1. The editor’s role and the limits of editorial investigation
An editor’s first responsibility when a research-integrity concern arises is to protect the integrity of the editorial process and the scholarly record. A concern may emerge during screening, peer review, revision, production or after publication. Its appearance does not establish that misconduct occurred, and an editorial inquiry should not begin from an assumption of guilt.
Editors should distinguish what a journal can assess—such as manuscript consistency, compliance with submission policies, originality of submitted material and the reliability of publication decisions—from matters that may require institutional authority, specialist investigation or legal advice. The journal can request explanations and evidence relevant to editorial decisions, but it should not represent itself as a disciplinary body.
The appropriate response is proportionate, documented and confidential. Where the handling editor has a conflict or lacks the authority to manage the concern, the matter should be transferred through the journal’s governance process.
2. Recognize and classify concerns before choosing a response
Concerns may involve overlapping publication, unattributed reuse, fabricated or falsified data, inappropriate image manipulation, authorship disputes, undisclosed competing interests, ethics approval or consent, trial registration, peer-review manipulation, undisclosed AI use or unreliable reporting. These categories help identify applicable policy; they are not verdicts about the people involved.
Start by describing the observable issue precisely. A duplicated image panel, for example, is a specific observation that may have several explanations. It is more useful to document the affected figures and their apparent similarity than to label the manuscript fraudulent before clarification.
Separate administrative omissions, correctable reporting problems, substantive reliability concerns and allegations requiring formal escalation. Different classes of concern demand different information, expertise, timelines and possible outcomes.
3. Secure the manuscript record and preserve evidence
Retain the original submission, relevant versions, reviewer correspondence, declarations, figures, supplementary files and decision history in approved editorial systems. Preserve the material that gave rise to the concern in a way that permits later verification without unnecessarily multiplying confidential copies.
Record when the concern was raised, by whom or through what channel, the specific material implicated and the actions taken. Internal notes should distinguish directly observed facts from interpretations and allegations. Avoid editing the original record or replacing earlier files without maintaining version history.
Limit access to people with a legitimate role in the case. External sharing, including with consultants or institutional contacts, should follow the publisher’s confidentiality and data-governance requirements.
4. Assess credibility and urgency without prejudging the outcome
Conduct a preliminary assessment to establish whether the concern is sufficiently specific and relevant to justify further inquiry. Consider whether the alleged problem can be located in the manuscript, whether the available information is coherent, and whether a straightforward explanation might resolve it.
Urgency depends on context. A potential participant-privacy exposure or an imminent publication of materially unreliable results may require prompt protective action. A minor citation discrepancy generally does not require the same response. Urgency should influence the timing and safeguards, not the fairness of the inquiry.
If the report is vague or anonymous, evaluate the evidence on its merits. Neither anonymity nor a named complainant is proof of accuracy or inaccuracy. Seek enough detail to identify the concern while protecting people from unnecessary disclosure.
5. Manage conflicts and assign an independent case handler
The handling editor, editor-in-chief or other authorized person should disclose relevant relationships that could impair impartial handling. A case involving an editor’s collaborator, institutional colleague or personal dispute may require reassignment or oversight by another authorized editor.
Decide who is responsible for communications, evidence custody, procedural decisions and the final publication-related outcome. Avoid parallel informal inquiries in which multiple editors contact the authors or institution without coordination.
Where the matter implicates journal leadership, use the publisher’s documented escalation or governance route. The legitimacy of the process depends on both fair treatment and a decision-maker who can act independently.
6. Seek focused explanations from authors
When an author response is appropriate, describe the specific issue and request the information needed to assess it. A focused inquiry might request original figure files, a clarification of overlapping text, a description of a data discrepancy or documentation of ethics oversight. Do not frame a request for explanation as a finding of misconduct.
Allow a reasonable opportunity to respond and consider whether coauthors need to be involved under policy. Authorship and institutional circumstances can be complex; an editor should not assume that the corresponding author can conclusively resolve every allegation involving the entire team.
Evaluate the substance and verifiability of the response. An incomplete response may justify a follow-up or escalation, but silence or delay should be interpreted in context and handled under the journal’s established process.
7. Handle plagiarism, text recycling and overlapping publication
Similarity reports identify text overlap; they do not establish plagiarism. Examine the source, extent, location, attribution and scholarly significance of matching material. References, standard methods language, legitimate quotation and properly disclosed reuse require different treatment from unattributed copying of original ideas or text.
