials, and more credible, there is no hindsight or publication bias.
He said that research shows that hypotheses are at least 3 times more likely to be disconfirmed in registered reports compared with regular articles so this means that published results represent a biased representation of our knowledge. There is no evidence that registered reports are cited less, and more and more journals have adopted registered report formats. A database of resources for authors, reviewers and editors is available on the COS website and will help incentivize preregistrations.
One of the current problems with the method section is that there is not enough detail, information is missing. This is not the researchers’ fault as you cannot go into much detail: the method section is about establishing credibility for your research; the paper is for communication purposes and not an instruction guideline. Martone said that it would be good if the methods section could take the form of a recipe for science: the materials you need and the method described as a list of steps. These steps can be shared, reviewed and you can question them. The research objects like data, workflow specification, complete results can be hosted on separate platforms. You could include a PID to link to a specific antibody for example, and add information that you don’t need to include in your paper. A 2013 study has shown that up to 50% of research objects are not identifiable from information provided in the article. Research Resource Identifiers (RRIDs) were introduced in 2014 to identify the resources and track who else has published with this resource. RRID is having an impact, it is possible to aggregate data in the research information network and to pull in citations.
Protocols.io is a community platform where each protocol receives a DOI. It takes a lot less time than writing an entire article, it also manages versions and people can ask questions; in a research paper that’s not possible. Researchers should be encouraged more to use these protocol services.
Prof. Martone concluded by saying that it is good to see that methods sections and reproducibility are receiving more attention. Methods reporting should be made more FAIR and more structured through linking of research objects.
Another problem is the explosion of different sorts of standards being used. The trap of yet another image standard should be avoided. Data management right from the start is key so the research value can be preserved. And Scientific Data Management is to empower people to do meta-analyses (increasing value of previously curated data). The challenge is how to incentivize researchers, just saying it is required or good for the field does not work.
Fraser said that capturing the entire workflow should be part of daily life; it will make researchers’ own experiments easier to do. The focus should be on transparency, context and long-term accessibility. They experienced the benefits of good data management in an experiment on “Synapse” organization and analysis pipeline (SOAP), where they documented right from the start what they were doing. Fraser shared best practices for transparency and reproducibility: these included among others: keep track of how results are produced, avoid manual (meta-)data manipulation, record intermediate results (using standardized formats), connect text/figure to underlying results, provide public access to scripts and results, add DOIs. By this, you follow the FAIR principles to make the process transparent (all the way, raw to polished) and you can make better experiments and the data and tools are reusable and interoperable. Fraser concluded: It provides the carrot as a motivation for good data management and there is no need for a stick.
At the start of the
Martone added: you have to prepare to share and prepare to share fair. It’s a mindset that has to be changed and data management plans have to be followed. Labs should organize data infrastructures as this has not happened on a large scale. This should be an investment. Mellor said there is already growing evidence for citation benefits for papers with datasets, funders could also value plans that use transparency and promote OS practices.
Crotty asked if it would be a negative thing when preregistered reports will include more negative results? Mellor agreed that more negative results will be published with preregistered reports workflow as it is a more effective way for getting null results out there. Martone said that not everything should be forced into the paradigm of research articles. You can open up venues that are more suitable for negative results e.g., in data repositories. With new data science practices there will be a rise of value for negative results. Fraser added that the paper should be a pointer to the database. When data management is part of the workflow it shouldn’t be extra work to publish negative results because you already have the data.
Crotty concluded that it is good to see so much work is actually being done and that the industry can continue to think about ways to better connect the dots in the research workflow, such as pulling value from subsequent works, e.g., grant applications, post-publication, peer review etc.
The OA Switchboard was launched to contribute to the solution as a neutral, independent intermediary, providing shared infrastructure, standards and back office services for funders, institutions and publishers. It is a central information exchange hub, taking charge of multi-lateral OA publication level arrangements. It does not get involved in invoicing, but it enables exchange of information about the OA related publication-level arrangements and ensures a financial settlement can be done. The OA Switchboard is a message hub in the middle, between different stakeholders. It is not building a database: it is a communication hub. The OA Switchboard validates the messages and enables and streamlines communications. This simple technology solution tackles multiple use cases in different steps of different workflows, not just manuscript submissions, also data. It is flexible and it is Open Source. Campfens added that this is essential for building trust and transparency. To deal with the heated topics, the project had to be inclusive from the start, they had to be clear and have clear and common goals, include all stakeholders and regular reporting to keep everybody involved.
