This policy is incomplete. You can help by suggesting content and changes directly via pull request.
-
Scholarly Peer Review refers to one or more other scholars providing feedback on a (usually) pre-published work, including presentation, methods, and findings.
-
A Peer Reviewer is an individual without conflict of interest and is a knowledgeable scholar in the field of a work who provides constructive, objective feedback on the (usually) pre-published work.
-
A Review Manager is an individual who assigns conference or journal submissions for review, oversees the reviewing process, and informs authors of decisions, typically a Program Chair, Associate Editor or Editor in Chief.
-
Reviewer Misconduct refers to participation in the peer review process, either as a peer reviewer or review manager, that produces, by action or inaction, adverse outcomes to relevant parties, including the authors, other reviewers, the publication venue (journal, conference, etc.), or scientific knowledge, as a whole. Examples of reviewer misconduct include inappropriate review content, rejection or acceptance of works when unjustified, or other inappropriate behaviours such as breach of confidentiality, impartiality, ethical standards or practices.
-
Submitting a work for peer review refers to the authors of a work following a defined submission process for a scholarly venue (e.g. conference, journal, edited book) overseen by one or more review managers, to seek peer review of and eventual publication of the reviewed work in the venue, by one or more peer reviewers.
- Avoid conflicts of interest. - The need to adhere to the venue's documented Conflict of Interest policies, the publisher or publishing society's overarching CoI policies, or declining to peer review a work that an objective peer review would not be possible
- Avoid discriminating against a protected group. The need to provide an objective peer review without discriminating based on author, institution and/or country characteristics, including but not limited to gender, age, ethnicity, language, culture, sexual orientation, polical affiliation, religious beliefs, physical or mental disability, cognitive difference, appearance, socio-economic status, educational background, ...
- Adhere to required guidelines and practices. The need to follow a venue's published, whether publically or internal to peer reviewers and review managers, guidelines and standards for peer review
- Do not submit fake reviews. Using AI to generate all or part of a review or submitting a review that deliberately mis-states the reviewer's true opinion of the work, usually in return for payment or other unacceptable benefits to the peer reviewer and/or as a favour to the authors or review manager, thus biasing the peer review process in favour (or against) the work.
- Inappropriate management of the peer review process. A review manager failing in their duty to implement the documented venue peer review process, Conflict of Interest requirements, discriminatory practices, or allowing a manifestly improper peer reviewer behaviour.
- Using a large language model (e.g. ChatGPT) or other tool to generate some or all of a review
- Using the review to promote work by the reviewer or their close colleagues, e.g., by demanding that authors add citations to one or more papers written by the reviewer or their close colleagues
- Failing to declare a conflict of interest and reviewing the submission anyway
- Reviewing a submission when knowingly unable to be sufficiently objective in conducting the review
- Prejudice dressed up as external validity concerns; e.g., dismissing a study because participants came from a developing country.
- Distributing the submission to anyone not explicitly approved, except where a supervisor completes a review with a student for pedagogical reasons.
- Failing to maintain confidentiality e.g., by sharing the unpublished manuscript on social media
- Submitting a review without reading the submission
- Submitting a manifestly inadequate review e.g. 1-2 lines; failing to address most review criteria
- Stealing ideas from a submission and rejecting it so the reviewer can publish a similar paper first
- Intentionally un-masking another anonymous reviewer to the authors
- Ad hominem attacks on the authors
- Accusing authors of misconduct (plagiarism, data fabrication, p-hacking, etc.) without compelling evidence
- Rejecting sound research to reduce acceptance rates (typically to increase perceived prestige of the venue)
- Deliberately praising the work and arguing for acceptance despite clear, serious flaws (as a favour to the authors)
- Rejecting sound research because results are non-significant
- Rejecting sound research because the conclusions contradict the reviewers' unjustified beliefs
- Submitting a review not at all related to the content of a submission e.g. copying verbatim from another review
- Completing fake reviews of a submission for payment or other material gain e.g. as a favour to the authors or review manager
- Deliberately assigning submissions to reviewers known to be overly hostile or overly favourable toward the authors or the submission topic
- Failing to treat genuine author concerns about a review with respect; failing to investigate a review complaint
- Procuring fake reviews (e.g. generated by a large language model) for a manuscript, whether favourable or unfavourable
- Suggesting the authors look at papers not written by the reviewer or close colleagues, even if the authors perceive the papers as irrelevant
- Publicly sharing thoughts about a manuscript under review without disclosing important elements of the manuscript; e.g., "Reviewing yet another manuscript that reports "experiments" with no control groups - so frustrating!"
- Failing to submit a review on time or at all
- Misunderstanding or deviating from a venue's peer review standards or procedures
- Missing important elements of a paper, skimming sections, failing to engage with a paper as deeply as one should
- Poorly explaining reasons for rejection
- Writing a review that is perceived as unduly rude or blunt by the authors
- Refusing to review a revision of a paper previously reviewed
- Revealing the reviewer's own identity, intentionally or accidentally, during or after review
It is the responsibility of every venue for scholarly publishing to:
- have a clear, publically available procedure for receiving, investigating and resolving complaints of misconduct against reviews, managers, and editors.
- keep records of complaints and their resolutions.
- overturn review decisions found to be improper.
Imagine a "line of misconduct." On the left side of the line are reviews that the authors perceive as rude, dismissive, unhelpful, or unreasonable. On the right side of the line are reviews that any reasonable person would find unduly hostile and destructive in tone and substance. This line is difficult to pin down. Different reviewers (and cultures) express themselves in different ways. Some authors are more sensitive; others more thick-skinned. We don't want to criminalize cultural differences. We want people to be respectful but we don't want to encourage toxic positivity or being so indirect that no one knows what anyone means. And yet, there exist reviews that are plainly over the line. Moving toward structured, standards-based reviews and away from unstructured free-text reviews may mitigate this problem.