Publishers’ requirements
Tsuyoshi Miyakawa, editor-in-chief of Molecular Brain, says that, in reference to the reproducibility crisis, he has rejected 40 out of 181 articles over the past two years due to a lack of supporting data. (1).
In 2019, a study titled “Effect of impact factor and discipline on journal data sharing policies” examined the data-sharing policies of 447 journals spanning various medical science disciplines. Only 12 of these journals — 2.7% of the sample — had a strict policy that made data sharing a condition for publication.
Two factors can influence the adoption of high standards for data sharing. First, the impact factor. Journals with a high impact factor are more likely to emphasize reproducibility and the submission of data for peer review (2). Second, the discipline. A study conducted by COPE conducted by COPE in 2015 examined 150 journals and found that 30% of biomedical science journals have a very proactive policy on data sharing, compared to 10% of physics and chemistry journals.
Among publishers of physics and chemistry journals, there are significant disparities in requirements and ambitions regarding data publication and verification. Below, we provide a summary of the practices currently in place within scholarly societies and general-interest publishers.
Les sociétés savantes
Scholarly societies
American Chemical Society
Since February 2020, the American Chemical Society (ACS) has launched a platform designed to facilitate the submission of nuclear magnetic resonance (NMR) data in accordance with FAIR principles. Currently, only two organic chemistry journals, the Journal of Organic Chemistry and Organic Letters, are participating in this pilot project by encouraging authors to use the platform to submit their FID files. Currently, the data is not stored on ACS servers, although this is a potential future development for the project.
Guidelines are available for each ACS-published journal in the Publications Center. General recommendations regarding NMR data are published here. Crystallographic data must be verified using the CheckCIF tool. A list of accepted file formats for ‘supporting information’ is also provided.
To explore the policies of individual journals in more detail, use the drop-down menu to consult the guidelines regarding ‘supporting information’ and ‘data requirements’. These requirements can vary significantly from one journal to another.
- The Journal of the American Chemical Society devotes a considerable amount of space to this topic in its author guidelines.
- The Journal of Organic Chemistry also provides extensive guidelines. For example: “The Editors would like to point out that reported yields above 95%, isomer ratios above 200:1, and enantiomeric excesses above 99% typically aren’t realistic without further explanation”.
- The Journal of Proteome Research emphasises the requirement for authors to deposit raw data and associated metadata in repositories such as ProteomeXchange, providing access details (URLs or passwords, if necessary) in the manuscript. “Access to the information will be kept confidential while the manuscript is under review but will be open to the public upon publication”.
- The ACS Nano journal outlines its ‘data requirements’ in its instructions to authors. In particular, authors must provide robust evidence to establish the identity and purity of the substances they are studying. The expected specifications depending on the types of data are specified.
Royal Society of Chemistry
RSC has a clear and detailed policy: ‘For all submissions to Royal Society of Chemistry journals, any data required to understand and verify the research in an article must be made available on submission.’
Depending on the nature of the data, the guidelines may vary. X-ray crystallography data, for example, must be deposited in a dedicated repository. Chemical compounds and spectra, on the other hand, may be included in the ‘supporting information’. So-called ‘additional’ data, which is not mandatory but may be useful for future research, may be deposited in a discipline-specific or general repository.
Spectral data must be provided in JCAMP-DX format. Detailed informations is provided to facilitate the description of substances based on approximately ten criteria, such as yield, melting point, spectra and refractive index. Crystallographic data must, of course, be in CIF format and accompanied by a CheckCIF report.
In addition to these general guidelines, RSC journals may include their own recommendations depending on their subject areas. For instance, ChemComm requires authors to provide a full description of the conditions in which magnetic measurements of samples were taken, including the type of capsule used. Meanwhile, the journal Energy & Environmental Science requires detailed information on the calibration protocol for solar converters.
American Institute of Physics
AIP publishes (or co-publishes) 46 journals, proceedings, and books.
AIP recommends that authors deposit their datasets in publicly accessible repositories or include them in the main manuscript so that readers can access all the datasets on which the article’s conclusions are based. AIP’s research data policy is available here.
At the time of submission, the author must provide a “Data Availability Statement” (see: Download Data Availability Statement Templates).
The journals generally align with AIP’s general recommendations and do not adopt their own standards, unlike ACS journals, for example.
Institute Of Physics
The IOP publishes (or co-publishes) 108 journals and conference series. The IOP is a member of the Committee for Publication Ethics (COPE).
General policy: “IOP Publishing supports the principles of transparency and openness in scientific research, with the reproducibility of research facilitated by the availability of data, code and research materials underpinning research articles”. The policy is available here. Two aspects regarding data standards and availability are detailed: The Standard data policy and the Data availability policy.
