Integrity overview
Protect the scholarly record without turning suspicion into a finding
Research integrity is embedded across submission, peer review, editorial assessment, publication, and post-publication maintenance. Treata Scholars uses a framework built around evidence, proportionality, confidentiality, appropriate escalation, and clear separation between editorial authority and formal institutional investigation.
The sections below explain how that framework applies to authorship, conflicts, data, images, text overlap, research ethics, peer review, AI, third-party concerns, institutional coordination, and post-publication action.
1. Research integrity as a publishing responsibility
Research integrity connects responsible research conduct, transparent reporting, fair editorial assessment, and the long-term reliability of the scholarly record. For a publisher, integrity is not limited to reacting to allegations; it is built into submission requirements, peer review, editorial decision-making, publication, correction, and post-publication oversight.
Treata Scholars approaches integrity through evidence, proportionate process, confidentiality, and clear responsibility. The purpose is to determine what the journal can reliably publish or continue to stand behind, not to convert editorial procedures into a substitute for institutional investigation.
Integrity processes should protect both the scholarly record and the fairness owed to people whose work or conduct is questioned.
2. The boundary between publisher and institutional authority
A journal can assess manuscripts, publication records, declarations, reviewer activity, submitted files, and information relevant to editorial decisions. It can request clarification and determine whether the available evidence is sufficient for publication or post-publication action.
Institutions, employers, ethics bodies, regulators, funders, or other authorized organizations may have access to laboratory records, employment information, participant records, or formal investigative powers that a publisher does not possess.
Treata Scholars therefore distinguishes editorial assessment from formal findings of research misconduct. Where a question exceeds journal authority, the appropriate response may include referral while the journal separately determines what action is needed to protect the publication record.
3. Evidence before conclusions
Similarity reports, image-analysis signals, statistical anomalies, automated flags, reviewer suspicions, or third-party allegations can identify questions that deserve attention, but they are not conclusions by themselves.
The meaning of a signal depends on context. Text overlap may reflect legitimate methods language; an image similarity may arise from a documented experimental design; an unusual statistical pattern may have a valid explanation. Editors should identify the observable issue before assigning a cause.
Integrity decisions should distinguish verified facts, unresolved uncertainty, interpretation, and allegation. This distinction should remain visible in internal records and external communication.
4. Fair and proportionate process
The response to a concern should match its seriousness, specificity, and potential effect on the scholarly record. A missing declaration, a disputed interpretation, a duplicated figure panel, and a credible allegation of fabricated data require different levels of inquiry.
Relevant parties should normally have an appropriate opportunity to respond to the specific issue being assessed. Requests should focus on information needed to resolve the publishing question rather than demanding broad personal or institutional records without a clear reason.
Fair process also requires conflict management. An editor whose independence could reasonably be questioned should not control the assessment of the concern.
5. Authorship and contributorship integrity
Authorship should reflect genuine scholarly responsibility and should be represented accurately. Concerns may involve omitted contributors, inappropriate inclusion, disputed order, submission without a contributor’s knowledge, or authorship changes during review.
Editors can review declarations, contribution statements, correspondence, and relevant documentation, but they may be unable to resolve private employment or laboratory disputes from the journal record alone.
Where contributors cannot reach agreement and the issue materially affects the publication record, institutional clarification may be required before the journal can proceed responsibly.
6. Competing interests and undisclosed relationships
Authors, reviewers, and editors should disclose interests relevant to their role in the publication process. These may include financial relationships, professional collaborations, institutional connections, personal relationships, intellectual commitments, or other circumstances that could reasonably affect independence.
A missing disclosure is not automatically evidence that the research itself is unreliable. The journal should assess the nature of the interest, whether it affected review or decision-making, and whether disclosure, correction, recusal, reassessment, or another response is appropriate.
Conflict management is intended to protect trust and decision quality rather than to imply wrongdoing whenever a relationship exists.
7. Data integrity and reliability
Editors may encounter unexplained inconsistencies in datasets, totals, reported sample sizes, analyses, tables, or results across manuscript versions. The first task is to identify the discrepancy clearly and determine whether it can be resolved through clarification or corrected analysis.
Where raw or supporting data are requested, the request should be relevant to the concern and should respect privacy, consent, contractual restrictions, and applicable data-governance requirements.
If the journal cannot establish that central claims are supported, it may need to pause or decline publication even without making a formal finding about why the data problem occurred.
8. Image integrity
Images can raise concerns involving duplication, inappropriate alteration, splicing, selective presentation, mislabeled panels, or inconsistencies between figure versions. Automated tools may help identify patterns, but expert assessment and contextual review remain necessary.
Editors should preserve the submitted files, identify the specific panels or features at issue, and request original or supporting material when proportionate and appropriate.
A correction may be sufficient for an honest and limited presentation error that does not affect the conclusions; broader concerns about reliability may require specialist input, institutional involvement, or stronger editorial action.
9. Plagiarism, text reuse, and publication overlap
Text-similarity percentages should not be treated as automatic plagiarism findings. Editors should evaluate the source, amount, location, attribution, scholarly significance, and whether the overlapping material represents original ideas, standard methods language, quotation, or previously published work.
Related manuscripts should be assessed for overlap in datasets, participants, analyses, conclusions, and disclosure. Legitimate preprints or secondary analyses should be distinguished from duplicate or redundant publication.
The appropriate response depends on the nature of the overlap and the stage of publication, ranging from clarification or citation correction to rejection, correction, or formal post-publication action.