For futureproof governance, it was decided to found a new foundation, the Stichting (Foundation) OA Switchboard. This legal structure was chosen because neutrality and independence are preserved.
Campfens concluded by saying that the OA switchboard will continue to build trust through transparency, efficiency and cost-effectiveness. It is about the ecosystem to work better and the researcher to focus on research. It will work best if all stakeholders participate. The years 2021/2022 are regarded as the ‘launch phase’, to achieve wide adoption, and allow time for (technical) integration and implementation, and for continuous improvement.
The reproducibility crisis causes money wastage and delays. To solve this, Sciscore has introduced resource identifiers; using them makes better (reproducible) papers. Sciscore can evaluate methods sections and measure the reproducibility of research in a few seconds. 6 automated screening tools and 55 algorithms check the methods section of an article, and you get a 100% report. The report shows if certain rigor criteria like blinding, randomization of groups, and power analysis are met. Sciscore Version 2 includes features such as code and data information. The Sciscore report includes a rigor table and resource table, and you get a score on a scale of 1–10 that publishers can use.
When the average Sciscore is increasing, it means better reproducible papers can be published. Sciscore not only creates dashboards for journals to check performance, institutions or funders can also have dashboards helping them to take better-informed policy decisions. In conclusion: the reproducibility of science is made measurable, publishers and editors are able to assess the reproducibility of their journals and Sciscore advises editors where to focus their efforts to improve their journal.
Hauer stated there are different challenges in science: it has become more difficult to make groundbreaking discoveries – productivity (the internal rate of return (IRR) on R&D spending) has declined 5 times over the last decade which has led to a $15 billion productivity gap. In addition, there is a reproducibility crisis: money is wasted each year on basic biomedical research that cannot be repeated successfully. Where do these problems come from? Different steps of the production workflow are happening in silo, there is no connection between different cycles and no integration across data silos. Labforward offers software to manage data across the scientific life cycle: Labfolder, Labregister and Laboperator. Different tools and platforms enable horizontal integration to make more out of the data that is generated. Vertical integrations allow the required depths at all stages of the R&D process. The software is developed for scientists, scientific organizations and STM publishers.
Schjøll Brede said we live in a world of exponential growth of (scientific) knowledge. How do we navigate all of this to find the information we need? When you use a citation system, there is a risk of citation bias so it is not a system that can be trusted. You can search with keyword queries but how do you use good keyword queries for interdisciplinary searches for example? When using Artificial Intelligence you need human-machine collaboration and trust this leads to proven improved results.
Isis.ai is a set of tools to semi-automate literature review with AI. It is an exploration tool to find a match with scientific papers across the globe. The tool uses text you inputted and builds a contextual fingerprint and gets an understanding of your problem and then finds results. You need to have human-machine collaboration. It has been proven that the tool works through a series of experiments, large scale review and hackathons with teams to compare performance. The tool reduces time and increases quality, and the tool connects to a variety of content outputs. It is a tool that can be trusted because it allows to machine-human collaboration in an efficient way, better than existing paradigms.
A lot of what we learn and read, will be forgotten very quickly. Therefore, MagmaLearning has developed ARI 9000: a personal AI tutor app powered by machine learning that helps to consolidate what you learn via personalized learning algorithms able to self-improve with experience. ARI will learn from the learner and vice versa in order to further personalize the learning process. It uses Natural Language Processing and generates micro-learning components like puzzles and summaries automatically. It ensures long-term retention with least effort and it’s the best way to consolidate knowledge.
Learning is very personal so to get this learning right, AI and machine learning are used so every person gets the right content at the right moment. Experiments have shown significant benefits for groups using the app. Knowledge visualization and continuous feedback are used so learners can see any knowledge gaps to keep them motivated. The app is used in many sectors, at schools, universities, hospitals, companies and non-profits. The personalized learning experience adapts to diversity and promotes inclusion, for example children with dyslexia or an international company with different backgrounds.
The problem is that there are too many papers and there is not enough time to read them all. In 2018 over 3M articles were published. It is a universal challenge for researchers; they are overwhelmed with the volume of research, trying to skim read to keep up but they might quickly forget what they have read.
Scholarcy offers the solution: articles distilled into summary flashcards. With deep learning technology, Scholarcy does the skim reading so researchers can determine if the paper is relevant for their research.