Overlapping submissions and publications require comparison of study populations, datasets, analyses, conclusions and disclosure of related manuscripts. A permitted preprint should not be treated as duplicate publication merely because it contains the same study. Journal policy determines what prior dissemination is acceptable.
Request related documents when necessary, record the substantive overlap and determine whether correction, clarification, editorial rejection or formal escalation is appropriate. Avoid reducing complex publication-history questions to a single similarity percentage.
8. Evaluate data, statistical and image-integrity concerns
Potential problems include incompatible totals, unexplained changes across versions, implausible distributions, repeated or altered image elements and mismatches between reported methods and underlying evidence. Editors should identify what can be checked directly and what requires technical expertise.
Specialist consultation may be appropriate for image analysis, statistics or discipline-specific methods. Consultants should receive only the information needed, under suitable confidentiality conditions, and should be asked focused questions rather than invited to speculate about author motives.
Original data or image files may help resolve a concern, but the journal may lack the authority or capacity to conduct a full forensic investigation. Where the evidence remains materially unresolved, consider institutional involvement and protect the publication decision from premature certainty.
9. Address human research, consent and participant protection
Ethics approval statements, informed consent, privacy safeguards and participant protections should be assessed against the study’s context and applicable journal requirements. Missing documentation may be a reporting problem, but contradictions or credible concerns about actual participant protection require closer attention.
Requests for documentation should avoid collecting unnecessary personal or identifiable participant information. Editors should not contact participants directly or distribute sensitive records to reviewers as an informal verification strategy.
Where immediate privacy or safety risks are credible, consult authorized publisher leadership and appropriate expertise promptly. Any communication with an ethics committee or institution should be coordinated, accurate and limited to the information necessary for the concern.
10. Address animal research, clinical trials and registration discrepancies
For animal research, evaluate concerns about reported oversight, welfare safeguards and consistency between the approved and described procedures. Specialist input may be required to understand whether a discrepancy is substantive or a reporting omission.
For clinical trials, compare reported registration details and relevant timing with the manuscript’s claims where required by journal policy. Differences between registered and reported outcomes or unexplained changes may warrant clarification; a discrepancy alone is not a determination of wrongdoing.
Document the exact inconsistency and give authors an opportunity to explain. When an issue falls beyond editorial verification, consider referral through the appropriate institutional or regulatory route without overstating the evidence.
11. Manage authorship, contributions and competing-interest disputes
Authorship disputes can involve contribution, order, consent to submission or changes during review. Editors should seek relevant statements and documentation under the journal’s authorship policy, but should not attempt to adjudicate private workplace disputes from incomplete evidence.
Where authors disagree materially, an institution or other responsible body may be better positioned to establish contributions. Publication progression may need to pause until the dispute is resolved sufficiently to support an accurate authorship record.
Undisclosed financial, professional or personal interests should be evaluated for relevance to the manuscript and editorial process. An omission may require correction or clarification; it should not automatically be treated as proof that the research findings are unreliable.
12. Respond to suspected peer-review manipulation
Warning signs can include fabricated reviewer identities, unverifiable email addresses, unusual clusters of author-recommended reviewers, suspicious invitation responses or relationships that were not disclosed. Treat these as signals for verification rather than conclusive evidence of manipulation.
Preserve invitation records, contact details and relevant correspondence. Independently verify identities and close access to questionable accounts when necessary. Protect legitimate reviewers and authors from unnecessary disclosure while the issue is assessed.
If manipulation is substantiated or remains a serious unresolved concern, follow the publisher’s escalation process and assess whether affected review reports can be relied upon. A replacement independent review process may be necessary, but it should not be used to conceal an unresolved integrity issue.
13. Handle AI-related concerns through evidence and policy
AI-assisted writing, analysis, image generation or automated review can raise distinct issues involving disclosure, attribution, confidentiality, accuracy and research integrity. Apply the journal’s policy and focus on what was done and whether the resulting manuscript is reliable and compliant.
Writing style and automated AI-detection scores are not reliable standalone evidence of undisclosed AI use. Request specific clarification when there is a concrete inconsistency, fabricated reference, unverifiable claim or disclosure issue that matters to the publication decision.