All journals published by IOP adhere to these guidelines. However, it is necessary to consult the specific policies of individual journals, which are listed in the ‘About the Journal’ section of their websites.
Supplementary materials: Files containing supplementary material may not exceed 10 MB each. Authors are advised to deposit larger files in a data repository.
Modalités de mise à disposition : IOP recommande la publication dans un entrepôt disciplinaire ou, à défaut, généraliste. Les entrepôts cités par IOP sont : Figshare, Dryad, Harvard Dataverse, Zenodo. A noter : IOP Publishing a développé un entrepôt dédié sur Figshare (voir : dedicated repository on the Figshare platform), utilisé par les revues Environmental Research Letters, IOP SciNotes, JPhys Complexity, Machine Learning: Science and Technology.
Availability requirements: The IOP recommends publication in a subject-specific repository, or a general-purpose repository if a subject-specific one is not available. The repositories cited by IOP are: Figshare, Dryad, Harvard Dataverse and Zenodo. Please note that IOP Publishing has developed a dedicated repository on the Figshare platform, which is used by the following journals: Environmental Research Letters, IOP SciNotes, JPhys Complexity and Machine Learning: Science and Technology.
These journals have a policy requiring the submission of a data availability statement, which can be found here.
Please note that any journal published by IOP Publishing on behalf of another company or organisation may have its own policy regarding research data. This is the case with the American Physical Society’s journals, for example, which have a ‘data guide’ available here.
American Physical Society
The American Physical Society (APS) publishes 18 journals, including Physical Review A, B, C, D, E, Physical Review Letters, Reviews of Modern Physics, Physical Review Applied. As stated in the ‘Editorial Policies and Practices (Data Availability)’ document, the Physical Review journals strongly encourage authors to share the data, code, software and other relevant materials that support their results by depositing them in open-access data repositories at the time of publication. Authors may indicate the availability of such materials in a Data Availability Statement in the paragraph preceding the acknowledgements.
Society of Photo-Optical Instrumentation Engineers
SPIE, a academic society that publishes 14 journals specialising in the study of light, states in its ethical guidelines (SPIE Guidelines for Ethical Publishing) that ‘the results of research should be recorded and maintained in a form that allows analysis and review.’. The instructions for authors state that SPIE journals require a code, data and materials availability statement for all accepted articles. This requirement took effect on 1 May 2023. These statements must describe how to access all the data necessary to reproduce or interpret the results described in the article. Authors are encouraged to make data and code available in repositories such as FigShare, GitHub and CodeOcean.
American Geophysical Union
The American Geophysical Union publishes 23 journals through Wiley, including Geochemistry, Geophysics, Geosystems, Reviews of Geophysics and JGR: Solid Earth.
The organisation has adopted a very detailed data policy. In its Position Statement on Data, the American Geophysical Union stated that ‘Earth and space science data should be widely accessible in multiple formats, and the long-term preservation of data is an integral responsibility of scientists and the institutions that sponsor them.’ Accordingly, the AGU politicy requires that ‘the underlying data and/or software or code needed to understand, evaluate, and build upon the reported research be available at the time of peer review and publication. Additionally, the code (e.g. Python, Jupyter Notebooks, R, MATLAB) used to perform any data analysis and to produce the manuscript’s figures should be made available in a free and open platform (e.g. Github) and preserved in a repository (e.g. Zenodo)’. AGU journals adhere to the FAIR data principles (see the guidelines for enabling FAIR data).
Therefore, authors are encouraged to identify and archive the data associated with their article in a widely used scientific community repository, retaining it for at least five years or providing access through a clear process. The AGU reserves the right to refuse publication in the event of non-compliance with these provisions.
Authors have access to comprehensive guidelines, including templates for indicating data availability and recommended repository types.
European Geosciences Union
The EGU publishes 19 open-access journals via Copernicus Publications, including Geoscientific Model Development and Atmospheric Chemistry and Physics (ACP). It is a signatory to the 2003 Berlin Declaration and the Open Access 2020 initiative, and has endorsed the commitment statement of the Coalition on Publishing Data in the Earth and Space Sciences (COPDESS) and the Enabling FAIR Data Commitment Statement in the Earth, Space, and Environmental Sciences. Each journal’s “Data Policy” (for example here) requires that data corresponding to articles be deposited in a repository and a data availability statement included in the final paragraph of each article. The EGU has also provided further guidance on big data and the Earth sciences.
