Scientific Data is exploring how peer-review mechanisms for sensitive human data can be improved. Here, we outline some of the initial feedback we received from leaders of human data repositories (HDRs), and some innovative alternatives to peer-review.

Data journals, like Scientific Data, stake their reputations on high standards for data peer-review. When it comes to data collected from humans subjects (for example patients, or participants in clinical trials), there can be complex issues surrounding the sharing of these data. These data have to be handled sensitively, ensuring that they are subject to appropriate levels of anonymity or pseudonymisation, and that their distribution does not put the subject at any risk of harm. Safeguards must be put in place to control against attempts at re-identification and other abuses. Access to such data is often limited to verified and trusted users, who may need to sign a Data Use Agreement (DUA) outlining the terms of use, access, and distribution allowed for the data.

These necessary controls raise challenges for anonymous peer-review. How can verifying an individual’s credentials be compatible with an anonymous peer-review process? At Scientific Data, we published a statement two years ago indicating that we would decline submissions where there was no mechanism for anonymous peer-review, and we have had to reject potentially valuable submissions as a result.

While other journals may not actively peer-review sensitive datasets, rising standards for transparent data availability statements are forcing editors to grapple with new questions, such as:

• What repositories are appropriate for these kinds of data?
• What kinds of data use agreements are acceptable?
• Where do our duties to check these elements begin and end?

As a first step in devising solutions for anonymous peer-review of sensitive human data, we consulted with leaders of a selection of HDRs frequently used by our authors. We asked what challenges they face, what peer-review processes they have in place (if any), and for any suggestions of how journals can work with and support HDRs to ensure data quality.

Frequent concerns were raised about the appropriateness of anonymous peer-review mechanisms for sensitive human data, indicating that they do not believe that the benefits outweighed the risk. The most significant barrier to investing in these is that many cannot envisage a mechanism compatible with their strict governance processes (it is “impossible to do it blind”), or one that fits within the resourcing they have available.

“The ability of the editor or journal playing a meaningful role in checking the sensitive data after submission to our repository is constrained by the current policy because access requires a research use statement that is consistent with the Data Use Limitation,” writes Mike Feolo of dbGaP, a repository at the US National Center for Biotechnology Information that hosts sensitive linked genotype and phenotype data.

There were, however, interesting proposals for alternative mechanisms which would require different kinds of journal-repository relationships, such as sharing the results of internal checks made by repository staff, which the referees could scrutinise rather than the datasets themselves. One HDR proposed to establish an external Data Review Board to provide independent perspective and checks on restricted data. To extrapolate from this, we might explore a mechanism by which a pool of verified and pre-approved ‘data reviewers’ at HDRs could be called upon to assess new datasets, expediting the peer review process at journals.

Overall, this exercise highlighted a need to initiate larger and more systematic consultation and to develop creative solutions, which we intend to do in the coming year. As such, it might not be possible to pursue a ‘one-size-fits-all’ approach to streamline this process. We may instead find it necessary to develop several ‘off-the-shelf’ mechanisms compatible with the diverse functionalities of these HDRs.

We invite our readers to consider the draft article that describes the work, its motivation, and relations to other initiatives, and to provide the authors with feedback via this form. The commenting period is open until the end of January 2020. 

Since the journal’s launch in 2014, Scientific Data has provided its authors with a list of recommended repositories, to help them find the most appropriate repository for their data. Scientific Data’s repository list is mirrored for use by all Springer Nature authors, and is also shared openly at figshare for wider use by all. The repositories we recommend are selected based on the maturity, community support, long-term preservation and openness of the resources (learn more about our criteria).

Our view is that clear and meaningful criteria for assessing public repositories are essential, and we therefore support initiatives like this one that aim to identify common principles across publishers. At the same, we would like to assure our authors and repository partners that our support for this initiative does not mean that Scientific Data will stop maintaining its own list, or that we will lower our own standards. At present there are number of ways in which Scientific Data’s criteria are stricter, or at least more specific, than these proposed cross-publisher guidelines. For example, we require that repositories offer licenses that meet common definitions of “open data” and we ask not just that repositories have a preservation plan, but that that plan is plausible and does not rely entirely on short term grant funding.

In our view, meaningful and concrete standards are necessary to build trust, especially if we want researchers to view dataset as a scholarly output that is on par with more traditional research publications.