Scholarcy distills a paper into flashcard, gives a headline summary, a set of significant terms with links to Wikipedia, and bullet points. Translations are covered by Google translate.
Research shows that Scholarcy cuts reading time by over 60% and makes complex information easier to digest. Researchers and students, libraries and research communications departments and content and discovery services are using Scholarcy in different ways, for example to share their own research or to generate lay summaries from complex articles to promote their research output. Publishers and aggregators use the Scholarcy API to create rich XML metadata from documents in any format, and to support the publishing process.
With so much content published each year it is difficult for researchers to stay up to date: they don’t want to miss anything. Publishers on the other hand have the challenge to reach and engage audiences. Researcher-app is designed to solve problems on both sides of the scholarly cycle.
In 2017, less than 10 % visits to journal websites were mobile, so Researcher app was built to solve the problem of discoverability on mobile devices and to make sure researchers will not miss anything.
Publications from different disciplines are indexed and content is pushed to scientists and researchers. Content offering is expanding all the time. Users also get to see additional relevant content e.g. call for papers, newsletters campaigns etcetera. The journey so far: Almost 2m users generate 3,8m content impressions per week.
After the 6 presentations, the audience was asked to vote for the different dotcoms by answering questions such as: which of these dotcoms would you like to collaborate with, in which dotcom would you invest your money, in 3 years who will be the most successful dotcom? At the end Smit declared Schjøll Brede (Iris.ai) as the winner.
Gerlof started the session by saying that this year, De Gruyter celebrates its 10 year OA book anniversary. In collaboration with partners, De Gruyter publishes 3000 OA book titles and 600 OA journals. In 2020, 10% is published gold OA which means that 90% of the book frontlist is not freely accessible so there is still a lot to do to full OA transition. This can be done through a collaborative approach and relevant partnerships. Gerlof added that the OA transition is challenging in HSS because of various reasons such as diversity of research areas, different types of content and HSS researchers’ affinity for print.
In his talk
He said that some alternative non-APC OA models have focused on a collaborative approach such as the consortium approach (SCOAP3 Model), the publisher driven PLOS Community Action Publishing (CAP) Model and the Subscribe to Open Model. The S2O model has been built on the basis of collective action theory to motivate participation, as opposed to models which rely on altruism or pro-group behavior. The model can support publishers to transition their subscription journals sustainably to OA.
He continued on the points underlying the model: 1: Selecting S2O is in an institution’s economic self-interest, it is offered at a discount over price of subscription; 2: It targets current, existing subscribers with expressed demand for content 3: It avoids and does not require collective coordination and leverages economic self-interest; 4: It uses existing procurement processes and maintains existing relationship between publishers and libraries; 5: Openness is guaranteed only with full participation, to reinforce that S2O is a subscription. Control of decision to publish OA remains with the publisher; and 6: Designed to recur annually, stability of the model requires continued participation of sufficient subscribers to meet revenue requirements.
Naim concluded with the S2O community of practice: this platform allows publishers to share experiences with the model and to familiarize the libraries with the model, open to all to see to which extent all offerings align with principals of this models in order to avoid confusion around the model.
He started off with the concept of trust: trust is relevant in organizational change processes and the success of the OA transformation is strongly related to trust. The National Open Access Contact Point in Germany was launched in 2017 as a complement to the DEAL project with the aim of creating the conditions for large-scale OA transformation. The project will end in the summer of 2021. Pieper said that an OA monitor has been created to get better insight into publications and the costs of OA transformation. Complementary business models for OA transformation in HSS and for books have been developed, and workshops have been organized with an emphasis on community building to carry on the OA transformation when the project ends, reports are available on OA2020-DE.
Pieper continued with 3 examples of what the OA Contact point has achieved:
1: Cooperative financing for transforming books to OA, this resulted in the participation of more than 40 libraries and academic institutions and the transformation of more than 180 books to OA.
2: Supporting editors and publishers to implement Subscribe to Open in Germany, this is relevant for smaller publishers, publishing content in the German language.
3: Creating the Enable! Community for OA transformation in HSS. This community is to develop an inclusive, stakeholder-driven OA culture, enables exchange of experience, and promotes fair OA business models for books and journals in HSS.
Trust in the in the large-scale OA transformation is extremely relevant. Librarians need to trust the stakeholders, the financial mechanisms, workflows and outcomes of the OA transition. Pieper concluded with a quote from Luhmann to illustrate that change processes are very complex and that we have to expand from personal trust into system trust.
