The reviewer’s role when an integrity concern arises
Reviewers may encounter apparent inconsistencies, overlap, unusual images, ethical omissions or suspicious review-related activity. Their task is to identify and describe concerns that may affect the reliability or legitimacy of the work, not to determine guilt or impose sanctions.
A concern can arise from an innocent error, inadequate reporting, a disputed methodological choice or possible misconduct. Keep these possibilities separate. The editor needs an accurate description of what was observed and why it matters, without premature allegations.
Distinguish an error, a reporting gap and possible misconduct
Missing methodological detail may justify a request for clarification; a numerical discrepancy may require correction or reanalysis; a pattern suggesting fabrication may require a confidential integrity assessment. Similar surface features can have different explanations.
Assess the severity and relevance of the observation. Ask whether the issue could change the results, ethics assessment, attribution or editorial decision. Do not equate disagreement over interpretation with misconduct.
Document observations before drawing conclusions
Record the precise manuscript section, figure panel, table cell, dataset description or citation that triggered the concern. Describe observable facts and, when useful, explain a plausible alternative explanation that the editor may need to check.
Use proportionate language such as ‘Figure 2 appears to contain a repeated region’ rather than declaring that an image was fabricated. Preserve the context necessary for editorial assessment, but do not create or circulate unnecessary copies of confidential files.
Raise concerns through the confidential editorial channel
Submit potentially sensitive integrity observations directly to the handling editor or designated research-integrity contact. If the review form separates comments to authors from confidential comments, use the latter for allegations or material that could compromise a fair assessment.
If an issue is simply a reporting deficiency, an ordinary author-facing comment may be appropriate. When uncertain about the channel, ask the editor rather than contacting the authors or their institution yourself.
Preserve confidentiality and procedural fairness
Unpublished manuscripts and the fact of an investigation may be confidential. Do not discuss suspicions with colleagues, post figures online, alert social-media communities or seek informal opinions by distributing the manuscript.
Confidentiality protects authors, participants and the integrity of the editorial process. It does not prevent appropriate escalation through the journal’s authorized procedures.
Text overlap, plagiarism and text recycling
Similarity between passages may reflect quotation, common methodological language, a preprint, legitimate reuse with attribution, redundant publication or plagiarism. The reviewer should identify relevant passages and known sources rather than treating a similarity percentage as a verdict.
Distinguish overlap in background or standard methods from duplication of findings or unattributed appropriation of original ideas. If the suspected source is publicly available, provide a precise citation or stable reference to the editor.
Redundant publication, salami slicing and related manuscripts
Multiple papers using a shared cohort or dataset can be legitimate when they address distinct questions and disclose overlap. Concern arises when related work is concealed or the same findings are presented as independent evidence.
Identify the potentially overlapping publication and the specific similarity in participants, dates, outcomes, figures or analyses. The editor can request related manuscripts, examine disclosures and determine whether further assessment is necessary.
Image integrity and potentially duplicated visual evidence
Look for unexplained duplicated panels, inconsistent labels, implausible image boundaries, selective presentation or a mismatch between legends and displayed evidence. Image processing can be legitimate when disclosed and applied appropriately.
Do not infer misconduct from compression artifacts, visually similar biological specimens or an unfamiliar imaging technique. State the panel and observation precisely; specialist image review and requests for original files belong to the editorial process.
Numerical anomalies, data integrity and statistical signals
Conflicting denominators, impossible dates, implausible distributions, unexplained exclusions or inconsistent summary statistics can affect confidence in the results. Some discrepancies arise from transcription errors or unclear reporting.
Describe the specific inconsistency and its consequence. Reviewers should not run an extensive forensic investigation using confidential data outside the agreed review scope; suggest specialist assessment when warranted.
Research ethics approval and consent concerns
Check whether the reported work includes ethics review information, consent arrangements and participant safeguards appropriate to the study. Approval, exemption, waiver of review, waiver of consent and consent for identifiable publication are different determinations.
If a statement is missing or inconsistent, identify the gap. Do not demand identifiable participant records or presume unethical conduct because a familiar approval format was not used; requirements can differ by design and jurisdiction.
Clinical trial registration, protocols and selective reporting
For clinical trials and other prospectively registered work, compare the reported registration details, protocol timing, outcomes and amendments when these are accessible and relevant. Changes can be legitimate but should be transparently explained.
Flag discrepancies between prespecified and reported outcomes, unclear registration timing or unexplained analyses. Avoid asserting intentional selective reporting when documentation is incomplete.