What do you think? Get involved in the debate by sending your feedback now…

(authors listed alphabetically by surname)

¹ Delft University of Technology, the Netherlands; ² Oxford e-Research Centre, Department of Engineering Science, University of Oxford, UK

Science is living in the era of data – the reuse of other people’s data can drive new research questions and products, and inspire new scientific discoveries. This was the motivation behind the FAIR Principles (published in 2016), which provide researchers with a framework to improve the quality of their research: making their data more Findable, Accessible, Interoperable and Reusable. To turn these aspirational principles into reality, however, we need to provide researchers with FAIR-enabling tools and services that make frictionless the (complex) technical machinery of (meta)data standards and identifiers that underpins FAIR. Service providers, librarians, journal publishers and funders, among others, are actively working to deliver the next generation framework for FAIR data, which is collaborative, interdisciplinary and sustainable. FAIR has also become the (core) mission of a growing number of initiatives – especially in Europe, USA and Australia – encompassing R&D projects and programmes, institutional, national and global service provision, alliances and societies, training and educational efforts. In particular, numerous policy makers from around the world have articulated a vision of global open science and embraced FAIR as the driving principles. In Europe, for example, the vision is being realised through the ambitious European Open Science Cloud (EOSC) programme, across all disciplines.

So where are we now? Is the road to FAIRness understandably or unnecessarily congested? Do we have sufficient bottom-up (discipline-specific) initiatives rooted in real data scenarios, and do these interact with top-down (cross-discipline) efforts? Have these initiatives reached the researchers, whose work FAIR is supposed to enhance? Do we have the necessary technological and social infrastructure? Have we done enough to foster cultural and policy changes to motivate, reward and credit researchers for disseminating and publishing high-quality, machine-readable data? The answer is simple: it is a work in progress. The 2018 State of Open Data Report found that just 15% of researchers were “familiar with FAIR principles”.

What is clear, however, is that there is no lack of coordination initiatives: funded and unfunded, via existing authorities or grass-roots born. In recent years there seems to be a growing number of calls for greater coordination between various research projects and initiatives, all in order to avoid duplication of efforts and wasted resources. Funding schema also seem to favour the coordination approach, sending the message that the fundamental building blocks to enable FAIR data already exist, and that we only need to bring them together. Noble initiatives and who would disagree with the rationale? However, how much effort and resource is now put into coordination, and is it effort well spent? Are these efforts appropriately funded to make the necessary technical and social penetration and have a real impact in terms of depth and breath? In the following sections we provide our opinion on the status quo, which we call the layer cake of FAIR coordination.

Coordination as a challenge

Long-standing, global and cross-disciplinary events, such as those organized by the Research Data Alliance (RDA) and CODATA (which also co-organizes every two years International Data Week), and more recently the Open Science FAIR 2019 conference, are great venues to learn about the wealth of coordinating initiatives. If, like us, you work to make open science and FAIR data a reality, participating in these events is almost essential, because they serve as a huge shop window of initiatives. Often the take home message is the desperate need for harmonisation: a menu to better understand the variety of cakes in the making; the working groups, the position papers, the synchronisation projects, the coordination groups. Interestingly, these projects are often funded by the same funding body. Is there a need to coordinate the coordinators?

Isn’t coordination necessary to avoid duplication of efforts and resource waste? In principle, surely. However, most funded projects already have their specific work plans and deliverables agreed when awarded. Typically, once approved by funders, projects are expected to deliver in accordance with the approved timetable.

So even if ‘coordinators’ identify duplication of efforts between the ‘coordinated’ projects, is it realistic to expect that these funded projects change their agreed schedules or content and format of deliverables? Is it realistic to believe that suddenly the ‘coordinated’ projects (which sometimes are themselves already ‘coordinating’ various other initiatives) voluntarily drop their agreed plans and embark on lengthy discussions and negotiations with their project partners and funders, to justify amendments to their agreed deliverables and to agree on new timelines? What’s their incentive to do this? Coordination is social engineering, and takes time and effort. How realistically can differences in views and approaches be discussed and easily resolved in the life span of a project?

Plus, coordination costs money. That’s not only people’s salaries, but also the costs involved in the administration of these projects, communication efforts, travel to various meetings, and, most importantly, time needed to participate in multiple conflicting telecons, face to face meetings, reviewing and commenting on work packages, deliverables, participating in coordination workshops, etc. All valuable (and expensive!) time which could have been otherwise spent on actual working.

There also seems to be a growing demand to have as many partners as possible involved in big international projects. It is not uncommon, especially in European projects, to see consortia that have received several million Euro funding, which may sound like a lot, but it is not when the funds are split between a large number of partners. Generally, this large participation is necessary to bring different expertise together and to avoid reinventing existing resources. Nevertheless, how much effort (in terms of people, time, money and other resources) must be devoted to mere coordination between all these different partners? We trust that these projects are motivated by a genuine desire to help researchers do their research more effectively. However, there is a risk that between these layers of coordinators who coordinate the coordinators the underlying motivation of these projects might get lost.

