Impartiality and Conflicts of Interest in Peer Review

“Impartiality” gets defined generically in ethics, law, and journalism contexts, but in peer review it has a specific, operational meaning: a reviewer’s assessment of a manuscript or grant application must be based solely on the scientific or scholarly merit of the work — its methods, evidence, reasoning, and clarity — and not on who wrote it, where they work, what the result implies for the reviewer’s own research, or any relationship between the reviewer and the author. This page covers what that standard requires in practice, the specific conflicts that put it at risk, what COPE and journal norms actually say about disclosing versus declining, and the editor-side controls that exist to catch what individual disclosure misses. For the broader classification of conflicts of interest across research generally, see CASRAI’s Types of Conflict of Interest guide and the Conflict of Interest (COI) dictionary entry; this page narrows to the peer-review context specifically. Note that every disclosure control described here assumes reviewer anonymity holds; the ICLR 2026 OpenReview breach shows what happens to that assumption when the platform itself leaks reviewer identities.

What impartiality means in peer review

An impartial review judges the work, not the author. That distinction sounds obvious stated abstractly, but it fails in predictable, recurring ways: a reviewer rating a manuscript more favorably because it comes from a well-known lab (prestige bias), less favorably because a result contradicts the reviewer’s own published position (confirmation bias), or declining to flag a weak methods section because raising it might delay a paper that cites the reviewer’s own work. None of these require dishonesty or bad faith — impartiality failures are usually a matter of unmanaged proximity, not misconduct. That is why the standard journals and funders apply is not “did you act in bad faith” but “could a reasonable person, aware of the relevant facts, question whether your assessment was influenced by something other than the work itself.” See Appearance of a Conflict of Interest for how that standard is defined and applied across research contexts, including peer review.

Because a review recommendation directly shapes whether a paper is published, a grant is funded, or a competitor’s priority is established, peer review is one of the highest-stakes points in the research system for impartiality to fail quietly — the reviewer’s identity and reasoning are typically hidden from the author, and in many models from the public entirely, which is precisely why professional norms around disclosure and recusal carry as much weight as they do.

Types of conflict that undermine reviewer impartiality

Reviewer conflicts cluster into a few recurring categories. Most journal reviewer guidelines and COPE’s own guidance describe some version of this list:

• Financial. The reviewer holds equity, consulting income, employment, or other financial ties to the authors, their institution, or a company whose product or competing technology the manuscript concerns. This is the most straightforwardly disclosable category and the one journals ask about most explicitly.
• Personal and relational. Recent collaborators, current or former mentors and mentees, co-investigators on a joint grant, close personal friends, family members, or anyone the reviewer has a supervisory or reporting relationship with. COPE’s guidance specifically flags recent (commonly a two-to-three-year window in journal policies) mentor/mentee and close-collaborator relationships as grounds to decline outright rather than disclose-and-proceed.
• Institutional. Sharing an employer with an author, particularly in the same department or research group, even absent a direct working relationship — the concern is less about actual bias and more about the appearance problem, plus the practical risk that the reviewer isn’t genuinely independent of local incentives.
• Intellectual and competitive. The reviewer is working on the same research question, using the same dataset, or developing a competing method or product. This is the hardest category to self-police because it’s also exactly the situation that makes someone the most qualified reviewer available — subject-matter proximity and competitive conflict often come from the same source. It’s also the category most linked to the failure modes below: a reviewer with a competing project has both the expertise to evaluate a manuscript accurately and the incentive to delay or undermine it.
• Ideological and philosophical. Strong prior commitment to a theoretical framework, methodology, or policy position that the manuscript’s findings would undercut. Harder to name and rarely covered by disclosure forms, but real — for example, an entrenched public position against a methodology can make it difficult to give a fair reading to a paper that uses it well.

Why the appearance of a conflict matters as much as an actual one: peer review is largely confidential — authors typically cannot see who reviewed their work or verify a reviewer’s stated absence of conflicts. That opacity is exactly why the field’s operative standard is the appearance test, not a private good-faith judgment by the reviewer alone. A reviewer who is confident they can “set aside” a conflict and review fairly is still expected to disclose or decline, because the system has no independent way to confirm that confidence was warranted, and the credibility of the outcome depends on it being verifiably not in question.