In his talk:
The JDH will be setting new standards in history publishing based on the principle of multilayered articles: the first layer will enable to produce transmedia narratives (narration layer); the second layer will explore the authors’ reflection on the methodological implications of using digital tools/data (hermeneutic layer); the third layer will give access to data and code through a professional infrastructure (data layer). Fickers said that JDH aims to build trust with historians in order set new standards in history publishing.
The journal team is driven by the idea of co-designing the journal with the publishing world and the world of historians. The core of the editorial process is based on Jupyter Notebooks: this open source research tool is used to build an interface to bridge the different article layers. The source files on Jupyter will be translated into specific elements of the journal article, such as abstract, methods etc. To create trust and optimal adoption, the platform is co-designed with authors. This is done through workshops, ongoing communication and agile development (writing the 1st issue and building the platform are linked processes). Fickers said that this process differs from ‘classical’ submissions to traditional journals; it is true co-development.
Fickers concluded with an example of how an article in the journal will look like. The article format will allow scalable reading; when you refer in the article to data, you can access a direct link to the data and interact with the data in real time. Scalable reading is the future for digital humanities, combining close and distant reading. Readers will be able to interact with the text, the code and data visualizations. The first issue will be online autumn 2021.
Pyne said despite the digital developments the changes are not that radical because evidently, the monograph already serves its purpose very well. The print monograph persists for both practical end sentimental reasons. In 2020 there has been a greater demand for e-books because physical libraries were closed. Pyne stated that even though more can be done to improve the e-book reading experience in this digital age, the focus should be on the transformation to OA to support the widest possible dissemination of HSS in a format valuable for the researchers and audience.
A study has shown that SN OA books have higher usage and a more diverse readership. SN OA books are downloaded 10 times more than non-OA books, and OA books are cited 2,4 more. OA books reach more readers, in more countries. Pyne thought this to be a very compelling argument to have more OA books. But in order to accomplish this, challenges have to be dealt with. A 2019 survey showed that the main reasons for not publishing a book OA are: inability to find funding for OA books, low awareness, concerns of quality. Pyne said this can’t be solved at once and the community should work together to achieve this.
A recent collaborative initiative was the development of the OAPEN OA Books Toolkit. It is a trusted source of information, providing answers to help authors in publishing an OA book. Workshops were organized to understand the community’s needs and an editorial advisory board with a wide range of perspectives was established. The toolkit was written collaboratively with different stakeholders and launched the end of 2020. The toolkit includes a wide range of information with introductions, info on OA licensing, funding etc. It is a living resource, and it is aimed to broaden the toolkit out, with more info and in other languages for example.
Collins continued on the role of journals and publishers: to advance the understanding of climate change. Despite the huge amounts of articles that are being published on climate change, nothing has happened, so it is a public duty to make the climate change science heard. Journals should also encourage co-creation to establish trust in science and breakdown barriers between and across disciplines. By providing additional context around the articles, publishers have the power and responsibility to impact the world.
Last year Elsevier published a report on SDG related research to verify if publishers are accumulating the right knowledge. Climate research is growing rapidly, not only in volume but also in number of citations. So, the knowledge is there and is being cited but needs to be converted into action. Climate research is dominated by high-income countries; this is a concern because climate change is felt most in low and middle-income countries. The research cannot be dictated by the Global North, there should also be more collaboration and publishers can play a role in this too.
Collins said that climate research lacks a social science focus. Climate change is a societal issue and addressing this will require shift in society and behavior so that’s a challenge. We need to see an increase of research in this area. There is low consideration of sex/gender in climate research and there is growing evidence that climate change has a different effect on genders so that is concerning. Collins concluded by emphasizing that journals can have an impact and added that the response to the COVID has shown that the publishing industry can contribute to positive change.
She continued on the barriers limiting policy impact and how to overcome them: Firstly, there is a culture clash between academia and policymakers. This can be overcome by simplifying the message or for example through collaborative authorship between policy makers and academics. Secondly, there is an information overload, this can be addressed by making papers stand out through free access or promotion, or grouping them in thematic areas. Thirdly, it is difficult to measure and assess and therefore value policy impact. The dominance of the impact factor is a disincentive to taking on papers that might have policy importance, but only limited citation prospects. This would require a change in the whole community. Additionally, there is a failure to breakout from Global North dominance of the academic publishing world. At climate policy they have tried to tackle this by providing more support to authors, actively encourage submissions and working with the Editorial Board to increase awareness. These structural problems cannot be addressed by individual journals.
