Reading science like a scientist
Preprints and peer review, explained
Every story on this site wears a label: Preprint or Peer reviewed. It tells you how carefully other scientists have checked the research so far. Here’s what the labels mean and why they matter.
In one minute
- Peer reviewedIndependent experts read the study and questioned it before a journal published it.
- PreprintThe scientists shared the paper publicly before that checking is finished.
- Both are real science. A preprint is like a first draft that anyone can read: exciting, but a “maybe” until it has been checked.
The journey of a study
Scientists ask a question, run the study and write it up as a paper. Then the paper can take two routes, and plenty of papers take both.
Route 1: the fast lane
Preprint
- 1
Post it online
The authors upload the paper to a preprint server.
- 2
Basic screening
The site usually checks that it’s real research and safe to share. It does not judge whether the science is right.
- 3
Anyone can read it
Within days, other scientists, journalists and you can see it, and often comment.
Route 2: the checked lane
Peer review
- 1
Send it to a journal
An editor decides whether it’s worth reviewing.
- 2
Experts check it
Usually two or three scientists who study the same area, but weren’t involved, read it closely.
- 3
Fix, accept or reject
The authors answer questions and often run extra checks. Then the journal publishes the paper, or says no.
Many papers go through both routes: a preprint first so others can see the work quickly, then a journal version after review. The journal version is sometimes quite different from the first draft.
What does peer review actually check?
The reviewers are the author’s “peers”: other experts in the field. They’re usually not paid for it, and they often stay anonymous so they can be honest. They ask questions like:
- Did the experiment really test the question it says it did?
- Were the numbers handled properly?
- Do the conclusions go further than the evidence allows?
It usually takes weeks or months, which is why preprints exist.
Side by side
| Question | Preprint | Peer-reviewed paper |
|---|---|---|
| Who has checked it? | Not yet formally checked by independent experts. | Independent experts have read and questioned it. |
| How fast is it shared? | Days. | Often months. |
| How sure can we be? | It’s early evidence. Promising, but wait for the checks. | Better tested, though not perfect. |
| Can it change? | Yes, often. Authors revise it, and some preprints are withdrawn or never published. | Yes. Corrections happen, and rarely a journal retracts a paper. |
Does “peer reviewed” mean “true”?
No. It means the study passed a quality check. Reviewers can miss mistakes, and a study can be honest and still be too small to be sure.
Science builds confidence slowly, as more teams repeat and test an idea. That’s why one study, reviewed or not, is a clue and not the final answer. Peer review is a useful filter. It isn’t a guarantee.
How we handle this on The Synaptic Report
- Every story carries a label, and preprint stories say so plainly.
- Our editor picks at most two preprints a day and prefers peer-reviewed work.
- Every story links to the original paper, so you can read it yourself.
- We write with AI help and check each story against the paper, but the original paper has the final say.
A checklist for reading science news
Try these questions on any science story, ours included.
- 1
Is it peer reviewed or a preprint?
A preprint is early evidence. A peer-reviewed paper has had a first round of expert checking.
- 2
Who or what was studied?
People, mice, cells and computer models are very different. A result in mice often doesn't carry over to people.
- 3
How big was the study?
Ten people can show a hint. Ten thousand can show much more.
- 4
Does it show a cause, or only a link?
"Linked to" means two things happen together. It doesn't mean one causes the other.
- 5
Has anyone else found the same thing?
One study is a clue. When several teams get the same answer, we can be much more sure.
- 6
Could anyone gain from the result?
Who paid for the study, and whether they'd like a particular answer, can matter.
Words to know
- Journal
- A publication, such as Nature or The Lancet, that publishes studies after they have been reviewed.
- Peer review
- Independent experts in the same field checking a study before it is published.
- Preprint
- A paper shared publicly before peer review is finished.
- Preprint server
- A website where scientists post preprints, such as bioRxiv, medRxiv or arXiv.
- Replication
- Another team repeating a study to see whether they get the same result.
- Retraction
- A journal publicly withdrawing a paper after finding serious problems with it.