Are all journal articles peer-reviewed?

No. ICMJE says a peer-reviewed journal does not have to send every submission out for review. Nature Portfolio reviews Articles, Letters, and several other types, but Correspondence and many corrections are usually not refereed. Always check the article type on the PDF and the journal's stated policy for that type.

Does "peer-reviewed" mean the findings are correct?

No. Berkeley's Understanding Science page states the limit directly: peer-reviewed work is not necessarily correct or conclusive. Nature Portfolio's quoted Nature Immunology editorial adds that reviewers cannot see data the authors omitted and are not expected to catch clever fraud. Read the methods and the limitations.

What is the difference between peer-reviewed and scholarly?

Scholarly describes the audience and the apparatus (experts writing for experts, with citations). Peer-reviewed describes a specific pre-publication process. DOAJ allows some arts and humanities journals to use two-editor editorial review instead of external referees. A university-press book can be scholarly without being a refereed article.

Is every peer-reviewed journal judging the same thing?

No. Nature Portfolio asks whether the work is technically valid and of significant interest. PLOS ONE asks whether it is technically rigorous and ethical, not whether it is important. A Registered Report judges the question and the methods before the results exist. "Peer-reviewed" names the process, not the editorial bar.

How long does journal peer review take?

There is no official global timeline. Nature Portfolio asks reviewers to meet an agreed number of days; ICMJE tells editors to reject promptly when they will not proceed. Berkeley's teaching page says the full path from submission to publication often takes more than a year. A journal that guarantees acceptance in days is failing Think. Check. Submit.

Is a preprint peer-reviewed?

Not by the journal that has not yet accepted it. ICMJE tells preprint archives to identify that work as not peer-reviewed. Cite the published version if it exists. If you must cite the preprint, label it as a preprint in the style your assignment names.

How is NIH grant review different from journal review?

NIH first-level review is a Scientific Review Group scoring a proposed project on a 1–9 scale; discussed applications receive an Overall Impact score from 10 to 90, then a council looks at mission relevance. For most RPGs due on or after 25 January 2025, reviewers score Importance and Rigor/Feasibility and mark Expertise/Resources as sufficient or not. A journal editor is deciding whether to publish a finished manuscript. A funded grant is not a peer-reviewed article.

Where can I read a real referee report?

On journals that publish transparent peer-review files. Nature Portfolio now posts that file with original research submitted on or after 16 June 2025. Nature Communications has done so for submissions since 1 November 2022. Luebber et al., Nature Communications 16, Article 9824 (2025), is one article whose page links a peer-review file. The file still omits confidential comments to the editor.

How do I spot a journal that only pretends to review?

Use ICMJE's test: the title accepts almost everything, charges a fee (often after a fake acceptance), claims review it does not do, and may copy a familiar journal's name. Then use Think. Check. Submit.: can you name the publisher, see the review model, see the fees, and find the journal in DOAJ or a database your library pays for?

Can I use a chatbot to peer-review my own draft?

That is not journal peer review. ICMJE Section V and Nature Portfolio both treat uploaded manuscripts as confidential and warn that AI output can be wrong, incomplete, or biased. Nature Portfolio requires any AI assistance in a real referee report to be declared. Use a human reader in your field, or a campus writing center, for a draft.

How Peer Review Works: Journal, Grant, and Predatory Checks [2026]

Quick Answer: What Peer Review Is

An editor screens a submission. If it fits the journal, the editor sends it to experts who are usually not on the editorial staff. Those reviewers write a critique. The editor then accepts, asks for revision, or rejects — and can ignore a reviewer's recommendation. A library filter that says "peer-reviewed" is a claim about the journal, not a guarantee that the particular item you opened (a letter, an editorial, a book review) was refereed. Selective journals also judge novelty and audience. Validity-only journals such as PLOS ONE ask whether the work is technically sound. Neither stamp makes the findings true.

ICMJE Recommendations (January 2026), §II.C.2.c

Official definition

Peer review is the critical assessment of manuscripts submitted to journals by experts who are usually not part of the editorial staff.

Who actually decides

A peer-reviewed journal is under no obligation to send submitted manuscripts for review, and under no obligation to follow reviewer recommendations, favorable or negative.