10. Human and animal research ethics
Research involving people or animals should be reported with the ethical information required by the journal and appropriate to the study. Missing statements may be correctable reporting omissions, while contradictions or credible concerns about actual oversight may require deeper assessment.
Editors should avoid collecting unnecessary sensitive information and should not contact participants directly as an informal verification method. Where documentation is needed, the request should be narrowly linked to the publishing concern.
When a question exceeds editorial expertise or authority, consultation with appropriate specialist or institutional bodies may be necessary before publication can proceed.
11. Clinical trials, registration, and protocol consistency
Where registration is required or relevant, editors should consider whether the manuscript’s reported design, outcomes, and timing are consistent with the available registration information and journal requirements.
Differences between registered and reported outcomes can have several explanations and should be clarified rather than immediately characterized as misconduct. Material unexplained changes may affect the reliability or transparency of the report.
Protocol, registration, and reporting information should be assessed as part of the overall evidentiary record, particularly when conclusions depend on selectively reported outcomes.
12. Peer-review integrity
The integrity of peer review can be compromised by fabricated reviewer identities, undisclosed relationships, manipulated contact details, inappropriate influence, breaches of confidentiality, coordinated reviewing, or misuse of privileged manuscript information.
Editors should independently verify reviewer identity and suitability where necessary, preserve invitation and correspondence records, and investigate unusual patterns proportionately rather than relying on author-supplied information alone.
If a review process is found to be unreliable, the journal may need to disregard affected reports, re-establish independent review, and consider whether the matter requires broader escalation.
13. Artificial intelligence and integrity
AI-assisted tools can create integrity questions involving disclosure, confidentiality, fabricated references, image generation, data analysis, attribution, and the reliability of generated content. The appropriate assessment depends on what the tool was used for and which journal requirements apply.
Automated AI-detection scores or stylistic impressions should not be treated as standalone proof of undisclosed use. Editors should focus on specific, verifiable problems such as false citations, inconsistent data, undisclosed image generation, or breaches of confidentiality.
Integrity assessment itself should not introduce new confidentiality risks by uploading unpublished manuscripts or sensitive case materials into unapproved external systems.
14. Anonymous and third-party concerns
Concerns may be raised by readers, reviewers, researchers, institutions, whistleblowers, or anonymous correspondents. The identity or status of the person raising the issue should not determine whether the evidence deserves assessment.
Anonymous reports can be difficult to clarify, but they should be judged on specificity, verifiability, and relevance. Named complaints are not automatically more accurate, and anonymity is not evidence of bad faith.
The journal should preserve appropriate confidentiality while avoiding promises that cannot be maintained if limited disclosure becomes necessary for a fair and effective process.
15. Communication during integrity assessment
Correspondence should describe concerns precisely and avoid declaring misconduct before a competent body has made such a finding. Neutral language helps preserve procedural fairness and reduces the risk that an editorial inquiry becomes accusatory.
Authors and other relevant parties should understand what information is being requested and why, subject to confidentiality limits. Where delays occur, the journal should provide procedural updates when possible without revealing protected details.
Internal notes should remain factual and professional because they may later need to support an appeal, institutional referral, or post-publication decision.
16. Institutional and external coordination
Some matters require information or authority held by a research institution, ethics body, regulator, funder, or other organization. Referral should describe the concern and evidence accurately without presenting unresolved allegations as established facts.
Parallel processes should be coordinated carefully. The journal may need to decide whether a manuscript can be published while an external inquiry continues, and an institutional investigation may not answer every publication-specific question.
Treata Scholars retains responsibility for the journal’s publication record even when another body conducts a separate investigation into research conduct.
17. Post-publication stewardship
Research-integrity responsibilities continue after publication. New evidence can reveal errors, unreliable results, missing disclosures, authorship problems, or concerns that were not visible during peer review.
The journal should assess the effect on the published record and select a response appropriate to the evidence. Possible actions can include correction, clarification, expression of concern, retraction, or no formal notice where the concern is not substantiated.
Public notices should explain the publication-related reason for the action and should remain linked to the affected work so that readers can understand the current status of the record.
18. Corrections, expressions of concern, and retractions
These mechanisms serve different purposes. A correction addresses a limited error when the work remains substantially reliable. An expression of concern can alert readers to a material unresolved issue while assessment continues. A retraction indicates that the work should no longer be relied upon in its published form.
The choice should depend on the reliability of the article and the evidence available, not on whether the situation is embarrassing or reputationally difficult.
Retraction and correction notices should not be designed as punishment. Their primary function is to maintain an accurate and intelligible scholarly record.
19. Documentation, retention, and auditability
Integrity cases should have a coherent record containing the concern, evidence considered, relevant manuscript versions, correspondence, specialist advice, conflicts, decisions, referrals, and publication actions.
Records should distinguish verified observations from allegations and should be stored in approved systems with access limited to people who have a legitimate role in the case.
Good documentation supports consistency, appeals, institutional communication, training, and later review without requiring staff to reconstruct events from memory.
20. A repeatable research-integrity framework
When a concern arises, first identify the observable issue and preserve the relevant record. Assess credibility, seriousness, conflicts, and the journal’s authority. Request focused clarification, obtain specialist input when needed, and escalate questions that exceed editorial authority.
Maintain a clear distinction between suspicion, evidence, unresolved uncertainty, and formal findings. Decide what the journal can responsibly publish or maintain in the literature and document the reasoning behind that decision.
The standard is not to eliminate every uncertainty. It is to use a fair, proportionate, confidential, evidence-based process that protects the reliability and intelligibility of the scholarly record.
Related guidance
Continue through the Treata integrity framework
Research Misconduct Policy → Handling Research Integrity Cases →