General publishers
Like specialist publishers, the major publishers of scientific journals such as Elsevier, Springer Nature and Taylor have established policies for access and distribution of data for articles published in their journals.
Springer Nature
As part of its standard policy, the Springer Nature Group (which publishes Nature Physics, Nature Chemistry, Nature Materials, Nature Photonics, La Rivista del Nuovo Cimento, among others) requires a data availability statement to be included with every published article. This statement must specify which data are available, where they can be found and under what conditions. It must also specify whether the data are original or reused. Authors are strongly encouraged to deposit their data in a repository at the time of publication. The publisher provides a list of recommended repositories organised by discipline. For certain types of data, such as proteomics data, small-molecule crystallography data or macromolecular structures, deposit in a public repository is mandatory. This must be accompanied by a citation in the article that includes the persistent identifier for the dataset.
Reviewers may request access to data and code when they deem it necessary during the evaluation of the manuscript. If data is too sensitive to be shared publicly, the conditions for accessing it must nevertheless be stipulated in the article.
The group has stated that it aims to have all its publications adopt this standard policy. This is already the case for titles published by BMC, Nature and Springer Open. However, as the transition is still underway for other Springer and Palgrave Macmillan titles, some journals may still have a less stringent data policy than the one outlined here.
Elsevier
“Our policy is clear in that we encourage and support authors to share their research data rather than mandating them to do so,” the publisher states. Among potential repositories, Elsevier recommends its own, Mendeley Data, as well as about sixty other more specialized repositories.
Elsevier’s journals uniformly refer to the data availability statement that authors are asked to provide and to the group’s overall policy. For the Journal of Computational Physics, co-submission to Data In Brief and SoftwareX is also offered. In chemistry, data can be highlighted in Chemical Data Collections.
Wiley
The level of requirement at this publisher can vary significantly depending on whether the journal ‘encourages’ data submission, requires it or mandates peer review of the data. Wiley has a partnership with Dryad, a paid data repository. Some of its journals in the medical and environmental sciences fields cover the associated costs.
Researchers are generally asked to provide a ‘statement on data availability‘, in which they can specify whether the data is openly accessible in a repository, subject to an embargo or available upon request. Further details on what is expected of authors regarding data can be found in an FAQ.
It is necessary to review the policies in place for each individual journal.
Both Angewandte Chemie International Edition and Chemistry, a European journal provide fairly detailed guidelines depending on the type of data to be submitted (NMR, infrared, and mass spectroscopy; catalysis; specific methods used for energy conversion and storage, etc.)
In contrast, Polymers for Advanced Technologies is not very specific on this point, stating simply that an article may not contain more than five figures—unless understanding the article depends on them—with any additional figures to be included in the “supplementary information.” Data should be made publicly available whenever possible, but this is not a requirement.
Taylor and Francis
This publisher offers a range of journals, the data disclosure policies of which can vary from optional to mandatory. As with Wiley, authors are required to submit a statement on data availability.
More stringent criteria apply to six géoscience journals, for which data must be deposited in appropriate repositories. As with ACS, a partnership has been established with Figshare to host the ‘supplementary information’ for articles via a dedicated subdomain.
A review of most chemistry journals reveals two recurring requirements. Firstly, the publisher requires the assignment of a DOI for any datasets associated with the article. Secondly, it notes that these datasets are not subject to peer review. ‘It is the author’s responsibility to ensure the validity of the data,’ the policy states.
EDP Science
As a publisher committed to open science, EDP Science outlines its data-sharing policy in the context of research funding. ‘EDP Sciences journals encourage authors to share and make data if legally and ethically possible.’
Authors are encouraged to deposit their data in an online repository to make it available for human and machine reading, thereby helping to accelerate scientific discoveries. They are asked to deposit their data in accordance with the FAIR data principles, and to provide a statement of data availability. These recommendations apply to the journals Acta acustica, Astronomy & Astrophysics, Journal of Space Weather and Space Climate.
Cambridge University Press
In line with the principles of transparency, the Journal of Fluid Mechanics, published by Cambridge University Press, sets out its requirements for research data in its ‘research transparency‘ section. All the information needed to reproduce the study must be provided either in the paper itself or in publicly accessible repositories.
- Miyakawa, T. No raw data, no science: another possible source of the reproducibility crisis. Mol Brain 13, 24 (2020). https://doi.org/10.1186/s13041-020-0552-2
- Cela viendrait du fait, selon les auteurs de l’étude, que les “revues ayant un haut facteur d’impact ont tendance à recevoir plus d’attention de la part des chercheurs et des médias”, induisant une pression accrue pour adopter des mesures.