Student checklist before you cite "a peer-reviewed source"

Confirm the item type on the PDF: original research, systematic review, or methods paper — not a letter, news item, or invited commentary.
Open the journal's "About," "Instructions for authors," or "Editorial policies" page and read the stated review model (anonymized, open, or transparent).
Ask what the journal is judging. Nature screens for interest and advance; PLOS ONE screens for technical rigor. "Peer-reviewed" does not mean the same gate.
For biomedical titles, check whether the journal is currently indexed in MEDLINE. MEDLINE asks whether the peer-review process is "explicit and sufficiently detailed."
For open-access titles you do not recognize, run the Think. Check. Submit. journals checklist (English text last updated June 2023) instead of trusting a "peer-reviewed" badge on the homepage.
Do not treat a preprint on arXiv, bioRxiv, medRxiv, or SSRN as refereed. ICMJE §III.D.3.a tells authors to choose archives that "Clearly identify preprints as work that is not peer reviewed."
If the article publishes a peer-review file, open it. Nature Portfolio now publishes that file with original research submitted on or after 16 June 2025.

The Official Rule, Not the Campus Shortcut

Campus assignments that say "use peer-reviewed sources" are asking for a publication process. ICMJE is blunt about what that process is for. Independent assessment is "an important extension of the scientific process." It gives a manuscript a hearing among people who work in the field, it helps an editor decide whether the paper belongs in this journal, and it often improves how the work is reported. It is not a certificate that the findings are true.

ICMJE on what review does — and does not — decide

The editor still owns the issue

The editor of a journal is ultimately responsible for the selection of all its content, and editorial decisions may be informed by issues unrelated to the quality of a manuscript, such as suitability for the journal.

Journals must describe the process

For this reason and as a service to authors, journals should publish a clear, transparent description of their peer-review process for all types of manuscripts.

That last sentence is the one to use on a journal website. If you cannot find a description of who reviews, how many people review, and whether identities are hidden, you do not yet have evidence of peer review. The Directory of Open Access Journals will not list a journal that hides this. Its application guide requires a quality-control system before publication and then sets a floor:

DOAJ Guide to applying (peer-review floor)

Minimum review

At least two independent reviewers must review each article, with an exception for arts & humanities journals (for which editorial review must be handled with a minimum of two editors).

The humanities exception is easy to miss. A history or literature quarterly can be a serious scholarly journal and still use editorial review by two editors rather than external referees. That is why "scholarly" and "peer-reviewed" are not synonyms — the distinction the scholarly versus popular guide draws at the publication-type level.

Journal Review Is Not Grant Review

Students meet the word "peer review" in two different systems and collapse them. A journal is deciding whether to publish a finished manuscript. A funder is deciding whether to pay for work that has not been done yet. The National Institutes of Health first-level review page (updated 28 April 2026) is the official description of the second system.

NIH first-level review (grants, not journals)

Who sits on the panel

The first level of review is carried out by a Scientific Review Group (SRG), also referred to as a study section, composed primarily of non-federal scientists who have expertise in relevant scientific disciplines and current research areas.

The score scale

The NIH utilizes a 9-point rating scale (1 = exceptional; 9 = poor) in whole numbers (no decimals).

For most research project grants with due dates on or after 25 January 2025, NIH also changed what those reviewers score. The simplified peer review framework (same 28 April 2026 update) keeps the five regulatory criteria but groups them so reputation is less likely to swamp the science.

NIH simplified review framework (R01 and related codes, due dates 25 January 2025 or later)

Three factors, not five separate scores

The Simplified Framework for NIH Peer Review Criteria retains the five regulatory criteria (Significance, Investigators, Innovation, Approach, Environment) but reorganizes them into three factors — two receive numerical criterion scores and one is evaluated for sufficiency.

Factor 1 — Importance of the Research:

Significance and Innovation, scored 1–9.

Factor 2 — Rigor and Feasibility:

Approach (including inclusion and clinical-trial timeline where they apply), scored 1–9.

Factor 3 — Expertise and Resources:

Investigators and Environment, marked sufficient or not. Gaps need an explanation. There is no separate 1–9 score for the investigator's general reputation.

