Peer review · Lesson 1 of 3

What it is for, and the forms it takes

Peer review checks competence and support. It does not verify data or guarantee replication.

7 min read · Free — no account needed to read

Peer review asks people with relevant expertise whether a piece of work is competently done, whether its conclusions follow from its evidence, and whether it contributes something. It is a filter and an improvement process, and it is performed by busy humans.

What it does well: catches methodological errors, unsupported claims, missing literature, overstated conclusions, and unclear reporting. What it does not do: verify that the data exist, detect most fraud, guarantee replication, or determine importance reliably.

Editorial screening comes first — an editor decides whether to send it out at all. Single blind: reviewers know the authors, authors do not know reviewers. Double blind: neither knows, which reduces bias by name and institution and is imperfect because work is often identifiable. Open review: identities known, sometimes with reports published, which raises accountability and can make junior reviewers cautious. Post-publication review happens in the open after release, and preprints have made it far more common.

Knowing the form matters when you read a review: an anonymous reviewer with a competing project is a real risk single-blind systems carry, and an open reviewer writing under their name may be gentler than the work requires.

See it happen

A paper is published in a well-regarded peer-reviewed journal.

The read that arrives on its own

It has passed peer review, so the methods and findings have been verified.

The same situation, with this idea

You keep the right expectation. Two or three people read it, usually without the data, usually unpaid, often in an evening. They checked whether it was competent and whether the conclusions followed. They did not re-run the analysis or verify the data exists. Peer review raises the floor; it does not certify the result, which is why replication and post-publication scrutiny still matter.

After this, you can

read "peer-reviewed" as a floor rather than a guarantee.

The takeaway

Peer review checks competence and support. It does not verify data or guarantee replication.

Easy to take the wrong way

That its limits make it worthless. The floor it sets is real — compare any literature with an unreviewed one.

Try this

Find the reviewer guidelines for a journal in your field and read what they actually ask reviewers to assess.