Authorship, contributorship and acknowledgements
Unusual author lists, missing contribution details, inconsistent acknowledgements or evidence of undisclosed writing support may warrant clarification. Reviewers generally cannot verify authorship from a manuscript alone.
Report concrete discrepancies rather than speculating about ghost or honorary authorship. The editor can seek contribution statements and coordinate any authorship dispute through appropriate channels.
Undisclosed funding and competing interests
A manuscript may describe funding or commercial relationships inconsistently across the title page, methods, disclosures and acknowledgements. Relevant interests are broader than direct study funding.
If you know of a material undisclosed relationship, provide the editor with a factual description and a reliable basis for the concern. Avoid investigating private personal information or treating an apparent omission as proof of deliberate concealment.
Data and code availability discrepancies
A data-availability statement should reflect actual access conditions and the research’s ethical or legal constraints. A repository link that does not resolve, missing analysis code or unexplained restrictions may warrant clarification.
Public release is not always appropriate, especially for sensitive participant data. Distinguish failure to meet a stated availability commitment from a legitimate controlled-access arrangement.
AI-generated content and suspected synthetic evidence
Fabricated citations, invented data, manipulated images or unsupported claims may arise with or without AI. Stylistic features and automated AI-detection scores are not reliable proof that authors used AI improperly.
Identify verifiable problems—such as a nonexistent reference or contradictory data—rather than alleging AI misconduct based on writing style. Apply the journal’s AI policy and refer any material concern to the editor.
Citation manipulation and inappropriate reviewer demands
Citation patterns may suggest irrelevant reference padding, excessive journal self-citation or omission of essential work. Evaluate relevance to the manuscript rather than relying on counts alone.
Reviewers must not exploit an integrity concern to demand citations to their own publications or collaborators. Any recommended references should have a clear scholarly purpose.
Peer-review manipulation and compromised identities
Suspicious reviewer suggestions, implausible contact details or signs that confidential material was circulated may warrant editorial attention. Reviewers rarely have enough information to establish manipulation independently.
Report specific observations privately. Do not attempt to access accounts, test suspected identities through deceptive messages or conduct an independent inquiry.
When the concern involves your own conflict
If a suspected problem overlaps with your competing work, prior dispute, collaboration or personal relationship, disclose that context immediately. The editor may seek an independent assessment.
Do not use the integrity process to settle scientific rivalry. A genuine concern can be important even when the reviewer has a conflict, but transparent disclosure is necessary to manage it.
Requests for raw data, images and supplementary evidence
A reviewer may recommend that the editor request original images, analysis code, protocols or supporting records when they are necessary for assessing a concrete concern. Requests should be targeted and proportionate.
Do not obtain identifiable participant records directly, ask authors to send files to personal accounts or bypass repository access controls. The editor should manage sensitive evidence and any specialist review.
Writing an actionable confidential integrity report
A useful report states the observation, its location, its potential relevance, what is currently unknown and what verification might resolve the issue. Separate directly observed facts from hypotheses.
Avoid legalistic accusations, certainty beyond the evidence and long lists of speculative concerns. A concise evidence-led report helps the editor decide whether ordinary clarification, specialist review or a formal process is appropriate.
Continue scientific review without prejudging the outcome
Depending on the nature of the concern, the editor may ask the reviewer to continue, pause or focus on a narrower issue. Follow those instructions and keep ordinary methodological comments distinct from confidential integrity observations.
An investigation can take time and may involve information the reviewer cannot access. The absence of immediate feedback does not establish that the concern was ignored.
Revisions, post-publication concerns and ongoing duties
A revised manuscript may resolve an apparent inconsistency or reveal a new issue. Assess the response on its evidence and do not treat a corrected error as proof of prior intent.
If a material issue comes to light after publication, use the journal’s established reporting channel. Corrections, expressions of concern and retractions are editorial mechanisms applied according to the evidence and applicable policy, not outcomes a reviewer independently imposes.
Treata integrity escalation standard for reviewers
Treata reviewers should recognize potential integrity signals, document observable evidence, protect confidential material and report concerns through authorized editorial channels. They should not accuse, investigate institutions independently or publish suspicions.
The editor coordinates assessment under the relevant publisher and journal policies. Reviewers should remain available for factual clarification while respecting confidentiality, conflicts and procedural fairness.
Continue reviewer guidance
Related reviewer guides
Publisher-wide standards