Same phrase, two jobs
Question	Journal peer review	NIH grant peer review
What is being judged?	A manuscript the authors already wrote	A proposed project, resource, training program, or center
Who chooses the reviewers?	The journal editor	A scientific review officer, who assigns at least three reviewers
Typical output	A decision letter: accept, revise, or reject	A summary statement with criterion scores and, for discussed applications, an Overall Impact score from 10 to 90
Does a good score publish the work?	No. Publication still depends on revision, ethics, and the editor	No. A second-level advisory council reviews mission relevance; an Institute or Center director awards the grant
Can you cite it as a "peer-reviewed source"?	Yes, once the version of record is published	No. A funded proposal is not a published article

NIH also states that assigned reviewers write a critique before the meeting, that reviewers with a conflict leave the room, and that applications judged less competitive — typically the bottom half — may be labeled Not Discussed and receive no final numerical Overall Impact score. None of that is how Nature or PLOS ONE handles a paper. If your methods course uses "NIH peer review" as a synonym for "journal article," the syllabus is mixing systems.

How a Selective Journal Actually Runs Review

The most detailed public description of a working editorial office is the Nature Portfolio peer-review policy. It is not a generic flowchart. It names which article types are refereed (Articles, Letters, Brief Communications, Matters Arising, Technical Reports, Analysis, Resources, Reviews, Perspectives, Insights), which usually are not (most Correspondence and many corrections), and what happens after the inbox.

Desk rejection is a real, official step

Every manuscript is read in-house first. Nature Portfolio sends only the papers that look likely to meet its criteria out to external reviewers. The rest are rejected without a formal referee report. That is the "desk reject" your professor mentioned, written as policy rather than rumor.

Nature Portfolio, "The review process"

In-house screen

All submitted manuscripts are read by the editorial staff. To save time for authors and peer-reviewers, only those papers that seem most likely to meet our editorial criteria are sent for formal review.

How many reviewers

Manuscripts judged to be of potential interest to our readership are sent for formal review, typically to two or three reviewers, but sometimes more if special advice is needed (for example on statistics or a particular technique).

The same policy page lists the four decisions an editor can send after those reports arrive. Learn the labels now; they are the ones that appear in real decision letters.

Nature Portfolio decision menu (verbatim options)

Accept, with or without editorial revisions
Invite the authors to revise their manuscript to address specific concerns before a final decision is reached
Reject, but indicate to the authors that further work might justify a resubmission
Reject outright, typically on grounds of specialist interest, lack of novelty, insufficient conceptual advance or major technical and/or interpretational problems

Reviewers may recommend a course of action. The editor is not taking a vote. Nature Portfolio says so in one sentence that students who imagine a "two out of three reviewers" rule need to see:

Nature Portfolio on conflicting reviews

Not a ballot

Editorial decisions are not a matter of counting votes or numerical rank assessments, and we do not always follow the majority recommendation.

A paper can therefore be rejected after two enthusiastic reviews if the editor judges it too specialized for the journal, or accepted after a split review if the technical objections do not hold. ICMJE makes the same point from the medical-journal side: an editor can reject even after acceptance if integrity concerns appear.

What the reviewers are actually asked

Nature Portfolio publishes the questions it wants in a report. That list is more useful than a generic "they check quality" paragraph, because it is the form a referee fills in.

Nature Portfolio reviewer prompts, condensed
Prompt	What the referee must do
Key results	Summarize the outstanding features of the work
Validity	Say whether any flaw should prohibit publication, and why
Originality and significance	If the conclusions are not original, supply the earlier references
Data and methodology	Comment on the approach, the data, and whether reporting is detailed enough to reproduce the results — including extended data and supplements
Statistics and uncertainty	Check error bars, tests, and how p values are described
Conclusions	Judge whether interpretation is robust, valid, and reliable
Suggested improvements	List experiments or data that would strengthen a revision
References, clarity, inflammatory material	Flag missing citations, an opaque abstract, or language that is inappropriate or potentially libelous

A Nature Immunology editorial that Nature Portfolio quotes on the same policy page states the job in two sentences. Keep them English; they are the official wording.

Nature Immunology editorial, as quoted by Nature Portfolio

What referees are for

Peer review is designed to select technically valid research of significant interest. Referees are expected to identify flaws, suggest improvements and assess novelty.

