Independently tested — no sponsored rankings
Subscribe
The Fine Print

How to Read a Study: A Plain-Language Guide to What the Research Really Says

This post may contain affiliate links. If you buy something through them, we may earn a commission — at no extra cost to you, and it never affects our ratings. Read our full disclosure.

Georden Jones, Founder, The Peer Review · Last updated: August 9, 2026

The Short Answer

You do not need a PhD to get useful information out of a scientific study — you need a short checklist and the willingness to look past the headline. Most health scares rest on a single paper that few people ever open, and the paper often says something more careful than the article about it. The questions that matter are answerable by a non-scientist: How many people (or animals, or cells) were studied? Was there a comparison group? Did the result hold up statistically, and was the size of the effect meaningful? Who paid for it, and has anyone repeated it? This page walks through each of those, so the next time you see “new study finds,” you can check the study yourself instead of trusting the person summarizing it.

Table of Contents

Start With the Abstract — but Don’t Stop There

Every study opens with an abstract — a short summary of what the researchers did and what they found. It is the fastest way to see what a paper claims, and it is free to read even when the full study sits behind a paywall. Read it first: it tells you the type of study, roughly how big it was, and what the authors concluded.

But the abstract is also where a study looks its most confident, because it compresses caveats out. The limitations, the messy numbers, and the “this needs more research” hedges live in the full paper, especially the Methods and Discussion sections. If a claim matters to you, the abstract is where you start, not where you finish. A study that sounds alarming in its abstract often turns out, in the Methods, to be a handful of cells in a dish (see the Evidence Hierarchy).

Who and How Many? (Sample and Design)

Two of the most revealing questions about any study are who was studied and how many of them there were.

Who tells you whether the result applies to real people. A finding in mice, or in isolated cells, does not automatically transfer to humans living normal lives — it is a signal to investigate, not a conclusion about you. A finding in thousands of people over years carries far more weight for human health.

How many — the sample size — tells you how much to trust the result. A study of 12 people can be knocked around by chance; a study of 12,000 is much harder to fool. Small studies are not worthless, but their findings are fragile and often shrink or vanish when someone runs a bigger version. When a headline rests on a study of a few dozen participants, that is worth knowing before you change anything about your life [1].

The design matters too: was this a one-time snapshot, a group followed over time, or a randomized trial where people were assigned to a treatment or a control? That single fact places the study on the evidence hierarchy and tells you how much it can prove.

Was There a Control Group?

A control group is the comparison — the people (or samples) who did not get the thing being tested, so the researchers can see what would have happened anyway. Without one, a study can describe but rarely prove.

Here is why it is so important: lots of things get better (or worse) on their own. If a study gives 50 people a supplement and reports that most felt better, that means little without a comparison group who got a dummy pill — because people often feel better regardless, from the placebo effect, from time passing, or from expecting to improve. The control group is what separates “this worked” from “this would have happened anyway.” When you read about a dramatic result, one of the first questions is: compared to what? If the answer is “compared to nothing,” treat the finding as preliminary [1].

“Significant” Doesn’t Mean “Big”

This is the single most misunderstood word in science reporting. When a study calls a result “statistically significant,” it does not mean the effect is large or important. It means the result is unlikely to be pure chance — usually measured by a p-value below 0.05, meaning under a 5% probability the finding is a fluke [2].

Two traps follow from this. First, “significant” says nothing about size: a statistically significant effect can be so small it makes no real difference to anyone. Second, with a large enough study, almost any tiny difference becomes “significant,” because significance partly reflects how many people were studied, not just how big the effect is. So “significant” is a gate a result passes through, not a measure of how much you should care. A study can be perfectly significant and completely trivial [2].

Effect Size: The Number That Really Matters

If significance tells you a result is probably real, effect size tells you whether it is big enough to matter — and it is the number most headlines leave out. An effect size is the actual magnitude of the difference: how many more cases, how much lower the blood pressure, how much bigger the risk in real terms.