Coordination as a necessity

Fundamentally, coordination is beneficial. There are multiple examples where important achievements have been reached through a reasonable two layers of coordination activities. The first layer encompasses efforts that are directly related to practitioners in a given discipline. Exemplars are the Integrated Research Programme on Wind Energy, ELIXIR in the life sciences and AGU in the earth and space science. These initiatives work with researchers and other stakeholders in their domains to make data FAIR. The second coordination layer offers these efforts the opportunity to engage with neighbouring disciplines as well as cross-disciplines and reach out to more stakeholders groups. But in reality, the two layers of coordination are just the ideal scenario, but not always a clear cut. The RDA, for example, is a successful global forum, which is also a network and a working platform where multiple players from various steps in the research life cycle engage to address research data problems. They subsequently deliver solutions that are first reviewed by relevant communities and then endorsed by RDA. Some initiatives from the first layer also participate in RDA to ensure their work can be re-used and aligned with other related activities, in order to avoid unnecessary duplications. For example, both ELIXIR and AGU have groups and representatives in RDA. However, for some, RDA is the first point of coordination where the community itself defines the problem and decides on specific solutions, e.g. for the Wheat Interoperability group. For others, RDA offers the opportunity to enlarge the stakeholder network, and this is the case of a resource like FAIRsharing (Sansone et al. 2019) that evolves and matures driven by community needs.

In all cases, problems and solutions are defined at research community levels, while coordination provides structure and support to catalyse these efforts.

Coordination is also a necessary evil. To be effective, a coordination effort has to be realistic and targeted at the right level. In addition, coordinators need to hold a position of authority in a given area, and be sufficiently resourced to be impactful. Therefore, funders should have a better oversight of what gets funded (refrain from funding several initiatives planning to do the same) and an overview of all funded proposals and their progress. For coordination-focused projects, less is more. Furthermore, the division between research and infrastructure funding programmes should be less drastic to better connect the need of the researchers with the FAIR-enabling tools and services.

A healthier cake with fewer layers

Better oversight on a funder level would have already helped a lot, but would not solve the problem when similar projects are funded by different funding organisations. This requires some bolder changes and introducing more transparency into the whole process of how funding is awarded. Perhaps project proposals should be published as soon as they are awarded and before the project gets going? In that way, anyone could read the proposals and interested communities could have an opportunity to reach out to each other and offer collaboration. Or perhaps, projects could be funded provisionally. After a provisional decision to grant funding is reached, projects would need to be made available for public consultation (public peer-review) on some sort of grant pre-print platform, and the final award only released after comments are adequately addressed.

Alternatively, funded projects could be encouraged to adopt a more agile approach. Detailed three year plans would no longer be necessary – three-year projects could outline high level goals they want to achieve, and then plan and execute in blocks of six or even three months. At the end of each block they would have time to react and respond to the demands for coordination.

Or to go completely in the other direction: maybe funders need to be more directive. Instead of providing a broad theme and inviting open responses, funders need to be clearer and more decisive about the direction they want a project to follow. Dictating not only the ingredients of a project, but also the recipe to be used. Handled badly, directive funding can risk killing innovation, but if managed with expertise, in particular for strategically-important topics, public funding can make transformative changes that are essential for today’s data-driven science.

Lastly, it is essential that coordination efforts do not become echo chambers. In this layer cake of FAIR coordination, the higher you go, the further away you are from the researchers. So something needs to change. Whatever option is taken, it’s clear that solving the current profusion of FAIR coordination projects simply by adding another layer of coordination might not be the best solution. While an extra layer may seem to offer a sweet way of bringing various ingredients together, the result can be a gooey mess.

This work is licensed under a Creative Commons Attribution 4.0 International License.

High-throughput 3D Screening

Organizer: Dr. Kaylene Simpson (Peter Mac)

Scientific Data is inviting submissions releasing and describing data from high-throughput screens employing cutting-edge 3D cell or tissue culture systems. Screens using a wide range of perturbations will be considered, including chemical libraries or functional genomic screens. Priority will be given to submissions that employ high-content imaging techniques, and which have particular value for methods development in this growing area.

Researchers will be asked to share, in full, any imaging data generated in the course of their screen, along with any other relevant measurements, through one or more data repositories approved by the journal. Researchers will also be asked to openly share any custom data processing code used in the work. Interested authors should contact the journal before submission so that we can advise them on suitable data and code repositories.

Submitted articles may be considered for inclusion in a special article collection to be launched in 2020.

To be considered for publication among the first papers in this collection, manuscripts should be submitted to the editorial office by 31^st January 2020; submissions after this date are welcome, and may still be considered for inclusion in this special collection on a case-by-case basis.