Editors and reviewers must also respect confidentiality when using AI tools in the inquiry itself. Do not upload unpublished manuscripts, reviewer reports or allegations to unapproved services merely to accelerate screening.
14. Coordinate with institutions and other journals when warranted
Some questions—such as access to primary laboratory records, employment-related authorship disputes or formal findings about research conduct—may require an institution’s authority. Contact should follow the publisher’s established process and should describe the evidence and requested clarification without asserting an unproven conclusion.
Cross-journal concerns may require coordination when related articles, duplicate publication or compromised review processes affect more than one record. Share only necessary information through authorized channels and respect applicable confidentiality and data-protection requirements.
Institutional responses may take time and may not resolve every publication question. The editor remains responsible for deciding what the journal can responsibly publish or retain, while distinguishing that decision from any formal finding of misconduct.
15. Choose interim actions and keep the process proportionate
Possible interim steps include requesting clarification, pausing peer review, restricting access to a compromised reviewer account, obtaining specialist advice or holding production. Each action should address a defined risk and be revisited as information develops.
A pause is not a sanction or proof of misconduct. Explain procedural delays to authors as far as confidentiality permits, avoid open-ended inactivity and establish a responsible person to monitor the case.
Do not use ordinary peer review to investigate allegations that require confidential integrity handling. Reviewers should not be asked to contact authors, institutions or third parties on the journal’s behalf.
16. Decide the manuscript’s editorial pathway after assessment
Once sufficient information is available, determine the publication-related outcome under applicable policy. Depending on the evidence and stage, this may involve resuming review, requesting a correction or explanation, declining a manuscript, or referring a matter for further authorized assessment.
State the basis of the editorial decision in terms of the evidence and journal requirements. An editor may determine that the available record does not support publication without claiming that misconduct has been proven.
Keep scientific assessment and integrity assessment analytically distinct. A scientifically weak paper is not necessarily misconduct, and an integrity concern may require action even when reviewers otherwise consider the work important.
17. Handle concerns discovered after publication
Post-publication concerns require an assessment of the published record and the potential effect on readers. Depending on the issue and applicable policy, a correction, expression of concern, retraction or no public notice may be appropriate after fair assessment.
The purpose of a public notice is to maintain an accurate and intelligible scholarly record. Notices should identify what is known and why the action is being taken without unnecessary personal accusations or disclosure of confidential material.
Where an investigation remains incomplete but readers face a material unresolved reliability concern, the journal should consider its established procedures for interim public communication. Any eventual action should remain linked to evidence and policy.
18. Communicate fairly and protect all parties
Correspondence with authors, reviewers, complainants and institutions should be accurate, neutral and appropriately limited. Avoid promising confidentiality beyond what the process can provide, and explain procedural expectations where possible.
People who raise concerns should not be dismissed solely because their report is uncomfortable, anonymous or critical. At the same time, authors and other affected parties should have a fair opportunity to respond to relevant allegations through the authorized process.
Separate confidential internal notes from external correspondence. Use professional language suitable for later procedural review, and disclose information only to those with a legitimate role.
19. Close the case, preserve records and learn from failures
At closure, record the concern, the evidence considered, responses received, specialist advice, relevant conflicts, decisions and communications. The record should permit an authorized reviewer to understand how the outcome was reached without reconstructing events from memory.
Apply the journal’s retention, access and data-protection requirements. Do not keep unnecessary sensitive material or circulate case files informally. Where the matter has been referred to an institution, record the scope and status of that referral.
Use recurring patterns to improve screening, reviewer verification, author instructions and editorial training. Process improvement should address systemic weaknesses without turning case records into unsupported judgments about individuals.
20. A repeatable integrity-case decision framework
First identify the observable issue and preserve the relevant record. Next assess credibility, urgency, conflicts and the journal’s authority to act. Choose the narrowest appropriate inquiry, give affected parties a fair opportunity to respond, and obtain specialist or institutional input when the question exceeds editorial expertise.
Maintain a clear distinction between suspicion, verified facts, unresolved uncertainty and formal findings made by authorized bodies. Document why the journal pauses, proceeds, declines or takes post-publication action, and ensure that the communication matches what the evidence supports.
A responsible integrity process is not defined by the severity of its outcome. It is defined by proportionality, independence, confidentiality, procedural fairness and protection of the scholarly record.