COPE guidance: disclose, or decline — not silent self-assessment

COPE’s Ethical Guidelines for Peer Reviewers set out the operative standard journals build their own reviewer policies around: reviewers should declare any potential conflicts of interest — financial, personal, intellectual, professional, political, or religious — that could affect their impartiality, and should not agree to review a manuscript from someone they are currently employed alongside, or have been a recent mentor, mentee, close collaborator, or joint grant holder with. The guidance is explicit that this is a duty to disclose or step aside, not to quietly weigh the conflict and proceed if the reviewer judges themselves unaffected — the reviewer is not the party positioned to make that call, the editor is. COPE also directs reviewers who discover a conflict mid-review, or anything else that might compromise a fair and unbiased assessment, to notify the journal immediately rather than complete the review and stay silent. On general standards, COPE’s guidelines state reviewers should not allow their assessment to be influenced by a manuscript’s origin, or by the authors’ nationality, religious or political beliefs, gender, or other characteristics — language that maps onto the documented bias patterns discussed below. Funder-side peer review (grant and award review) applies the same disclose-or-recuse logic; see, for comparable operational detail from the funder side, agency guidance such as NIH’s management of conflict of interest in peer review. For CASRAI’s broader walkthrough of COPE’s Core Practices, see COPE Guidelines Explained; for the practical mechanics of declining an invitation you shouldn’t accept, see How to Decline a Peer Review Invitation, and for the reviewer role generally, see Peer Reviewer: Role, Responsibilities, and How to Become One.

Editor-side controls

Individual disclosure is necessary but not sufficient — journals build structural safeguards around it because self-disclosure alone depends on reviewers correctly recognizing their own conflicts, which is exactly the judgment the appearance standard exists to not rely on:

• Reviewer selection. Editors are expected to screen for obvious institutional and collaboration overlap before inviting a reviewer, using citation history, co-authorship records, and shared-affiliation checks — not just wait for the reviewer to self-report.
• Author-requested exclusions. Most journals let authors name specific individuals they’d prefer not be asked to review — typically for genuine, nameable conflicts — and editors generally honor these within limits. The limits matter: journals do not let authors exclude entire competing labs or effectively hand-pick the reviewer pool, and editors retain final discretion over who reviews a submission regardless of an author’s request.
• Submissions from editorial board members. When a board member or the editor-in-chief submits their own work, standard practice routes the manuscript to a different, independent editor — often a deputy or associate editor — who handles the entire process, including reviewer selection and the final decision, without the submitting editor’s involvement. See CASRAI’s Editorial Board Responsibilities guide for how this fits into the board member’s broader duties, and the Editorial Independence dictionary entry for the underlying principle.
• When the editor has a conflict. The same logic applies one level up: an editor with a personal, financial, or institutional conflict on a specific manuscript is expected to recuse themselves from handling that submission and hand it to a co-editor or a member of the editorial board instead, rather than merely disclosing the conflict and proceeding.

Review models and what they do to impartiality

The identity-disclosure model a journal uses changes which conflicts are easiest to catch and which are hardest to prevent. CASRAI covers the mechanics of each model in detail in Single-anonymized vs. double-anonymized vs. open peer review; briefly, as it bears on impartiality specifically:

• Single-anonymized (traditional “single-blind”). The reviewer knows the author’s identity, the author does not know the reviewer’s. This leaves prestige and affiliation bias fully exposed — the reviewer can be influenced by a recognizable name or institution before reading a single result.
• Double-anonymized (“double-blind”). Both parties’ identities are withheld. Evidence on its effect is mixed but generally favorable for reducing prestige and affiliation bias — several studies of journals that adopted double-anonymized review found shifts in acceptance patterns consistent with reduced favoritism toward well-known authors and institutions. Its limitation is well documented: self-citation, distinctive datasets, preprint availability, and narrow subfields routinely make authorship identifiable anyway, especially in small research communities where a handful of labs work on a given problem. Anonymization also does nothing for competitive or intellectual conflicts — a reviewer who recognizes the work as a rival’s project doesn’t need the author’s name to have a stake in the outcome.
• Open review. Reviewer (and sometimes report) identities are disclosed, often alongside publication. Proponents argue visibility disincentivizes unfair or careless review because the reviewer’s own name is attached to the judgment; critics note it can make reviewers, especially early-career ones, reluctant to criticize senior authors they may depend on later — a different impartiality risk, not an absence of one.
• Post-publication review. Commentary and critique occur openly after publication, sometimes under real names, sometimes pseudonymously depending on the platform. It doesn’t replace pre-publication gatekeeping but adds an ongoing, less time-pressured check that can surface conflicts or errors the original review missed.

No model eliminates the conflicts described above — they redistribute which ones are easiest to detect and which are structurally invisible to the process.