Interested in publishing in this special collection?

If you have any questions or would like to make a pre-submission enquiry, please contact the Chief Editor, Andrew Hufton, with appropriate details: scientificdata@nature.com.

Scientific Data is an open-access, peer-reviewed journal for descriptions of scientifically valuable datasets. Our primary article-type, the Data Descriptor, is designed to make your data more discoverable, interpretable and reusable.

Scientific Data’s Editorial Board continues to grow, now numbering over 250 individuals from fields across the natural and social sciences. With submissions to the journal also growing, our Senior Editorial Board will help us maintain a high-level of service to our authors, while furthering the journal’s mission of establishing data publication as a first-class research output and promoting open and FAIR data sharing throughout the wider research community. To this end, our Senior Editorial Board will work with the in-house editorial team to develop consistent journal policies, handle difficult cases in a fair manner, and further expand our existing board membership. They will also help strengthen our connections with community research data experts across multiple disciplines.

The board so-far consists of prominent scientists covering a diverse range of areas including palaeontology, neuroscience, genomics and computational materials science. We are glad to welcome the first additions to our Senior Editorial Board:

• Harry Dowsett, US Geological Survey, USA
• Stephen Eglen, University of Cambridge, UK
• Jingjing Liang, West Virginia University, USA
• Nicola Marzari, École Polytechnique Fédérale de Lausanne, Switzerland
• Russell Poldrack, Stanford University, USA
• Lynn Schriml, University of Maryland School of Medicine, USA
• Kaylene Simpson, Peter MacCallum Cancer Centre, Australia
• Charles Wang, Loma Linda University, USA

Click here to browse the full membership of our Editorial Board

At this time, we are also retiring our Advisory Panel. The diverse members of this panel — representing researchers, funders, librarians and curators — were instrumental in the launch of Scientific Data and we are very grateful for their support during these foundational years. Many of the original members of the Advisory Panel will continue to work with the journal as members of our Editorial Board and we are excited to have their assistance in the continued development of the journal.

Susanna-Assunta Sansone will continue to serve as the ‘Honorary Academic Editor’ of the journal, guiding the strategic development of Scientific Data in partnership with the in-house editorial team, led by Andrew Hufton, Chief Editor of the journal.

Correction: The original version of this post included an out-of-date affiliation for Russell Poldrack. The post as since been updated with the correct affiliation.

Special Article Collection on
Reproducible data processing

In collaboration with Harvard Dataverse

Scientific Data is inviting submissions that provide compelling examples of how portable computing technologies can be used to create transparent, reproducible descriptions of data processing workflows. Submissions considered for this collection should describe valuable research datasets that involve some form of computational processing in their production. Authors should provide source code for all data processing steps in a way that would allow others, including referees, to easily understand and execute all processing steps.

Submitted articles may be considered for inclusion in a special article collection.

Interested researchers are encouraged to first review our criteria for publication for Data Descriptors. Submissions may be from a wide range of research disciplines including fields in both the natural and social sciences.

Authors will have the option to host their data and processing code at Code Ocean. Code Ocean is a computational reproducibility platform that hosts code and data in a cloud-based environment. Users can view, execute and modify the computational workflow in the browser and reproduce the outputs described in the accompanying Data Descriptor without the need to install software locally. Authors may also choose to submit their Data Descriptor as a fully integrated “reproducible manuscript”, using technologies such as Jupyter or knitr.

Researchers will be encouraged to use the Harvard Dataverse to host their data, whenever appropriate. Staff at Harvard Dataverse will assist participating authors in passing code and data files to Code Ocean.

Interested authors should contact the journal before submission so that we can advise them on the best way to host their data, code and manuscript files.

Acceptance for publication will be based on the reuse value of the described dataset, the technical rigour of the procedures used to generate the data, and the transparency of all data processing steps.

Upon publication, data and code must be released to the wider research community under appropriate open data and open source licenses.

To be considered for publication in the potential special collection, manuscripts should be submitted to the editorial office via our online system by 15^th April 2019; submissions after this date are welcome, and may still be considered for inclusion in this special collection on a case-by-case basis.

Interested in publishing in this special collection?
If you have any questions or would like to make a presubmission enquiry, please contact the Chief Editor, Andrew Hufton, with appropriate details: scientificdata@nature.com

Scientific Data is an open-access, peer-reviewed journal for descriptions of scientifically valuable datasets. Our primary article-type, the Data Descriptor, is designed to make your data more discoverable, interpretable and reusable.

