Why Research Integrity Cannot Depend Only on Journal Editors

Getting your Trinity Audio player ready...

Research integrity cannot depend only on journal editors because editors see only one stage of a much larger research process. They generally evaluate manuscripts after experiments have been designed, data have been collected, analyses have been selected, and authors have decided what to report.

Editors remain important, but they cannot independently verify every dataset, reproduce every calculation, inspect every laboratory procedure, or monitor a published paper indefinitely. A reliable scientific system therefore needs distributed, continuing, and transparent scrutiny, rather than a single editorial checkpoint.

Journal Editors Are Gatekeepers, Not Universal Auditors

Journal editors typically decide whether a manuscript fits a journal, whether it should be sent for peer review, which reviewers should assess it, and whether revisions are sufficient for publication.

These are consequential responsibilities. The Committee on Publication Ethics recommends that journals maintain clear policies covering peer review, conflicts of interest, complaints, corrections, and research misconduct. However, such policies do not transform editors into investigators with unlimited expertise, evidence, time, or enforcement power.

Editors ordinarily rely on information supplied by authors and reviewers. Even a conscientious editor may not have access to:

• complete raw data and laboratory records;
• undocumented analytical choices;
• source code and computational environments;
• private communications among collaborators;
• evidence held by universities or funders;
• specialist expertise covering every claim in a multidisciplinary paper.

Publication decisions should therefore be treated as informed assessments—not permanent certifications that every claim is correct.

Peer Review Has Structural Limits

Peer review can identify unclear arguments, methodological weaknesses, missing references, unsupported interpretations, and some statistical errors. It is an essential quality-control mechanism, but it is not equivalent to replication or forensic investigation.

Reviewers normally work with limited time and incomplete access to the research process. They may also have conflicts of interest, intellectual commitments, or direct competition with the authors. The National Academies notes that peer review must balance impartiality against the fact that the most qualified reviewers are often researchers working in closely related areas.

A manuscript can therefore pass peer review while still containing:

• honest mistakes;
• fragile statistical conclusions;
• irreproducible computational results;
• omitted experimental details;
• exaggerated causal claims;
• manipulated or fabricated evidence that reviewers cannot detect.

Conversely, an unconventional but valid paper may be rejected because reviewers misunderstand it, consider it too risky, or evaluate it through the assumptions of an established school.

Research integrity needs peer review, but it should not confuse peer review with proof.

Editors Face Institutional Conflicts

Editors do not operate outside the publishing economy. Journals compete for prestige, citations, influential authors, institutional subscriptions, publication fees, and public attention.

These incentives do not make editors dishonest. They do mean that editorial judgment occurs inside an institution with its own interests.

A journal may be reluctant to investigate a highly cited paper that strengthened its reputation. Editors may also face legal uncertainty, incomplete cooperation from authors’ institutions, or disagreement over whether a problem justifies a correction, expression of concern, or retraction.

COPE guidance recognizes the need for recusal and independent handling when editors themselves have conflicts—for example, when an editor submits work to the same journal. The broader lesson is that no integrity system should assume that one office can always investigate itself impartially.

Research Integrity Begins Before Submission

Many threats to integrity arise long before a journal receives a manuscript:

• poor supervision;
• inadequate recordkeeping;
• undisclosed conflicts;
• selective reporting;
• inappropriate statistical methods;
• pressure to produce positive results;
• inaccessible data or code;
• authorship disputes;
• institutional retaliation against criticism.

Universities and research organizations are better positioned than journals to examine laboratory records, interview personnel, audit equipment, and determine whether institutional procedures were followed.

Funders can require preregistration, data-management plans, open outputs, independent audits, and reporting of negative findings. Professional associations can establish disciplinary standards. Researchers can preserve records and disclose limitations. Software repositories and data archives can make important parts of the work inspectable.

The National Academies consequently treats integrity as a collective responsibility grounded in honesty, openness, fairness, accountability, objectivity, and stewardship—not merely as an editorial function.

Publication Must Begin Scrutiny, Not End It

A conventional publication decision is concentrated in time: a small number of reviewers examine one manuscript before publication. Scientific knowledge, however, develops over years.

Problems may become visible only when:

• another laboratory attempts replication;
• researchers apply the method to new data;
• readers inspect the source code;
• a specialist notices an image or equation problem;
• later evidence contradicts the conclusion;
• whistleblowers provide previously unavailable records.

Research integrity therefore requires post-publication review. Critiques, replications, corrections, updated datasets, alternative analyses, and formal responses should remain connected to the original work.

A paper should not retain an unchanged appearance of authority after substantial evidence against it has emerged. At the same time, criticism should be attributable, evidence-based, and open to challenge. Transparent post-publication systems can make both the original claim and its subsequent evaluation visible.

This principle is compatible with the World Science DAO proposal for a journal using post-publication moderation, where publication and continuing evaluation are treated as distinct processes rather than compressed into one editorial decision.

Independent Reviewers Can Operate Without Journals

Scientific review does not logically require a journal. Qualified researchers can publish signed reviews of preprints, datasets, proofs, software, replications, and self-published manuscripts.

Within the AI Internet-Meritocracy model, some of these independent reviewers may operate as science marketers: specialists who examine useful research, explain its importance, disclose weaknesses, and place their own reputations behind signed assessments.

In this limited sense, science marketers can replace part of the journal’s reviewing and discovery role. Instead of one journal privately selecting reviewers, multiple independent evaluators could publish competing assessments that remain visible and attributable. The reviewer’s historical accuracy and judgment could then become part of their reputation.

The broader AIIM concept of independently reviewed and promoted scientific work should not mean that promotion substitutes for verification. A science marketer who reviews research must disclose conflicts, distinguish technical assessment from advocacy, cite evidence, and accept public scrutiny of the review.

This is not a proposal to give every reviewer automatic authority. It is a proposal to make reviewing open, plural, attributable, and separable from journal ownership.

A Distributed Integrity System

A stronger research-integrity architecture assigns different responsibilities to different participants:

Participant	Primary responsibility
Researchers	Preserve evidence, report methods honestly, disclose uncertainty and correct mistakes
Universities	Investigate conduct and examine internal records
Funders	Set transparency requirements and monitor funded work
Reviewers	Evaluate claims within their competence and disclose conflicts
Replicators	Test whether findings survive independent repetition
Data and software auditors	Check evidence, calculations, and reproducibility
Journal editors	Coordinate editorial assessment and correct the published record
Independent reviewers and science marketers	Publish attributable evaluations outside journal control
Readers and scientific communities	Challenge claims and update collective judgment

No component is sufficient by itself. The objective is not to weaken editors, but to prevent the scientific record from depending on the capacity, incentives, or judgment of a small editorial group.

Research Integrity Requires Checks and Competition

Scientific knowledge becomes more trustworthy when claims can be evaluated through several independent channels. A journal’s acceptance should be one positive signal. Successful replication, open data, verified code, signed expert reviews, methodological audits, and continued usefulness should provide additional signals.

Likewise, one editor’s rejection should not permanently suppress a work. Researchers should be able to publish openly and seek independent evaluation elsewhere.

The central principle is straightforward:

Research integrity is a property of an accountable scientific ecosystem, not a service that journal editors can provide alone.

Editors should remain participants in that ecosystem. They should not be its sole judges, investigators, or custodians. Distributed review, institutional accountability, open evidence, post-publication criticism, and independent evaluators can create a system that corrects mistakes more reliably while reducing the power of any single gatekeeper.