For Referees

Guide to referees

About Scientific Data

Scientific Data is a peer-reviewed open-access journal for descriptions of datasets, and research that advances the sharing and reuse of scientific data. 

Data Descriptors, the primary article-type published at Scientific Data, provide detailed descriptions of experimental and observational datasets, including the methods used to collect the data and technical analyses supporting the quality of the measurements. Data Descriptors should not contain tests of new scientific hypotheses, extensive analyses aimed at providing new scientific insights, or descriptions of fundamentally new scientific methods. Scientific Data's peer-review process evaluates the technical quality of the experiments used to generate the data and the completeness of the description of the data. The actual data files are stored in one or more public, community-recognized repositories, and must be made publicly available when the Data Descriptor is published.

Criteria for publication

For all article types, the report should be technically sound, scientifically valid and make an original contribution to the field. Beyond this, all relevant previous literature should be correctly cited and the methods for data sharing appropriate for the data type. Scientific Data publishes across all science and data-oriented social science areas. There is no restriction on scale or niche as long as the dataset is perceived as being useful to someone.

Data Descriptors

These are descriptions of open research datasets, presented in a manner that promotes reuse, without reporting whether datasets support hypotheses or conclusions. 

Scientific Data accepts descriptions of research datasets that are not described at the same level of detail as elsewhere and do not contain flaws in experimental design, data collection or processing that would inhibit their re-use. Referees may request additional information, analyses or experiments when they are needed to convincingly support this. In rare cases reviewers may consider a dataset to be so simplistic, routine, or understandable in its own right as to not require an associated Data Descriptor, however if there is any perceived advantage in disclosing the methods, data availability, or details or usage or technical validation of the data being described then publication of a Data Descriptor should always be considered. Authors are required to provide referees with access to all data (complete as per the description in the manuscript), and full public release of the research data through a trusted community data repository is required before publication (please see our data deposition policies). For human-derived data, please see our section on Human Data.

Evaluation of the submission should not be based on the perceived impact or novelty of the findings associated with the datasets. Data Descriptors should stay focused on describing dataset creation, supporting the quality and rigour of data-generating procedures and providing any information that other researchers need for data interpretation and reuse. Authors should not impose specific interpretations on the data, and Scientific Data will not be a place to publish preliminary or weakly supported results.

In line with this focused scope, referees may ask authors to remove any analyses or claims that appear to go beyond this scope.


These are reports on data policy, repositories, standards, ontologies, workflows, or any topic relating to the mechanics of data sharing. Scientific Data does not publish traditional research articles using data to validate regular scientific hypotheses

Scientific Data accepts Articles that are technically sound on the topic of research data sharing. There are no criteria for perceived significance or impact as long as the reported system has some perceived use to the community, even if the audience is niche.


These are re-analyses or re-assessments of existing open data.

Scientific Data accepts papers that provide new insights or interpretations of existing datasets, either through re-processing, aggregation, or other methods. They should not report 100% new data however we expect any results of re-processing to be shared. There are no criteria for perceived significance or impact as long as the analysis has some perceived use to the community, even if the audience is niche.


These are short commentaries or opinions on research data policy, workflows or infrastructure that can be shorter and more speculative than Articles and do not need to report a specific/extant technology or finding.

Scientific Data accepts Comments that are technically sound on the topic of research data sharing. There are no criteria for perceived significance or impact as long as the reported system has some perceived use to the community, even if the audience is niche.

The review process

Scientific Data is an inclusive journal and will consider all submissions for peer review that fit our scope. In addition to this, members of our editorial team will assess submissions for compliance with our data policy and will communicate any changes required to prepare the paper for review. On completion of this step an Editorial Board Member will be assigned to oversee the peer-review process. Editorial Board Members may choose to recommend rejection without peer review for works that do not meet Scientific Data’s scope or standards, based on their own experience and expertise, however we expect the majority of publications to be assessed by reviewers. 

During peer review the Editorial Board Member will invite one or more referees to evaluate the submission. After consideration, the Editorial Board Member will make one of the following decisions:

After consideration, the Editorial Board Member will make one of the following decisions:

  • Accept, with or without editorial revisions
  • Request a minor revision, giving authors one month to revise their manuscript to address specific concerns
  • Request a major revision, giving authors three months to revise their manuscript to address significant concerns and perhaps undertake additional work
  • Reject 

The Editorial Board Member or Editor may occasionally need to go back to the referees for further advice. We therefore ask that referees be willing to provide follow-up advice as requested. When agreeing to review a paper, we may occasionally ask referees to review a subsequent revision; however, Editorial Board Members will not send resubmitted Data Descriptors to referees if it seems that the authors have not made a serious attempt to address the referees’ criticisms.

Find out more about our Editors and Editorial Board.