Special Article Collection on
Multiomics data

Organizers: Ana Conesa, Sonia Tarazona

Scientific Data is inviting submissions that release and describe datasets from studies that employed multiple ‘omic’ profiling technologies, including, but not limited to, genomics, epigenomics, transcriptomics, proteomics and metabolomics. Submitted articles may be considered for inclusion in a special article collection to be published at the journal.

To be considered, submitted multiomics datasets  should include measurements on different types of biomolecules. For example, manuscripts comparing expression microarrays and RNA-seq of the same samples may not be appropriate for this special collection, while manuscripts that include data from different metabolomics platforms that profile distinct sets of molecules would be welcome.

Submissions may describe data from previously unpublished studies, or may provide more detailed accounts of datasets from studies that were already published. Authors submitting the latter should download our complementary Data Descriptor checklist.

Submitting authors will be asked to deposit each of their omic datasets to an open data repository that is approved by the journal and which is appropriate for the data-type (Browse our recommended data repositories). Researchers will also be encouraged to provide additional integrated data products whenever it may help others reuse their data. Authors are encouraged to write to us before submission for data hosting advice. Any data processing or integration steps should be transparently described. Authors will be encouraged to share supporting source code whenever possible.

If the datasets described were derived from humans and require special protections to ensure ethical use or protect privacy, authors should send a presubmission enquiry to determine whether the dataset would be suitable for consideration at the journal. For more information, please see our recent editorial, and our checklist for sensitive human data.

To be considered for publication among the first papers in this collection, manuscripts should be submitted to the editorial office by 4^th March 2019; submissions after this date are welcome, and may still be considered for inclusion in this special collection on a case-by-case basis.

Interested in publishing in this special collection?
If you have any questions or would like to make a pre-submission enquiry, please contact the Chief Editor, Andrew Hufton, with appropriate details: scientificdata@nature.com.

Scientific Data is an open-access, peer-reviewed journal for descriptions of scientifically valuable datasets. Our primary article-type, the Data Descriptor, is designed to make your data more discoverable, interpretable and reusable.

Complementary Data Descriptor Checklist

Data Descriptors are designed to be complementary to traditional research articles. Researchers can describe and release their data in a more complete manner, and may be able to reduce their reliance on supplementary material, which can be hard to find and poorly accessible during peer-review, and which offers authors little additional credit. About half of the Data Descriptors published at Scientific Data so far are linked to one or more research articles at other journals.

As of July 24th, our publications were linked to 615 publications at 288 journals. In the pie chart above, each wedge is proportional to the number of Data Descriptors linked to each journal’s content. The biggest wedges belong to the highest volume Nature Research journals: Scientific Reports, Nature and Nature Communications. Data and code needed to replicate this is analysis of the journals linked to our publications are available in our GitHub repository.

Related papers are shown in the “Associated Content” box in the top right corner of each Data Descriptor. When the associated paper was published at another Nature Research journal, a link to the Data Descriptor will show on that paper too (see, for example Shan et al, 2018 and Mi et al, 2017). Related papers are also encoded in our ISA-tab metadata and viewable through ISA-explorer (learn more here).

Editors at other Nature Research journals will encourage authors to submit a Data Descriptor when they feel it may help drive wider use and understanding of important datasets. Not surprisingly, many of our publications are linked to publications at other Nature Research journals. But, we are also glad to consider Data Descriptors that build on articles at other publishers, as is evident in the pie chart on the right.

Naturally, authors should ensure that they adhere to basic scientific ethical standards and respect copyright when drafting a complementary Data Descriptor. When you make a submission to any journal, it is always good practice to alert the editor to any related unpublished papers you have in preparation or consideration elsewhere. Any overlap between related publications should always be well-acknowledged, and you should make sure that you have the right to reuse any shared content, especially if the papers will be at different publishers. Our checklist covers these and other important points.

When considering complementary Data Descriptors, our Editorial Board must be convinced that the Data Descriptor will add sufficient value to the scholarly literature to justify the new peer-review process and publications. Interested in submitting your own complementary Data Descriptor? Send us a presubmission enquiry by email (scientificdata@nature.com). We’d be glad to learn about your data and to advise you on whether a Data Descriptor might be suitable.

A growing awareness of the lack of reproducibility has undermined society’s trust and esteem in social sciences. In some cases, well-known results have been fabricated or the underlying data have turned out to have weak technical foundations.

{credit}Quan Hoang Vuong{/credit}

Many researchers have investigated the plausibility of findings in the social sciences and humanities. A typical example is the mysterious Critical Minimum Positivity Ratio 2.9013 by Fredrickson and Losada (2005), which claimed to show that there exists such a positivity ratio and that “an individual’s degree of flourishing could be predicted by that person’s ratio of positive to negative emotions over time”. This ratio had once been a well-known, highly influential and greatly admired psychological “constant” until it was shown by Brown, Sokal and Friedman (2013) to be an unfounded, arbitrary and meaningless number.

