Wellness / Last reviewed 12 August 2026

How to read a health headline

Six questions that let you dismiss most fitness news in the time it takes to read the sub-headline.

The method, not the verdict

This is the piece that explains how we work. We do not summarise studies for you, because a summary is a thing you have to trust. We would rather you could check.

Six questions, in order.

1. Who was it about?

The first paragraph says "adults". The study may have been thirty men aged 20 to 25 in a university sports department, or mice. Both are legitimate research and neither is a finding about you. This question alone disposes of a great deal.

2. Is that a relative or an absolute change?

The most reliable way to make a small finding sound enormous.

> Suppose something happens to 2 people in 1,000, and an intervention takes it to 1 in 1,000. > As a relative change that is a halving - a 50% reduction. > As an absolute change it is one person in a thousand.

Both descriptions are accurate. Only one of them is how it will be reported. When you see a large percentage with no underlying counts, you are looking at a relative change with the denominator left out.

3. Am I reading the paper or the press release?

Usually the press release, filtered through a journalist working at speed. University press offices write to be picked up. The gap between the abstract and the headline is where most of the distortion lives, and it is often introduced before any reporter is involved.

4. Does the wording say "linked to"?

"Linked to", "associated with", "tied to" and "may be connected to" all mean the study observed two things occurring together. That is a real and useful kind of finding. It is not a finding that one caused the other, and the headline verb will frequently suggest otherwise.

5. Compared with what?

A result needs a comparison to mean anything. Better than nothing, better than the standard approach, and better than a placebo are three completely different claims. If the comparison is missing, the claim is not yet a claim.

6. Can I get to the source in four clicks?

Try it. Find the researcher's name or the journal, search for the title, open the abstract, read the "participants" line and the "limitations" line.

If the article does not name the study at all, that is your answer, and you have spent forty seconds finding it out.

What this means for us

We hold ourselves to the same test. We publish no statistics of our own, we quote no effect sizes, and where a number matters we either show the arithmetic on the page or link you to the body that published it. Our standards page sets out that rule in full.

The limits of this piece

Settled enough to state

How to find what a piece of health news is actually based on, and how relative and absolute change differ enough to reverse the impression a headline gives.

Depends on you

Whether any particular finding applies to you. Who was studied is the question, and the answer is usually not people like everyone.

Not something a webpage can answer

Whether a specific claim is true. This page teaches you to find the source; it cannot referee the science for you, and neither can any other webpage.

Corrections and clarifications are published here, dated. There are none on this page. Write to contact@nouraro.com if something is wrong.