This is where the difference between relative and absolute risk lives. “Doubles your risk” is a relative effect that can hide a tiny absolute one: a study might show a real, significant, reproducible effect that still only raises your absolute risk from 0.1% to 0.2% — true, and also almost nothing. Always look for the real-world number behind a claim: how much does this change the outcome for a real person? If a study or the article about it only gives you percentages and ratios, and never a plain “X out of 1,000,” the most important number is missing.

Peer Review, Preprints, and Where It Was Published

Peer review is the process where other scientists in the field check a study before a journal publishes it. It is a quality filter, not a guarantee of truth — reviewers miss things, and plenty of peer-reviewed work is later overturned — but a peer-reviewed paper in a reputable journal has cleared a bar that a blog post or a press release has not [3].

A preprint is a study posted publicly before peer review. Preprints are valuable — they share findings fast — but they have not yet been vetted, so they deserve extra caution. If a scary claim traces back to a preprint, it is early, unchecked evidence. It is also worth glancing at where a study was published: predatory journals will print almost anything for a fee, so a paper in an obscure pay-to-publish outlet carries less weight than one in an established journal. And no matter where it appeared, a single study is a data point — the real signal comes when multiple studies, ideally pooled in a systematic review, point the same way (see the Evidence Hierarchy).

One more question belongs here and gets its own page: who funded the study? Funding does not automatically invalidate research, but it is part of reading a paper honestly — covered in Who Funded This?.

A Quick Field Checklist

When a study crosses your feed, run these in about a minute:

  1. What kind of study is it? Cells, animals, or people? Snapshot, followed group, or randomized trial? (See the Evidence Hierarchy.)
  2. How many, and who? A few dozen or many thousands? The population that applies to you, or not?
  3. Compared to what? Was there a control group, or just a before-and-after with no comparison?
  4. Significant vs. big. “Significant” means probably-not-chance, not large. Where is the effect size?
  5. What’s the real number? Look for the absolute effect, not just “doubles” or “50% more.”
  6. Vetted and repeated? Peer-reviewed or a preprint? One study, or many agreeing?
  7. Who paid, and did they benefit? Not disqualifying, but part of the picture.

You will not answer all seven perfectly every time, and you do not need to. Even two or three of these questions will deflate most of the alarming health content online, which counts on you never opening the study at all.

FAQ

Do I need to read the whole paper?

Usually no. The abstract plus a glance at the Methods (how many people, what design, was there a control) answers most of what you need. The full Discussion section is where authors list limitations if you want more.

What does a p-value really mean?

Roughly, the probability that you would see a result this strong if there were truly no effect. A p-value under 0.05 is the common cutoff for “statistically significant,” but it says nothing about how big or important the effect is [2].

Is a peer-reviewed study always trustworthy?

No — peer review is a filter, not a guarantee. It catches many problems but misses others, and peer-reviewed findings are sometimes overturned. It raises confidence; it does not settle a question on its own [3].

What’s wrong with a small study?

Nothing, as a starting point — but small studies are easily swayed by chance and often shrink or disappear when repeated at larger scale. Treat a small study as a lead, not a verdict [1].

What’s a preprint?

A study shared publicly before peer review. Useful for speed, but unvetted — so a claim resting only on a preprint is early evidence that has not yet been checked by other scientists.

The Bottom Line

Reading a study is less about expertise than about asking a few stubborn questions: who and how many, compared to what, is the effect real and big, was it vetted, and has anyone repeated it. The headline is written to be shared; the study is where the honest, hedged version lives, and it is more accessible than most people assume. You do not have to become a scientist — you just have to open the paper and refuse to be impressed by “a study found” until you know what kind of study, how large, and how it compares. That habit alone will make you harder to scare and harder to sell to.

Stay curious, stay critical.
Georden

References

  1. Health Knowledge (UK public health textbook). “The hierarchy of research evidence.”
  2. Wasserstein RL, Lazar NA. “The ASA Statement on p-Values: Context, Process, and Purpose.” The American Statistician.
  3. “Peer review: what is it and why do we do it?” Health Knowledge.

Georden Jones is the founder of The Peer Review. Read the full story.