How to Become a Peer Reviewer and Get Credit

Editors can only invite reviewers they can find and trust. Most early career researchers are qualified long before they are visible.

Guide · 11 min read · Directive Publications Editorial Team

What editors need before they invite you

Some journals appoint reviewers to a formal panel, and their reviewer guidelines will say so. Otherwise, an invitation starts with an editor who has a manuscript on their desk and needs someone who works on that specific question or method, can judge it without bias, and can be reached at contact details that are genuinely theirs. If you meet those three conditions and the editor can see that you do, invitations tend to follow. If you meet them and the editor cannot see it, they do not.

The COPE Ethical Guidelines for Peer Reviewers put it from the editor's side: editors must match reviewers to the scope of a manuscript, and potential reviewers should give journals accurate information that fairly represents their expertise, with verifiable contact details. The same guidelines treat impersonation during review as serious misconduct, which is one reason editors are cautious about contact details they cannot verify.

Nothing in that list is about seniority. An editor who needs someone to judge a propensity score analysis in a registry cohort wants a person who has done one recently, and a postdoc who has published that analysis is a better match than a professor whose last such paper was a decade ago. What holds early career researchers back is almost never the expertise. It is that the expertise is invisible, unverifiable, or attached to an email address that bounces.

Make your expertise findable

The people an editor can find most easily are those who have already published on the manuscript's topic. Even one first-author paper puts your name next to a method and a question. Editors may also draw on recommendations from colleagues, reviewers suggested by authors, their submission system's reviewer records, and commercial tools such as Web of Science Reviewer Locator, which searches researcher profiles alongside publication and citation data. Every one of those routes runs on your public record, so the job now is to make that record easy to connect to you.

• ORCID. List your works, your current employment and an institutional email address. Our guide to ORCID, affiliations and authorship covers setting it up properly.
• Journal accounts. You get an account in a journal's submission system when you submit a paper, and most systems also let you register directly without submitting anything. Either way, if the account has fields for expertise or keywords, fill them in.
• Specific terms. "Oncology" matches thousands of manuscripts and tells an editor nothing. "Survival analysis in observational cancer cohorts" tells them exactly when to think of you.
• One email address. Use the institutional address on your papers for journal accounts as well, so the record, the publications and the reviewer all line up.

Start with a co-review, with permission

A common way to learn is to work on a real manuscript alongside someone who already reviews. That is legitimate only when the journal knows about it. COPE's guidelines say reviewers must not involve anyone else in a review, including early career researchers they mentor, without first getting the journal's permission. A supervisor who wants to involve a student must ask the editor and abide by the decision. Where the student does the review under guidance, the student should be acknowledged as the reviewer of record.

Sharing a confidential manuscript with someone the journal did not invite is a breach of confidentiality, even inside your own lab. Getting permission takes one email, so ask for it.

A short request to your supervisor makes this easy to say yes to:

If your supervisor rarely reviews, practise on a preprint or a published open access paper in your field and ask for feedback on your report. Structured options exist too: the Web of Science Academy runs peer review courses aimed at PhD students and postdocs who have started publishing, including one built around co-reviewing with a mentor.

Tell editors you are available

Once you have a publication or two and some supervised review practice, tell editors directly. Pick journals you read and cite, where your own work would fit. If a journal or publisher has a reviewer registration page, use it. Otherwise write to the editor:

Keep it that short. Send it from your institutional address. Do not send the same note to every journal you can find: an editor can tell when a message was not written for their journal, and a focused note to a few journals in your field is worth more.

Check the invitation before you say yes

Declining a review you cannot do well helps the editor more than accepting it: a late or shallow report costs the editor weeks they would have spent finding someone else. Before you click accept, work through this list:

1. The manuscript sits within your expertise, or you can tell the editor clearly which parts you cannot assess
2. You do not work at the same institution as any author, and you have not recently been a mentor, mentee, close collaborator or joint grant holder of any author (COPE gives the past three years as an example)
3. You are not preparing a closely similar manuscript of your own
4. You can return the report by the deadline
5. You know the review model: whether identities are anonymised, whether reports are published, and whether you would sign
6. You know the journal's policy on AI tools
7. The journal is real and in your field

The last item matters because a reviewer's name lends credibility. An invitation from a journal you have never heard of deserves the same checks as a journal you are thinking of submitting to, and our list of predatory journal red flags covers them. If a report might be published, read our guide to transparent peer review before you decide whether to sign.