Documented biases in peer review

Beyond named conflicts of interest, research on peer review has documented several systematic bias patterns that operate even without any specific relationship between reviewer and author:

• Prestige and affiliation bias. Manuscripts from well-known authors or highly ranked institutions have been shown in multiple studies to receive more favorable treatment than blinded evaluation of the same work would predict.
• Nationality and language bias. Non-native-English manuscripts and submissions from researchers outside a small set of dominant research economies have been associated with harsher review outcomes independent of scientific quality, a pattern that has driven journal-level interest in equitable review practices; see CASRAI’s Equitable Peer Review dictionary entry.
• Gender bias. Studies using name-based and blinded comparisons have found evidence of differential outcomes correlated with perceived author gender in some fields and review contexts, though effect sizes and consistency vary across the literature.
• Confirmation bias against null and unexpected results. Reviewers, like researchers generally, tend to scrutinize surprising or null findings more skeptically than results that confirm existing expectations — a pattern that contributes to publication bias in the literature even when no single review decision looks unreasonable in isolation.

Failure modes: appropriation, delay, and coercive citation

Some impartiality failures go beyond biased judgment into outright misconduct:

• Idea and data appropriation. A reviewer using confidential information from a manuscript under review — an unpublished method, dataset, or hypothesis — in their own work before the original is published is both a breach of the confidentiality every reviewer agrees to and a research-integrity offense in its own right, not merely an impartiality lapse.
• Deliberate delay. A reviewer with a competing paper or grant application in progress who slow-walks a review to gain a priority advantage is exploiting exactly the intellectual/competitive conflict described above. Editors monitoring review timelines and enforcing deadlines is a partial safeguard, but the underlying incentive is structural.
• Coercive citation. A reviewer or editor requiring authors to add citations — typically to the reviewer’s or journal’s own work — as an implicit or explicit condition of a favorable recommendation. See CASRAI’s Coercive Citation dictionary entry for how this is defined and how publishers and indexers have responded to it; it is distinct from a legitimate suggestion that relevant, missing literature be cited, which is a normal and appropriate part of review.

What to do when impartiality fails

Authors who believe a review was compromised by an undisclosed conflict or evident bias have a defined escalation path, not an informal one:

• Author appeal to the handling editor first. The initial route is a direct, evidence-based appeal — specifying what in the review suggests bias or an undisclosed conflict, not simply disagreement with the outcome. See CASRAI’s How to Write an Appeal Letter to a Journal Editor for the mechanics of doing this well.
• Formal journal complaint. If the handling editor’s response doesn’t resolve the concern, most journals and publishers maintain a separate complaints process, often routed to the editor-in-chief or a publication-ethics committee independent of the original editorial handling.
• COPE flowcharts. For cases journals are themselves handling — for example, a suspected conflict discovered after acceptance, or a dispute over reviewer conduct — COPE publishes discussion documents and flowcharts editors use to work through publication-ethics scenarios systematically rather than case-by-case improvisation; see COPE Guidelines Explained for how these fit into COPE’s broader framework.
• Institutional escalation. Where the concern rises to research misconduct — appropriated ideas or data, or a pattern of conduct rather than an isolated lapse — the appropriate route shifts to the reviewer’s or author’s home institution and its research integrity office, alongside the journal process rather than instead of it.

None of these routes are fast, and appeal outcomes are not guaranteed — which is precisely why the preventive side (disclosure norms, editor-side screening, and review-model design) carries more practical weight than the remedial side. See also Peer Review Fraud for the more severe end of this spectrum, where impartiality failure crosses into fabricated or manipulated review altogether.

Frequently asked questions

What’s the difference between a conflict of interest and bias in peer review?

A conflict of interest is a situation or relationship that creates the risk of biased judgment — it’s structural and identifiable (a financial tie, a recent collaboration). Bias is the actual or perceived skew in the judgment itself, which can result from an unmanaged conflict or from unconscious patterns like prestige or confirmation bias that don’t trace back to any single disclosable relationship.

Should a reviewer disclose a minor conflict or just decline?

COPE’s guidance points toward declining for the categories it names explicitly — current shared employment, and recent mentor, mentee, close-collaborator, or joint-grant relationships — rather than disclosing and proceeding. For conflicts outside that list, disclosure to the editor is the minimum step, and the editor, not the reviewer, decides whether the review can proceed, be reassigned, or be weighted accordingly.

Does double-anonymized review solve reviewer impartiality problems?

No. It reduces certain biases tied to recognizing an author’s name or institution, but self-citation, distinctive datasets, and narrow-subfield familiarity routinely defeat anonymization in practice, and it does nothing for competitive, intellectual, or ideological conflicts where the reviewer doesn’t need to know who wrote the manuscript to have a stake in its outcome.

Can authors choose their own reviewers?

Authors can typically suggest reviewers and request exclusions, but editors retain final discretion over who is actually invited and are not bound to use suggested reviewers or honor every exclusion request — the point of editorial control over reviewer selection is precisely to prevent authors from curating their own review outcome.