Peer review
Declared on the three independent axes of NISO Z39.106-2023. The terms "single blind" and "double blind" are not used, because they compress three separate decisions into one word and are read differently by different people.
- Identity transparency
- Double anonymised — the reviewers are not told who the authors are, and the authors are not told who the reviewers are
niso-z39.106:identity-transparency = double-anonymised - Reviewer interacts with
- With the editor only
niso-z39.106:reviewer-interacts-with = editor - Review information published
- None. Reports and reviewer identities are not published.
niso-z39.106:review-information-published = none - Post-publication commenting
- Moderated. Comments are reviewed before they appear and are not anonymous.
The numbers the software enforces
- Minimum completed reviews before acceptance
- 2. This is a guard on the accept transition, not a target: the platform refuses the decision if fewer reports are on file.
- Time allowed for a review
- 21 days from acceptance of the invitation.
- Reviewer credit
- Reviews are credited to the reviewer’s ORCID record.
- Review DOIs
- Reviews are not deposited as citable records, which follows from not publishing them.
Anonymity is enforced, not requested
Under a double-anonymised policy the reviewer's copy of the manuscript is stripped of author names, affiliations, ORCID iDs, ROR identifiers, email addresses and institution names before it is rendered — the reviewer's response never contains those fields, rather than containing them behind a filter applied in the browser. The platform additionally scans free text for author surnames and affiliation strings and warns the editorial office when a manuscript would identify its authors through the text itself.
The editor decides, and must say why
Reviewers recommend; the editor decides. Where an editor accepts a manuscript that a reviewer recommended rejecting, the platform requires a written reason for departing from the recommendation before it will record the decision, and that reason goes onto the permanent file. This follows COPE's position that editorial decisions must be documented and defensible — not that they must agree with reviewers.
Conflicts of interest
Reviewers are screened automatically against the authors' institutions, recent co-authorship and shared funding. A blocking conflict removes the role rather than flagging it. An editor who is an author on a submission holds no editorial role on that submission and is not offered the decision. The screen and its severities are set out on the ethics page.
What the workflow will refuse
Every one of these is a condition checked by the platform at the moment of the decision. There are 18 guarded transitions in the current workflow; these are the ones bearing on review.
- Reject without review: Only the handling editor or the editor-in-chief may desk-reject. (ICMJE: decisions rest with the named editor)
- Request revision: Only the handling editor may issue this decision.
- Reject: Only the handling editor may issue this decision.
- Accept: Acceptance requires 2 completed reviews; — are on file. (NISO Z39.106 declaration: minimum reviews)
- Accept: — reviewer(s) recommended rejection. Acceptance is permitted — the editor decides, not the reviewers — but a written reason for departing from the recommendation must be recorded on the file. (COPE: editorial decisions must be documented and defensible)
- Accept: The record still carries — blocking problem(s); an accepted article cannot.
- Accept: De-identification has not been signed off.
- Accept: Consent is not on file.
- Publish: De-identification sign-off is not on file. Nothing publishes without it. (Journal policy: de-identification)
- Publish: Consent for publication is not on file. (ICMJE: patient consent for publication)
- Publish: The record still carries blocking problems.
- Publish: The article is below the adequacy threshold.
- Publish: No DOI has been minted for this article. (Crossref deposit requires a DOI)