To address the plausibility problem, I suggest that a combination of open data, open peer-review and open community dialogue, could serve as a feasible option for the social sciences.

Open data
Ethical standards for open data can be difficult to meet. As I discuss below, my own experience of publishing with Scientific Data (Vuong 2017) showed me firsthand the issues with managing ethical concerns. However, addressing these issues is not impossible, and underneath the difficulty and challenges lies the beauty of shared values and open thinking.

Still, psychological and financial barriers to open data remain. Although it sounds simple, the practice of pursuing an open data policy has turned out to be complicated and hard to achieve, so that “nudging scientific practices toward greater openness requires complementary and coordinated efforts from all stakeholders” (Nosek et al 2015). Unfortunately, the scientific community has not always taken this seriously (Begley and Ellis 2012), forsaking the opportunity for their valuable data to stand the test of time.

Now with the existence of reliable data repositories such as Harvard Dataverse, Open Science Framework, Mendeley, UK Data Archive, depositing data for public use and replication has never been easier. Open publishing platforms such as F1000Research even helps to securely deposit multiple data sets to their own systems without reliance on a third party’s service.

Scientific Data goes even further to the ideal of removing restrictions on use of open data for commercial purposes, by advocating the practice of generating economic value through commercial enterprise, and showcasing the potential benefit for society and the economy (Scientific Data Editorial, “Open for business“).

A benefit to authors of sharing data with a scholarly article is the opportunity for their data to be examined by other scholars, and possibly even lead to further scientific discoveries. The authors have the opportunity of having their colleagues verify both the soundness of data and robustness of the scientific process—as part of a cumulative process of uncertainty reduction (Aarts et al 2015). This is how the scientific workhorse of self-correction operates.

My own experiences of open data
Sometimes the process of data reuse can be quite fast, as is my own experience. When publishing a data article in Data in Brief, with open data sets deposited to Mendeley Data (Vuong and Nguyen, 2016), I was not sure if the data would be reused. To my surprise, a researcher at the University of Pittsburgh used them to make a quality contribution to International Journal for Equity in Health (Nguyen 2017). How quickly did this all happen? My paper was published on 30 September 2016 and Nguyen’s manuscript was submitted on 15 November 2016!

Publishing data that was later reused was a special experience. Work value is transmitted via rigorous peer-review, to data users, i.e., other researchers, for the sake of advancing human knowledge. My own questions about the data’s value were answered not just once, with the editorial and peer-review process, but twice and perhaps more than twice, with subsequent scientific publications that used the contributed data.

My subsequent experience while preparing data for publication at Scientific Data was even more special and unique, in the sense that meeting the most rigorous and demanding criteria boosted the sharpness, clarity, transparency, ethical values and reusability of my data. Even a highly accomplished author can find that the publishing process at Scientific Data enhances their own learning attitude and scientific exploration.

Open review
In the course of encouraging transparency and open science for the sake of better science for all, open review has emerged to be a commendable solution. This may sound strange to researchers in social sciences, where double-blind review has been predominant; however, theoretically the system has its advantages. Evidence from Walsh, Rooney, Appleby and Wilkinson (2000) supports open peer-review since signed reviews are of higher quality, more courteous, and more thoughtful. In addition, social values of open peer-review cannot be ignored as “Improvement of journal policies can help those values become more evident in daily practice and ultimately improve the public trust in science, and science itself,” (Nosek et al 2015).

Heavyweight funders also see the value in the transparent peer-review process, public judgments, and the system’s ability to address the issues of interdisciplinary research review as raised in Bammer (2016). That is why the young platform F1000Research received substantial support not long after its debut (Van Noorden, 2013), from Wellcome Trust in 2016 (Butler 2016) and Bill & Melinda Gates Foundation in 2017 (Butler 2017). As Begley and Ellis (2012) put it, social sciences now need that open review system, which helps facilitate a transparent discovery process leading to significant social benefits.

Open community dialogue.
We refer to an enlarged notion of “dialogue” that consists of technical expert discussions about scientific methods and computer codes (Eglen et al 2017) and the research communication processes that can be made available to the community for evaluation, critique, reuse or extension (Nosek et al 2015).

PubPeer has further incentivized the need for open community dialogue, as flagging a paper on the site is now perceived (incorrectly) by many as a threat. This is a technical expertise community that can help fix problems, deal with statistical weaknesses (that is to improve overall quality of manuscripts), and serve the quality gatekeeping for scientific outlets. PubPeer has been pushing what Eglen et al (2017) advocate, “Share the methods and computer codes.” In social sciences, their values help update and verify, “stylized facts,” in a Bayesian probabilistic world where, “an erroneous argument does not necessarily lead to a wrong conclusion,” due to Gödel’s theorem (Jaynes 2003).