Depledge concluded with her key recommendation: the publishing industry should seriously commit to widening the geographical reach of academic publishing by setting up a mentoring programme to provide training on academic writing and publishing expectations for researchers from the Global South, and consider providing free or low cost academic translation services, translate abstracts & key messages, compile and disseminate strategic special issues on and by under-researched regions.
90% of respondents indicated that their work either currently contributed to tackling real-world problems such as SDGs or that it would be a priority for them in the future. In the Earth and Environmental Sciences, this is real incentive. But why? For researchers, it is important that their research contributes to tackling big real-world problems and they care about impact of their work, such as citations from within their field, contribution to the advancement of science and readership downloads. A comparison between the results in the US, China, UK & Europe and India showed there are changing global priorities and that there are issues hampering researchers achieving their desired impact.
Kelly said that publishers can address this and bridge the gap by communicating the link between highly focused projects and live policy issues, including “lay summaries” and “policy highlights” and fostering interdisciplinarity. Additionally, publishers can build a pipeline to policy through more-robust feedback mechanisms, education for advocacy and for speaking the policy-maker’s language, improved translation and dissemination beyond personal networks and interoperable standards to link research outcomes. Publishers can lean into the role of bridge-builder, within and without academia and foster public engagement. The academic currency could be changed through clarifying policy applications and rewarding and recognizing policy impact. Kelly concluded that in order to fully support researchers’ ambitions to addressing real-world problems, a collective ambition and multiple solutions from all stakeholders are required.
Kolman said that COVID 19 worsened the situation in achieving the goals. Reaching SDG 4 (education) was already lingering but now because of the school closures, the situation has worsened. And because of the lockdown, there has been an increase of domestic abuse and violence to girls and women, which makes achieving SDG 5 (gender equality) more problematic too.
To achieve the SDGS by 2030, raising awareness should be a continuous effort, with a focus on accelerating progress during the “decade of action” (2020–2030). Publishers can be agents of change through publications and accessible content. Publishers should have their own house in order, travel less for example. And in the area of gender equality there is room for improvement too.
The SDG Publishers compact was launched at FBF 2020: The Compact is designed to inspire action among publishers. Launched in collaboration with the International Publishers Association, the Compact aims to accelerate progress to achieve the Sustainable Development Goals (SDGs) by 2030. There are already over 60 signatories and many are from STM. Aldis concluded by saying that publishing is already very much aligned to SDGs and that many publishers are already contributing to SDGs without realizing.
In 2019, In Review was launched: a service integrating preprint posting with journal Peer Review. This provides a way for authors to share their research as a preprint while under review. The preprint is citable and offers real time updates. This provides a possibility to extend transparency in Peer Review, from submission through publication. Peer reviews versions are released in real-time, preprint versions are directly linked to the article on the journal platform. When Open Peer Review is offered, reports are available and connected. This service is now available for 476 Springer Nature journals.
She concluded by saying that preprints are clearly shifting the status quo. Elements of registration, certification, dissemination and archiving that were originally focused on research articles, are now more uncoupled, with preprints allowing the community to explore new ways of Peer Review. With this shift, publishers are also forced to innovate.
Lemberger explained the workflow: After authors have submitted their research to Review Commons, it is sent out for PR, and the full set of review reports is circulated across referees. Review Commons does not make any editorial decisions but after PR, the authors can submit their work to journals of the consortium, up to 4 times. When the work is published in a journal, the full revision process can be posted alongside the article. He said that it turned out that all 17 journals participating in the experiment, have accepted manuscripts with Review Commons reviews. In 95% no new PR was done. Only 30% of authors posted reviews alongside the preprint so more work needs to be done to improve transparency.
He concluded that the audience of review reports now also includes readers, so not just the authors and editors. Clarity and transparency are crucial for the adoption of new PR models. And expert PR can be combined with automated curation methods that for example build knowledge graphs and show emerging topics. He added that human expertise is still needed to meet consistent scientific and editorial standards and that’s why journals will continue to play an important role.
She continued on the changing publication models; when preprints are published, you get access early on but you still miss the review process and then you still have to wait for the end of the process to view the revised, published article. It would be more transparent to have the PR reports published and linked to the publication. With F1000 they try to address that by being completely transparent and allowing versioning. She said there are many different forms of pre-review publication, and hence of post-publication peer review.