What referees cannot see

Journal editors do not expect peer review to ferret out cleverly concealed, deliberate deceptions. A peer reviewer can only evaluate what the authors chose to include in the manuscript.

A Validity-Only Journal Judges a Different Question

Not every refereed journal is trying to be Nature. PLOS ONE is the official counter-example students need when a syllabus treats "peer-reviewed" as a prestige rank. The journal still runs external review. It does not ask reviewers whether the paper is trendy.

PLOS ONE on what review is for

Journal information

We evaluate research on the basis of scientific validity, strong methodology, and high ethical standards—not perceived significance.

Reviewer guidelines

Unlike many journals which attempt to use the peer review process to determine whether or not an article reaches the level of 'importance' required by a given journal, PLOS One uses peer review to determine whether a paper is technically rigorous and meets the scientific and ethical standard for inclusion in the published scientific record.

That is why a replication, a negative result, or a careful methods paper can be a legitimate PLOS ONE article and still be desk-rejected at a selective title for "specialist interest." Both can be peer-reviewed. They answered different editorial questions. PLOS ONE's seven publication criteria (original research; not already published; high technical standard; conclusions supported by the data; intelligible English; ethics; reporting and data-availability standards) are the checklist its academic editors and reviewers actually use.

Three review jobs students collapse into one word
Model	When the review happens	What has to be true for acceptance
Selective journal (Nature Portfolio)	After the paper is written; many submissions never leave the in-house screen	Technically valid and of significant interest to that journal's readers
Validity-only journal (PLOS ONE)	After the paper is written; the editor still screens for the seven criteria	Technically rigorous and ethical. Novelty and perceived impact are not the test.
Registered Report (Center for Open Science format)	Stage 1 before data collection; Stage 2 after results are in	The question and methods were sound. The result — positive, negative, or null — is not a reason to revoke the in-principle acceptance.

Registered Reports review the plan, not the plot

The Center for Open Science describes a format more than 300 journals now offer. Reviewers see the question and the methods before anyone knows how the data came out. That is the opposite of the usual incentive to dress a messy result as a clean story.

Center for Open Science, Registered Reports

The format

Registered Reports is a publishing format that emphasizes the importance of the research question and the quality of methodology by conducting peer review prior to data collection. High quality protocols are then provisionally accepted for publication if the authors follow through with the registered methodology.

What has to be in the journal's policy (COS minimum):

Peer review occurs prior to observing the outcomes. Manuscripts that survive Stage 1 receive an in-principle acceptance that will not be revoked because the results were disappointing — only for quality failures, protocol deviations, or unresolvable reporting problems.

If your assignment asks whether a study was preregistered, that is a methods question, not a substitute for journal review. A Registered Report is journal review, moved earlier. Cite the published Stage 2 article, not the private protocol, unless the journal has posted the Stage 1 acceptance as a citable object.

Two Worked Cases: One Famous Skip, One Published Review File

Watson and Crick (1953) is a poor model of modern review

The most famous biology paper of the twentieth century is a poor example of modern peer review, which is exactly why it belongs here. J. D. Watson and F. H. C. Crick, "Molecular Structure of Nucleic Acids: A Structure for Deoxyribose Nucleic Acid," Nature 171, no. 4356 (25 April 1953): 737–738, doi.org/10.1038/171737a0, was accepted by Nature without a systematic external referee report. Nature's emeritus editor John Maddox later said the paper "could not have been refereed"; Stanley Scher's 2004 correspondence in Nature 427, 584 (doi.org/10.1038/427584c) treats that editorial decision as a matter of record. Melinda Baldwin's history in the British Journal for the History of Science (PMC4528400) adds the institutional fact: Nature did not make external refereeing a requirement until editor David Davies did so in 1973.

Cite the paper you mean, and do not invent a review history

The 1953 letter (real bibliographic data):

Watson, J. D., & Crick, F. H. C. (1953). Molecular structure of nucleic acids: A structure for deoxyribose nucleic acid. Nature, 171(4356), 737–738. https://doi.org/10.1038/171737a0

What this case teaches:

A paper can be foundational and still have skipped the process your syllabus now requires. "Published in Nature" is not, by itself, evidence of external peer review — not in 1953, and not today for every section of a Nature Portfolio journal. Check the article type and the journal's current policy.

A 2025 paper that shows you the referee file

You do not have to guess what a modern review looks like. Nature Communications has published the referee comments and author rebuttals with every primary research article submitted from 1 November 2022. Finn Luebber, Sören Krach, Frieder M. Paulus, Lena Rademacher, and Rima-Maria Rahal, "Lottery before peer review is associated with increased female representation and reduced estimated economic cost in a German funding line," Nature Communications 16, Article 9824 (6 November 2025), doi.org/10.1038/s41467-025-65660-9, is a real article about funding review that also has a published journal-review file. The article page states that a peer review file is available. Open the PDF of the paper for the claim; open the Transparent Peer Review file for the argument the referees actually had.

A refereed 2025 article you can audit

APA 7 (version of record):

Luebber, F., Krach, S., Paulus, F. M., Rademacher, L., & Rahal, R.-M. (2025). Lottery before peer review is associated with increased female representation and reduced estimated economic cost in a German funding line. Nature Communications, 16, Article 9824. https://doi.org/10.1038/s41467-025-65660-9

What the paper is (so you do not mis-cite it):

An empirical report on the Freiraum funding line at Stiftung Innovation in der Hochschullehre. It is journal peer review about a grant-allocation experiment. It is not itself an NIH study-section score, and the published review file is not a grant summary statement.

What the review file will not show:

Confidential comments to the editor, and the editors' private discussion. Nature Portfolio states that the published file is a view into the process, not the entire decision record.

Types of Peer Review (Use the Journal's Words)

Older textbooks say "single-blind" and "double-blind." Current publisher pages say anonymized. Elsevier's reviewer explainer is the clearest public glossary of the models journals actually run. Nature Portfolio's policy uses the same vocabulary and, from 16 June 2025, treats transparent review as the default for original research rather than an author opt-in.

Review models, with the names journals print
Model	Who is hidden	What to watch for
Single anonymized ("single-blind")	Reviewers hidden from the author	Elsevier calls this "the traditional method of reviewing and … the most common type by far." Reviewer prestige bias against the author is the usual complaint. PLOS ONE uses this model unless a reviewer signs.
Double anonymized ("double-blind")	Author and reviewer hidden from each other	Elsevier: "Both the reviewer and the author are anonymous in this model." Anonymity often fails in small fields because of self-citation and topic. Some Nature-branded journals offer this as an opt-in.
Triple anonymized	Author, reviewer, and decision-making editor hidden from one another	Rare. Elsevier warns that the handling editor who invites reviewers still sees identities, and that writing style can unmask authors anyway.
Open review	Identities known; sometimes reports are signed	Elsevier treats "open review" as an umbrella: named reviewers, published reports, or a public discussion after a pre-check. Accountability goes up; so does fear of retaliation.
Transparent peer review (TPR)	Reports published beside the article; reviewer names may stay hidden	Nature Portfolio: the published file can include reviewer comments, author rebuttals, and decision letters. Confidential comments to the editor stay unpublished.

Elsevier, "Types of peer review"

Single anonymized

In this type of review, the names of the reviewers are hidden from the author. This is the traditional method of reviewing and is the most common type by far.

Double anonymized

Both the reviewer and the author are anonymous in this model.

The Nature Portfolio change that most syllabi have not caught: transparent review is no longer a courtesy opt-in on the flagship titles.

Nature Portfolio, "Transparent peer review" (policy page, 2025 rule)

The 16 June 2025 default

For submissions received on or after 16th June 2025, this peer review file will be published alongside all our original research articles.

What that file is:

Reviewer comments to the authors, author rebuttal letters, and in some cases the editorial decision letters, posted as a supplementary peer-review file. Manuscripts submitted before that date could still opt out.

What that file is not:

A complete transcript of the decision. Transfers from another Nature Portfolio journal do not republish the earlier journal's reports.

Post-publication commentary is a separate channel, not a replacement for the pre-publication report. Nature Portfolio formalizes challenges, clarifications, and replications as Matters Arising, which are themselves peer-reviewed. ICMJE §II.C.2.c says some people believe "true scientific peer review begins only on the date a paper is published" and tells medical journals to keep a mechanism for reader comments. Neither sentence means an unreviewed blog post is peer review.

How Long It Takes, and Why "Fast Acceptance" Is a Warning

There is no official global clock. Nature Portfolio asks reviewers to respond "promptly within the number of days agreed" and to warn the editor if they will be late. ICMJE §II.C.2.b tells editors to reject quickly when they have no intention of proceeding, so authors can try another journal. UC Berkeley's Understanding Science explainer — a teaching page, not a journal policy — still gives students the scale they need: review plus publication "frequently involving more than a year between submission and publication," and it reports Science accepting less than 8 percent of submissions and the New England Journal of Medicine publishing about 6 percent. Treat those rates as Berkeley's classroom figures, not as a live acceptance dashboard.

A journal that guarantees acceptance, or a decision in a few days, is failing the Think. Check. Submit. peer-review test. The official journals checklist (March 2023, amended June 2023) asks two things that belong on this page rather than on a style-comparison guide:

Think. Check. Submit. journals checklist — peer-review items

The heading

Is the journal clear about the type of peer review it uses?

The speed test

The journal doesn’t guarantee acceptance or a very short peer review time.

Follow-up questions on the same checklist:

Does the website mention whether the process involves independent/external reviewers, how many reviewers per paper? Is the publisher offering a review by an expert editorial board or by researchers in your subject area?

How to Tell Whether the Item in Your Hand Was Reviewed

This is the failure mode that sends students to the wrong PDF. A database "Peer reviewed" limiter, as How to find scholarly sources already warns, is usually a flag on the journal. Editorials, letters, and book reviews still sit in that result list. Confirm the article type on the PDF first. Then confirm the journal.

Read the journal's own policy

ICMJE requires a published description of the process for every manuscript type. Nature Portfolio's page is the model: it names the types that are reviewed, the typical number of reviewers, the anonymity rule ("We do not release referees' identities to authors or to other reviewers unless a referee voluntarily signs their comments"), and the AI rule. If a journal's "peer review" page is two marketing sentences and a stock photo, keep looking.

Ulrichsweb and MEDLINE are journal-level tools

The University of Texas at Dallas library guide How to Determine if a Journal Is Peer-Reviewed states the librarian's first reminder — "not every scholarly journal is peer reviewed" — and then the Ulrichsweb check: search the journal title, look for the striped referee-shirt icon, and open the record's Refereed field (Yes or No). Treat that badge as a catalog field, not an independent audit of the journal's review. For the icon itself and the difference between "academic" and "refereed," use the worked Ulrichsweb section in Scholarly vs popular sources.

For health and life sciences, NLM's MEDLINE journal selection (last reviewed 24 July 2026) is a stronger signal than a homepage badge. Consultants reviewing a title are asked, among other editorial-policy questions:

NLM MEDLINE Scientific Quality Review — editorial-process prompt

What the consultants look for

Is the peer review process explicit and sufficiently detailed, including information on the type of peer review and the number of reviewers typically assigned to a manuscript?

MEDLINE indexing is not a claim that every item in the journal was refereed. It is a claim that NLM judged the journal's scientific and editorial character, including that process description, worth inclusion. Look the title up in the NLM Catalog journals subset and prefer "currently indexed in MEDLINE" over a publisher's marketing line.

Preprints, books, and magazines

Preprints. Posted before journal review. ICMJE §III.D.3 tells journals to publish a preprint policy and tells authors that posting can affect a journal's interest in the paper. Cite a preprint only when you must, and label it. If a version of record exists, cite that.
Books and book chapters. University-press monographs are edited; they are not "peer-reviewed articles." Some presses send manuscripts to external readers. That is a different process with different documentation. Do not apply a journal filter to a book.
Magazines, newspapers, trade publications, and most websites. Not journal peer review. Sort them with scholarly versus popular before you argue about credibility.
Grey literature — agency reports, working papers, white papers — can be authoritative without being refereed. That is grey literature, not a failed journal article.

Predatory and Pseudo-Journals

ICMJE §II.C.1.a is the official medical-publishing definition. Use it instead of a crowd-sourced "blacklist" that you cannot audit.

ICMJE Recommendations, §II.C.1.a

What "predatory" or "pseudo-journals" do

These journals (“predatory” or “pseudo-journals”) accept and publish almost all submissions and charge article processing (or publication) fees, often informing authors about this after a paper’s acceptance for publication. They often claim to perform peer review but do not and may purposefully use names similar to well-established journals.

ICMJE adds that these titles may claim ICMJE, COPE, or WAME membership they do not have, and that authors "should avoid citing articles in predatory or pseudo-journals." The student version of that duty is: do not pad a bibliography with a paper whose only home is a journal that sells acceptances. Check the publisher name, the editorial board, the stated fees, and whether the journal is in DOAJ (if it is open access) or indexed in a database your librarians actually use. Then finish the Think. Check. Submit. list. For open-access business models that are legitimate, see open access research.

Limitations You Should Keep in the Paper, Not Just the Lecture

ICMJE says the "actual value of peer review is widely debated" in the same paragraph that defines it. Berkeley's Understanding Science page is the classroom wording students remember:

UC Berkeley, Understanding Science — "Scrutinizing science: Peer review"

The limit

Peer-reviewed work isn’t necessarily correct or conclusive, but it does meet the standards of science.

It is a filter, not a replication. Reviewers read what the authors submitted. They do not usually re-run the experiment or re-analyze the data from scratch.
It is weak against concealed fraud. Nature Portfolio's quoted Nature Immunology editorial says editors do not expect referees to catch "cleverly concealed, deliberate deceptions." Retractions exist because review missed things. See how to evaluate sources after you have a refereed paper in hand.
It is slow, and it is biased toward what already looks like a paper. ICMJE tells editors not to reject work merely because findings are negative or inconclusive. Publication bias is a documented failure mode, not a conspiracy theory. Registered Reports exist because of that bias.
Most reviewers are unpaid. Nature Portfolio thanks "many tens of thousands" of referees and "the time that reviewers devote to assessing the manuscripts we send them." NIH's first-level page describes reviewers as scientists, mostly from academia, who serve on study sections. Fatigue and uneven thoroughness are part of the system, not a side issue.
Impact factor is the wrong proxy. ICMJE §II.C.2.f is explicit, and it is the reason this page will not tell you to "check the impact factor" as a peer-review test.

ICMJE Recommendations, §II.C.2.f

Impact factor

The journal impact factor is widely misused as a proxy for research and journal quality and as a measure of the importance of specific research projects or the merits of individual researchers, including their suitability for hiring, promotion, tenure, prizes, or research funding.

If you need metrics at all, use them the way citation metrics teaches — as one signal among others, never as a substitute for reading the methods. For political, commercial, or methodological slant inside an otherwise real journal, use bias in sources.

AI in the Review File

Confidentiality is the binding constraint. ICMJE treats submitted manuscripts as the authors' private property. The January 2026 Recommendations keep the reviewer rule in §II.C.3 and add a dedicated Section V on AI in publishing. Reviewers still need the journal's permission before using a tool, and they must not upload a manuscript into a system that cannot keep it confidential. Nature Portfolio is stricter on the upload and requires a disclosure if any part of the evaluation used an AI tool.

Official AI rules for reviewers

ICMJE §II.C.3

Reviewers must request permission from the journal prior to using AI technology to facilitate their review.

ICMJE on AI output

Reviewers should be aware that AI can generate authoritative-sounding output that can be incorrect, incomplete, or biased.

ICMJE Section V (January 2026) — confidentiality

Editors, reviewers, or publishers should not upload submitted manuscript into AI systems where confidentiality cannot be assured without authors’ explicit permission.

Nature Portfolio, "AI use by peer reviewers"

If any part of the evaluation of the claims made in the manuscript was in any way supported by an AI tool, we ask peer reviewers to declare the use of such tools transparently in the peer review report.

DOAJ Guide to applying — AI assessments

Peer reviewers and editors should not use generative AI to create their assessments, due to risks such as breaches of confidentiality, superficial and non-specific feedback, bias, hidden prompts, and false information such as fake references; editing and rewriting may be acceptable if disclosed.

A student using a chatbot to "peer review" a classmate's draft is not doing journal peer review. The confidentiality and hallucination problems are the same. If you used a tool on your own draft, that is an authorship-disclosure question for how to cite AI, not a substitute for a human reader in the field.

What This Means When You Write the Paper

Professors ask for peer-reviewed sources because they want you inside the conversation that other researchers already had, not because a referee stamp makes a sentence undebatable. Read the abstract, then the figures, then the limitations, using how to read academic papers. If the assignment allows a mix of sources, say why a newspaper investigation or an agency report is the right type for that claim. If you cannot find a refereed article, that is a search problem or a topic problem — ask a librarian before you decide the field is empty.

When you are ready to cite the article you actually used, format the version of record, not the accepted manuscript sitting in a repository, unless that is the only file you have. Start with how to cite a journal article in APA or MLA, and use the generator only after you have confirmed the article type and the journal.

Cite the version of record, not the filter

A "peer-reviewed" database hit is a starting point. Confirm the article type, then format the published journal article — not the preprint, not the news write-up of the trial. If a peer-review file is posted, read it before you treat the paper as settled.

Cite a journal article in APA 7 Open the citation generator

Frequently Asked Questions

Are all journal articles peer-reviewed?: No. ICMJE says a peer-reviewed journal does not have to send every submission out for review. Nature Portfolio reviews Articles, Letters, and several other types, but Correspondence and many corrections are usually not refereed. Always check the article type on the PDF and the journal's stated policy for that type.
Does "peer-reviewed" mean the findings are correct?: No. Berkeley's Understanding Science page states the limit directly: peer-reviewed work is not necessarily correct or conclusive. Nature Portfolio's quoted Nature Immunology editorial adds that reviewers cannot see data the authors omitted and are not expected to catch clever fraud. Read the methods and the limitations.
What is the difference between peer-reviewed and scholarly?: Scholarly describes the audience and the apparatus (experts writing for experts, with citations). Peer-reviewed describes a specific pre-publication process. DOAJ allows some arts and humanities journals to use two-editor editorial review instead of external referees. A university-press book can be scholarly without being a refereed article.
Is every peer-reviewed journal judging the same thing?: No. Nature Portfolio asks whether the work is technically valid and of significant interest. PLOS ONE asks whether it is technically rigorous and ethical, not whether it is important. A Registered Report judges the question and the methods before the results exist. "Peer-reviewed" names the process, not the editorial bar.
How long does journal peer review take?: There is no official global timeline. Nature Portfolio asks reviewers to meet an agreed number of days; ICMJE tells editors to reject promptly when they will not proceed. Berkeley's teaching page says the full path from submission to publication often takes more than a year. A journal that guarantees acceptance in days is failing Think. Check. Submit.
Is a preprint peer-reviewed?: Not by the journal that has not yet accepted it. ICMJE tells preprint archives to identify that work as not peer-reviewed. Cite the published version if it exists. If you must cite the preprint, label it as a preprint in the style your assignment names.
How is NIH grant review different from journal review?: NIH first-level review is a Scientific Review Group scoring a proposed project on a 1–9 scale; discussed applications receive an Overall Impact score from 10 to 90, then a council looks at mission relevance. For most RPGs due on or after 25 January 2025, reviewers score Importance and Rigor/Feasibility and mark Expertise/Resources as sufficient or not. A journal editor is deciding whether to publish a finished manuscript. A funded grant is not a peer-reviewed article.
Where can I read a real referee report?: On journals that publish transparent peer-review files. Nature Portfolio now posts that file with original research submitted on or after 16 June 2025. Nature Communications has done so for submissions since 1 November 2022. Luebber et al., Nature Communications 16, Article 9824 (2025), is one article whose page links a peer-review file. The file still omits confidential comments to the editor.
How do I spot a journal that only pretends to review?: Use ICMJE's test: the title accepts almost everything, charges a fee (often after a fake acceptance), claims review it does not do, and may copy a familiar journal's name. Then use Think. Check. Submit.: can you name the publisher, see the review model, see the fees, and find the journal in DOAJ or a database your library pays for?
Can I use a chatbot to peer-review my own draft?: That is not journal peer review. ICMJE Section V and Nature Portfolio both treat uploaded manuscripts as confidential and warn that AI output can be wrong, incomplete, or biased. Nature Portfolio requires any AI assistance in a real referee report to be declared. Use a human reader in your field, or a campus writing center, for a draft.