Nowadays, protecting trust in social sciences also means engaging in open community dialogue, which helps to safeguard robustness and improve efficiency of research methods. The combination of open data, open review and open dialogue may sound simple but implementation in the real world will not be straightforward. However, in view of Begley and Ellis’s (2012) statement that, “the scientific process demands the highest standards of quality, ethics and rigour,” they are worth implementing. More importantly, they are feasible to work on and likely will help to restore plausibility to social sciences research. Therefore, I feel it likely that the triplet of open data, open review and open dialogue will gradually emerge to become policy requirements regardless of the research funding source.

References
Aarts AA et al (2015) Estimating the reproducibility of psychological science. Science; 349 (6251): aac4716, doi:10.1126/science.aac4716.

Bammer G (2016) What constitutes appropriate peer review for interdisciplinary research? Palgrave Communications; 2: 16017, doi:10.1057/palcomms.2016.17.

Begley CG and Ellis LM (2012) Drug development: Raise standards for preclinical cancer research. Nature; 483 (7391): 531–533, doi:10.1038/483531a.

Brown NJL, Sokal AD, and Friedman HL (2013) The complex dynamics of wishful thinking: the critical positivity ratio. American Psychologist; 68 (9): 801–813, doi:10.1037/a0032850.

Butler D (2016) Wellcome Trust launches open-access publishing venture. Nature News

Butler D (2017) Gates Foundation announces open-access publishing venture. Nature; 543 (7647): 599, doi:10.1038/nature.2017.21700.

Eglen SJ et al (2017) Toward standard practices for sharing computer code and programs in neuroscience.    Nature Neuroscience; 20 (6): 770-773, doi:10.1038/nn.4550.

Fredrickson BL and Losada MF (2005) Positive affect and the complex dynamics of human flourishing. American Psychologist; 60 (7): 678-686, doi:10.1037/0003-066X.60.7.678.

Jaynes ET (2003) Probability Theory: The Logic of Science. London, UK: Cambridge University Press.

Nguyen HM (2017) Inequality in healthcare costs between residing and non-residing patients: evidence from Vietnam. Int. Journal for Equity in Health; 16 (1): 76, doi:10.1186/s12939-017-0581-3.

Nosek BA et al (2015) Promoting an open research culture. Science; 348 (6242): 1422-1425, doi:10.1126/science.aab2374.

Scientific Data Editorial (2017) Open for business. Scientific Data; 4: 170058, doi:10.1038/sdata.2017.58.

Van Noorden R (2013) Company offers portable peer review. Nature; 494 (7436): 161, doi:10.1038/494161a.

Vuong QH (2017) Survey data on Vietnamese propensity to attend periodic general health examinations. Scientific Data; 4: 170142, doi:10.1038/sdata.2017.142.

Vuong QH and Nguyen TK (2016) Data on Vietnamese patients’ financial burdens and risk of destitution. Data in Brief; 9: 543-548, doi:10.1016/j.dib.2016.09.040.

Walsh E, Rooney M, Appleby L, and Wilkinson G (2000) Open peer review: a randomised controlled trial. British Journal of Psychiatry; 176 (1) 47-51; doi:10.1192/bjp.176.1.47.

Alexander Weiss

Guest post by Alexander Weiss of the University of Edinburgh, United Kingdom

Early on in her behavioural observations of the chimpanzees at what is now known as Gombe National Park, Jane Goodall was struck by their personalities, which were as distinct as our own¹. However, upon sharing her observations with a ‘respected ethologist’, she was told that, yes, animals differed in their behaviour, but that this was best ‘swept under the carpet’ (pp 11-12)².

Goodall, thankfully, felt differently and in 1973 Peter Buirski, then an Assistant Professor in the Department of Psychology at John Jay College of Criminal Justice, and a student in the Psychoanalytic Institute of Postgraduate Center for Mental Health, travelled to Gombe. While there he asked seven students and post-doctoral researchers to use the ‘Emotions Profile Index’ to describe the personalities of 24 chimpanzees³. These descriptions were used to assign scores on eight dispositions – Trustful, Distrustful, Controlled, Dyscontrolled, Aggressive, Timid, Depressed and Gregarious – to each chimpanzee. Ratings of the same chimpanzees by two or more researchers were similar. The ratings also showed differences in dispositions between the sexes; some of which resembled observed gender differences in humans⁴ (e.g. female chimpanzees were more Trustful than males), and others that did not (e.g. male chimpanzees were more Gregarious than females). In addition, personality in males was associated with their position in the social hierarchy⁴. Finally, these data predicted a series of dramatic events; Passion (a chimpanzee with a ‘deviant’ profile) together with her daughter, kidnapped and ate the infants of other chimpanzees⁵.