Different publisher approaches (Elsevier’s First Look, Springer Nature’s In Review and Wiley’s Under Review) are compared with F1000, on key elements of the different services such as editorial review and commercial model. Are review status, peer review commentary and revision life cycle transparent? The key difference is around post rejection, with F1000 the VOR remains publicly available; flaws are transparent. This is a broader look and takes into account the advances for society. She concluded with the key benefits of post-publication Peer Review: the accountability and credit for reviewers, reproducibility, one connected process, all articles go through Peer Review, and it’s a living publication.
The subsequent
Lemberger added that every domain has to find its own set of rules, there are different cultures of research assessment. Lawrence said some communities already tend to be more collaborative in the 1st place. They are used to sharing and having open discussions, e.g. in medicine there is an obvious need and benefit for society. Swaminathan said that it is already happening organically, and it will help when journals provide infrastructures e.g. around data sharing.
Skipper said that even though it is a positive development that research is available publicly, there could be a risk of confusing the public. Bloom said that educating the public about how science is done, is crucial. Journalists are crucial translators. The community should build a general awareness that scientists build trustworthiness together. Lawrence agreed that educating the public that there is not just one view is important. One way to do this is by publishing peer review reports, there are different views and it shows how research evolves. She added that the community needs to change the way how retractions are perceived, we need to change the research culture. Swaminathan added that openness around preprints could help realize this.
Another question focused on mandating preprint deposition. Bloom said she was never in favor of a journal imposing something before the community was ready. Lemberger agreed; they saw that the community is not ready yet for posting review reports for example. However, he felt that a clear vision for the future is needed but to mandate it now would be too early. Swaminathan added that when thinking ahead, the community should not just keep ‘dominant’ researchers in mind but also early career researchers or researchers from the Global South. Lawrence said the broader scholarly system should be addressed if we want to move to preprints more broadly, in conjunction with the research culture and awarding system. Jackson agreed that tools for culture change are needed, researchers also need to trust their own communities. Skipper said that culture is changing over time and maybe mandating is not necessary: it will just happen. Bloom concluded: posting preprints should not be a goal in itself, the goal should be to advance science.
Skipper said that peer review models are already changing, e.g. with Open Peer Review, she asked if it would be easy to change the culture. Lemberger said it has to be done carefully, it has to be made useful for the readers too and the review process should not be too complicated. Lawrence thought that for research with immediate impact, publishing the review reports is immediately beneficial. E.g. for COVID 19 papers providing context for the media; it brings richness to the discussion. When comparing transparent PR in 2013 with the current situation, researchers have changed, reviewers are much more comfortable. Key is getting peer reviewers credit. Bloom added that the main difference they saw was in the tone of the reports. (Financial) conflicts of interests should also be exposed. Transparency is also key when dealing with inappropriate referee reports for example. Jackson said that the discussion should be open otherwise it’s a loss to science. Skipper concluded the session by emphasizing that if we want trust in science, transparency on how it is done and evaluated and grows over time, is needed.
The
Today’s physicists can never publish all information that has been collected. That would be too much for writers and readers. But the community should do better and collect the ‘trash’ that is currently not included in publications. E.g. Synthesis recipes: synthesis of perfect crystals may not be included because of human reasons: people keeping secrets for competitive advantages, or the metadata is incomplete because of hand-written notes. She said that publishing only success stories leads to bias in research. Failures should also be published. A lot of data is thrown into ‘trash’ because it is not needed for the current research, but it may be useful for something else.
The Novel Materials Discovery (NOMAD) Lab has been created for the community to enable data-sharing. NOMAD is an internationally embedded FAIR data infrastructure, and has collected more than 100 million calculations. A human accessible form for the archive is available.
She said interoperability is the biggest issue and FAIR data infrastructures are a must. The German initiative Nationale ForschungsDatenInfrastruktur (NFDI) is an important step towards this. Changing our publication culture is key: only if we build a comprehensive FAIR data infrastructure, capturing all aspects of data – from synthesis to experiment and theory – and interweaving data, metadata and processing tools with novel concepts of AI, materials science will reach a new level.
She closed with her vision for the library of tomorrow: it will still include books, as well as research journals connected to data. The library will be enhanced by interactive tools for data sharing, metadata will be centralized so the community has access.