Selecting Referees

Referee selection is critical to the review process, and the Editorial Board Member should base his or her choice on many factors, including expertise, specific recommendations and previous experience. When inviting referees to review manuscripts, the Editorial Board Member will send confidential information, which should be treated as such, with the invitation letter.

Writing the review

A referee’s report should provide members of the Editorial Board with the information needed to reach a decision and should instruct authors on how they can strengthen their paper to the point where it may be acceptable for publication.

Referees should feel free to request additional experiments or analysis when needed to support the data. Referees, however, should avoid asking for extensive follow-up experimentation, or confirmation of specific hypotheses or interpretations, which would fall outside the scope of a Data Descriptor manuscript. Indeed, referees may ask authors to remove in-depth analyses or new scientific conclusions from submitted Data Descriptors.

Evaluation of a Data Descriptor manuscript should not be based on the perceived impact or novelty of the findings associated with the datasets. The peer-review process should remain focused on data quality and reusability, not specific interpretations.

When preparing a report, we ask referees to consider and comment on the following questions:

Experimental Rigour and Technical Data Quality

  1. Were the data produced in a rigorous and methodologically sound manner?
  2. Was the technical quality of the data supported convincingly with technical validation experiments and statistical analyses of data quality or error, as needed?
  3. Are the depth, coverage, size and/or completeness of these data sufficient for the types of applications or research questions outlined by the authors?

Completeness of the Description

  1. Are the methods and any data-processing steps described in sufficient detail to allow others to reproduce these steps?
  2. Did the authors provide all the information needed for others to reuse this dataset or integrate it with other data?
  3. Is this Data Descriptor, in combination with any repository metadata, consistent with relevant minimum information or reporting standards?

Integrity of the Data Files and Repository Record

  1. To the degree that you have viewed the actual data files, did they appear complete and do they match the descriptions in the Data Descriptor?
  2. Have these data files been deposited in the most appropriate available data repository?


We ask all Editorial Board Members and external referees to treat the review process as strictly confidential, and not to discuss the manuscript with anyone not directly involved in the review. It is acceptable to consult with laboratory colleagues, but please identify them to the Editorial Board Member. Consulting with experts from outside the referee's own laboratory may also be acceptable, but please check with the Editorial Board Member before doing so, to avoid involving anyone who may have been excluded by the authors.


Scientific Data is committed to rapid editorial decisions and publication, and we believe that an efficient editorial process is a valuable service both to our authors and to the scientific community as a whole. We therefore ask referees to respond promptly (within ten days of receiving a manuscript, although this may be either longer or shorter by prior arrangement). If referees anticipate a longer delay, we ask them to let the Editorial Board Member know so that we can keep the authors informed and, where necessary, find alternative referees.


We do not release referees' identities to authors or to other referees, except when referees specifically ask to be identified. Unless they feel strongly, however, we prefer that referees remain anonymous throughout the review process and beyond.

We ask referees not to identify themselves to authors without the Editorial Board Member's knowledge. If they wish to reveal their identities, this should be done via Scientific Data's publishing team.

We deplore any attempt by authors to confront referees or determine their identities. Our own policy is to neither confirm nor deny any speculation about referees' identities, and we encourage referees to consider adopting a similar policy.

Editing referees' reports

As a matter of policy, we do not suppress referees' reports; any comments that were intended for the authors are transmitted, regardless of what we may think of the content. On rare occasions, we may edit a report to remove offensive language or comments that reveal confidential information about other matters. We ask referees to avoid saying anything that may cause needless offence; conversely, authors should recognize that criticisms are not necessarily unfair simply because they are expressed in robust language.

Competing interests

Our normal policy is to avoid Editorial Board Members and referees whom the authors have excluded, for whatever reason. We also usually try to avoid referees who have recent or ongoing collaborations with the authors, who have commented on drafts of the manuscript, who are in direct competition to publish the same finding, whom we know to have a history of dispute with the authors, or who have a financial interest in the outcome. It is not possible for the Editorial Board or publishing team to know of all possible biases, however, so we ask referees to draw our attention to anything that might affect their review, and to decline to review in cases where they feel unable to be objective.

We recognize, however, that competing interests are not always clear-cut, and the above circumstances need not automatically undermine the validity of a report. Indeed, the people best-qualified to evaluate a paper are often those closest to the field, and a sceptical attitude towards a particular claim does not mean that a referee cannot be persuaded by new evidence. Editorial Board Members try to take these factors into account when weighing referees' reports.

Referees who have reviewed a paper for another journal might feel that it is unfair to the authors for them to re-review it for Scientific Data. We disagree; the fact that two journals have independently identified a particular person as well-qualified to review a paper does not, in our view, decrease the validity of his or her opinion.

Online manuscript review

Referees must submit their comments via our online submission system by following the link provided in the editor's email. For help with the system please contact the MTS helpdesk.