If a conflict comes to light after you accept, tell the journal. In anonymised review, do the same if you recognise the authors and that creates a conflict. COPE advises not looking at the manuscript again until the editor replies.

To decline, reply promptly and suggest someone if you can:

Suggest people on expertise alone, and do not tell them anything about the manuscript.

Write a report the editor can use

Read the journal's reviewer instructions, the whole manuscript and the supplementary files before writing anything. For clinical studies, check the reporting against the guideline the journal requires, such as CONSORT for randomised trials, and compare the reported outcomes with the registration record. Then structure the report so an editor can act on it:

1. Summary. Two or three sentences on what the study asked, how and what it claims. This shows the editor you read the paper the way the authors meant it.
2. Major concerns. Number them. For each one, give the location (section, table or line), explain why it affects whether the conclusions are supported, and say what would resolve it. Say which requests are essential and which would only strengthen the work, because extending the study beyond its scope is not the reviewer's job.
3. Minor points. Unclear wording, missing definitions, labelling problems. Suggest clearer phrasing, but do not rewrite a sound paper into your own style, and be respectful of authors writing in a second language.
4. Confidential comments to the editor. Raise integrity concerns here. Do not investigate them yourself, and do not contact the authors without the journal's permission. Say which parts of the manuscript you assessed and which you could not, such as the statistics, so the editor knows where to send it next; that note can go to the authors as well, since most of your feedback belongs in the report they will read.

Keep your recommendation consistent with what the authors will read. Do not ask authors to cite your own work unless there is a genuine scientific reason. If the journal later asks you to review the revision, try to say yes, since you are best placed to judge whether your concerns were addressed.

Confidentiality applies throughout. The ICMJE responsibilities for peer reviewers say reviewers must keep the manuscript confidential, must not use the authors' ideas before publication, and should destroy their copies after submitting the review. Under ICMJE's guidance on AI use by reviewers, you must follow the journal's AI policy or ask permission before using AI. Confidentiality can rule out uploading the manuscript to AI tools where it cannot be assured, unless the journal explicitly allows it, and any AI use should be disclosed to the journal.

Record the review so you get credit

Credit for review work only exists if something records it. Start your own log the day you accept your first invitation: journal, date invited, date submitted, and whether it was a first round or a revision. File the journal's thank-you or decision email with it. That email is your evidence.

ORCID's peer review help page says peer reviews can only be added to your record by a trusted organisation using Member API credentials, with your explicit permission: either the organiser of the review, such as the publisher, society or funder, or a third-party recognition service such as Web of Science. You cannot add them yourself, and the section does not appear on your record until one of those organisations puts a review there. To record serving as a reviewer for a journal in general, ORCID asks you to use a Service affiliation instead.

Web of Science, which took over the Publons reviewer profiles in August 2022, does not accept an invitation or an acceptance email as a receipt, because accepting a review is not the same as completing one. Once your accounts are connected, it can also export your reviews to ORCID.

If you review for a journal that publishes signed reports, the report itself can become a public, linkable part of your record.

List review work on your CV

Give peer review its own section and keep it factual:

• Reviewer for [Journal A] ([number] manuscripts, [years]) and [Journal B] ([number] manuscript, [year])
• Co-reviewer, with editor permission, for [Journal C] ([number] manuscript, [year])
• Verified record: ORCID iD [link]; Web of Science Researcher Profile [link]

Do not list manuscript titles, authors or decisions, because the review was confidential, unless the review was published with the article: a signed, published report is already public, and Web of Science lets you show the article title for reviews of that kind. Decide once whether a revision round counts as a separate review, and count the same way every time. Keep editorial board roles in a separate section, since editors and reviewers do different jobs.

If you want to see how a journal sets out what reviewers assess and what it expects of them before you volunteer, the Directive Publications peer review process page describes both, including how to express interest in reviewing.