There is an annual growth of 21% of articles linked with datasets. Even though this growth is substantial, compared to the total output of research articles, this is still a minimal amount. There are areas where publishers can support data sharing. And the The State of Open Data Report 2020 confirms that publishers still have a key role in FAIR data sharing. Researchers would be more motivated to share data when they get full data citation or co-authorship on papers. Researchers do not get sufficient credit for sharing data.
Milne said a lot has been done already and STM wants to keep momentum going with SHARE-LINK-CITE. Activities to encourage authors are scheduled. Milne added we cannot work alone we must start earlier in the process, before the manuscript is submitted.
The STM 2020 research data year, launched at last year’s APE, started with 6 and ended with 21 participating publishers. The dashboard reports on progress as the initiative to promote data sharing has not finished yet. The whole research cycle and additional topics, such as Data Peer Review, need to be addressed. The sharing, linking and citing of research data should be an integral part of scholarly communication.
The EOSC cannot live without the European e-infrastructure organizations; they belong together. Together they form a web of scientific insight; it is a federation of existing services for storing and interoperable data, making use of the quality mark: data made in Europe. The boundary conditions for EOSC are: The funding comes from EU, it has to be inclusive for all stakeholders, the core follows the subsidiarity principle: it should provide us with a shared purpose. In addition, we have to realize that countries have different structures and accommodate those, it should be self-inclusive, hardware agnostic and the focus is on FAIR data.
The core functions for the EOSC Association 2020+ are to develop and govern a federating core that will manage a compliance framework, trusted certification, the Authentication and Authorization Infrastructure (AAI). Other functions are: PID policies and outreach to stakeholders, monitor services and transaction, manage EOSC trademark(s) and contribute to Horizon EU policy.
The overarching principle for developing EOSC is that research should be in the heart of EOSC, including a multi stakeholder approach and openness. The EOSC association is established to govern EOSC, it is a European co-programmed partnership with 187 member organizations, including research funding, research performing and service providing organizations. All member countries are represented.
The draft MoU between the EOSC association and the EU is a contractual arrangement, not legally binding, the duration is until end 2030, and partners have to commit to openness and transparency and set up and implement an effective reporting and monitoring system. The minimum viable EOSC is a federated set of data used by a core that makes it possible to exchange data. When looking into the future, it is aimed to include not just publicly funded research but to include other public sectors as well, such as education or the private sector.
Sens asked all panelists what they think 2031 would look like. Draxl thought that full interoperability could be reached within materials sciences and with AI we could be a big step further. Luyben thought it would take 20 years before 50% of all relevant research data is as FAIR as possible. Milne emphasized that the behavior of scientists towards data sharing has to change too. It should be kept simple and be harmonized. To motivate scientists to make data fair, tools that will save time have to be provided, this also requires a change of culture. Luyben added that data stewards are important, there should be an optimal mix of expertise available in both libraries and labs. Not one size fits all. In order to avoid the risk of scientists getting confused where to deposit data, it was emphasized that infrastructures should be fully interoperable.
Speakers:
Universities are becoming more sophisticated in their management of the research enterprise, a significant element of their mission and also a major source of revenue that remains strong even during this year’s disruptions. This year, at Ithaka S+R we have been examining the state of the academic research enterprise — how it is managed, the strategic priorities that universities are pursuing for it, and the disruptions caused by the pandemic.
You can also read the two papers from which we presented. Our landscape review of the pandemic’s disruptions to the research enterprise emphasized financial and budgetary impacts, research project impacts, and the human impacts. This project, sponsored by Springer Nature, was principally conducted by Jane Radecki, and Roger C. Schonfeld contributed to it.
Additionally, Oya Y. Rieger and Roger C. Schonfeld examined the role of the senior research officer, a generic title for the vice provost, vice president, or vice chancellor of research. We interviewed 44 of these research leaders from the largest research universities in the US, examining the nature of the role itself, key responsibilities and collaborators, and strategic priorities and challenges. Ex Libris sponsored this project.
Eric Merkel-Sobotta (Berlin Institute for Scholarly Publishing) and Arnoud de Kemp (Founder of APE and Chairman of the Program Committee) closed this year’s APE by thanking the program committee and speakers for realizing this program. They apologized for some of the technical challenges and hoped next year would be in real life again. Merkel-Sobotta added that the Berlin Institute for Scholarly Publishing will start running workshops and seminars this year.
Please note: