Publication Ethics & Malpractice Statement

ScienceBank Publishing Society follows the principles of the Committee on Publication Ethics (COPE). This page summarises how those principles apply here and links to the policy that governs each in full.

1. Publication ethics

Every submission is assessed on the quality of its research and reporting, not on the reputation of its authors or their institutions. Editors and reviewers treat manuscripts in strict confidence, and must not use material or information gained through peer review for any other purpose.

The peer-review mode a community's manuscripts are reviewed under is set by ScienceBank when the community is created. A community owner or administrator can ask ScienceBank to change it — there is no self-service control for the mode in a community's settings — and the change takes effect only once ScienceBank or the Society Editorial Board has approved it. A change applies to the community's whole review history and not only to manuscripts submitted after it: were a community's mode changed from double-blind to open, identities withheld under the old mode would become visible on reviews that had already been completed. Every ScienceBank community currently reviews double-blind: author identities are withheld from reviewers, and reviewer identities from authors. The handling editor's decision is given to the authors in a decision letter that sets out the reasons for it.

Editors and reviewers declare any interest that could be seen to influence their judgement, and step aside where one exists. For reviewers this is enforced: a reviewer must confirm they have no conflicting interest before their review can be submitted, and the platform refuses the review without it. For editors it is an obligation handled outside the platform — ScienceBank does not check an editor for conflicts and does not record an editor's declaration. The platform checks for a conflict before a reviewer is invited: anyone listed as an author of the manuscript cannot be invited to review it, and neither can anyone who has already worked on that manuscript through ScienceBank's editorial-assistance service with a person involved. Assistance carried out by an AI tool alone is recorded but does not stop the invitation. A shared institutional affiliation does not stop the invitation — it is recorded against the reviewer's assignment and shown to editors on the manuscript's reviews list after the invitation has been sent, and no justification is requested or recorded. What counts as a competing interest, and how it is handled, is set out in full in our Competing Interests policy.

The complete standards for editorial conduct are in Ethical Publishing Practice, and the expectations placed on editors and reviewers are in the Editor and Reviewer Guidelines.

2. Malpractice

Submissions must be the authors' own original work. Plagiarised work is not published.

A similarity check is started when the manuscript is submitted, and is carried out with iThenticate (Turnitin). Its result is shown to editors on the manuscript itself; the community's editorial queue carries only a flag, and only where the reported similarity was above the threshold. It is not a gate: an editor can desk-check a manuscript, invite reviewers and issue a decision while its check is still pending or has not run. Where the reported similarity exceeds ScienceBank's threshold, the manuscript is marked as flagged, and the handling editor, the Editor-in-Chief, the community owner or a ScienceBank administrator records whether the overlap is acceptable, requires revision, or warrants rejection. No notification is sent, so a flag is seen when an editor next opens the queue. Where a report cannot be produced, the manuscript is marked as needing manual editorial review — it is never recorded as cleared.

A clear result does not end the matter — cases can still warrant correction by an editor, and authors who repeatedly attempt to submit plagiarised work are investigated and may have their account suspended. Authors who disagree with a decision taken on the basis of a similarity check may ask us to review it, using the route in section 5.

Fabrication or falsification of data, duplicate submission, and undisclosed competing interests are treated with the same seriousness. Where a concern is raised about published work, we follow the process in Post-publication Procedures.

3. Corrections, retractions and expressions of concern

Substantive changes to published findings are issued as corrections, expressions of concern or retractions, each published as its own citable record. Separately, ScienceBank administrators, editors and typesetters can amend the text of a published article in place. Every such amendment is permanently recorded, with the text before and after it, in an append-only internal log — the platform provides no way to alter or remove an entry once written; it is not currently shown to readers as a change to the article.

• A correction is issued where an error affects part of the record but not its conclusions.
• An expression of concern is published where a serious question has been raised and is not yet resolved.
• A retraction is issued where the findings are unreliable or the work breaches research or publication ethics. ScienceBank publishes transparent retractions: the article remains available and citable, marked as retracted, with the reason stated.

Each of these is published as a record in its own right, linked to the article it corrects, questions or retracts, so the article and the notice can each be cited and neither replaces the other in the record.

Content is permanently removed only in the rare circumstances set out in Post-publication Procedures, which also covers the rights and responsibilities of authors, communities and ScienceBank.

4. Authorship

ScienceBank defines an author as a person who makes a substantial contribution to the submitted work and takes responsibility for it. Contributors who do not meet that threshold are credited in the acknowledgements instead.

The submitting author lists every co-author at submission and names the corresponding author. It is the submitting author's responsibility to confirm, before submitting, that all listed authors have agreed to their inclusion, to the order in which they are listed, and to the manuscript as submitted. Every co-author is notified when the manuscript is submitted. ScienceBank does not collect Contributor Roles Taxonomy (CRediT) roles from authors at submission; where CRediT roles are recorded for an article during production, they appear in the article's JATS XML and in its PDF.

The full criteria, the CRediT role definitions and the corresponding author's responsibilities are in our Authorship policy.

5. Raising a concern

Concerns about published work, editorial conduct or a decision — including a decision taken on the basis of a similarity check — may be raised with us at admin@sciencebank.com. We acknowledge every concern and tell you what happens next.