My interest in chimpanzee personality began after I started my PhD in Psychology at the University of Arizona. Having initially planned to study moth learning, Jim King introduced me to personality research by inviting me to work on data on captive chimpanzee personality that he had collected. He ended up becoming my PhD supervisor and a long-time collaborator. Since then, I have investigated personality in captive chimpanzees and other non-human primates as well as studying personality, health and ageing in humans, the latter of which I started during my post-doctoral training under Paul Costa at the National Institute on Aging.

Chimpanzee in a patch of sunlight in 2010

Shortly after my post-doc and after starting my current position at Edinburgh, I spoke to Anne Pusey and Mike Wilson about collecting chimpanzee personality data at Gombe. Both were enthusiastic and encouraging, and in 2009 I obtained enough funding to make it possible. Things then moved quickly. Mike and Anne translated the Hominoid Personality Questionnaire, which I developed with Jim King, into Swahili. They also identified who among the Tanzanian field assistants would rate which chimpanzees. We included deceased individuals as these included famous chimpanzees, such as Flo, which were familiar to the older field assistants.

My first visit to Tanzania took place in 2010 from late September to early November. I spent most of that time at the Gombe Stream Research Centre in Kigoma, and gathered data on the chimpanzees of the Kasekela community. Anthony Collins (the scientific director), Shadrack Kamenya and Deus Mjungu contacted and arranged visits from the field assistants. Bernadetha Tungu, my capable translator, assisted with the data collection. The sheer height of the stack of questionnaires that each field assistant was asked to complete would intimidate many people, but they took the task seriously and diligently worked through the paperwork. I then travelled to Gombe National Park where I watched the field assistants follow and record the behaviours of individuals. It was also the first time that I saw chimpanzees in the wild. Highlights included an impressive, fearless high-ranking male, playful juveniles and a chimpanzee in a patch of sunlight, with a beatific look on his or her face.

My second trip to Tanzania took place over two weeks in the late summer of 2011 and I was joined by my new wife, Emily. We spent nearly the whole time in Gombe. Again, Anthony, Shadrack, Deus, and Bernadetha provided crucial support, and the field assistants came through, this time rating members of the smaller Mitumba community. At the end of both trips, I had a staggering amount of data on perhaps the most famous wild chimpanzees in the world.

In their papers, Buirski and his colleagues provided personality scores for the chimpanzees^3,5. It was therefore possible to test how well the ratings I collected lined up with those collected in 1973. The answer turned out to be ‘surprisingly well’, especially considering the use of different questionnaires by different people, who would have had distinct experiences with the chimpanzees. However, there were also surprises, such as the lack of correspondence between a few pairs of similar personality scales.

I enjoy sharing my memories of Gombe, the chimpanzees, and the field assistants, and I wanted to share these data too, so that others can study wild chimpanzee personality and its attendant fitness consequences. I was thus delighted to receive a Sabbatical Scholarship from the National Evolutionary Synthesis Center in 2014, which enabled me to work closely with Anne Pusey, Steffen Foerster, and others to make these data available and thereby share the field assistants’ knowledge with the wider scientific community. Our personality data, along with our comparisons to the 1973 ratings, were published today in Scientific Data⁶.

The study of animal personality no longer has the pariah status it once “enjoyed”. However, less work has focused on large, long-lived mammals in the wild. By making these data freely available and accessible, I hope to make studying one such species possible for current researchers and for future generations too.

References

1. van Lawick-Goodall, J. In the shadow of man. (Houghton Mifflin, 1971).
2. Goodall, J. Through a window: 30 years of observing the Gombe chimpanzees. (Weidenfeld & Nicolson, 1990).
3. Buirski, P., Plutchik, R. & Kellerman, H. Sex differences, dominance, and personality in the chimpanzee. Anim. Behav. 26, 123-129 (1978).
4. Costa, P. T., Jr., Terracciano, A. & McCrae, R. R. Gender differences in personality traits across cultures: Robust and surprising findings. J. Pers. Soc. Psychol. 81, 322-331 (2001).
5. Buirski, P. & Plutchik, R. Measurement of deviant behavior in a Gombe chimpanzee: Relation to later behavior. Primates 32, 207-211 (1991).
6. Weiss, A. et al. Personality in the chimpanzees of Gombe National Park. Sci. Data 4, 170146 (2017).