How to Read a Nutrition Study Without Being Misled
Six questions that separate a finding worth acting on from a headline. Useful for everything you'll read, including this site.
Nutrition headlines are unusually unreliable, and the reason is structural rather than malicious. Here's how to evaluate one in about ninety seconds.
1. What kind of study is it?
Randomized controlled trial — people assigned to different conditions. The strongest design for cause and effect, usually small and short.
Cohort study — people followed over years, associations observed. Large and long, but cannot establish cause. Most nutrition headlines come from these.
Mechanistic or animal study — explains how something might work. Frequently reported as though it applies to humans. It usually doesn't yet.
If the headline says a food "causes" something and the study is a cohort, the headline has overstated it.
2. How many people, and for how long?
A two-week trial in nineteen young men tells you something. It doesn't tell you what a food does to a 62-year-old woman over a decade.
Small and short is not worthless — it's preliminary.
3. What was actually measured?
Outcomes people care about — heart attacks, diabetes diagnoses, deaths.
Surrogate markers — cholesterol, insulin, inflammation markers.
Surrogates are faster and cheaper to measure and they don't always translate. A study that improved a marker has not shown it improved anyone's life.
4. Relative or absolute?
"Doubles your risk" sounds alarming. If the risk went from 1 in 10,000 to 2 in 10,000, that's a doubling and it's a very small absolute change.
Almost all frightening health headlines are relative risk. Ask what the absolute numbers were.
5. Compared with what?
A diet that "improved outcomes" improved them relative to something. Often the comparison is a standard Western diet, which is a low bar.
6. Who funded it, and does that matter?
Funding source is worth knowing and isn't disqualifying on its own. Industry-funded research is more likely to find favorable results, and plenty of it is well conducted.
Be more skeptical when the funder sells the thing being tested and the finding is surprising.
The pattern worth noticing
The strongest claims usually rest on the weakest evidence. A confident, dramatic assertion about a single food is nearly always built on a mechanism, an animal study, or a small trial.
Well-supported findings tend to sound boring: eat more vegetables, move after meals, sleep enough.
Applying it here
Every research claim on this site links its source. If a claim ever appears without one, that's a signal — and it's fair to apply these six questions to anything written here.
Try this week: next time you see a nutrition headline, find the study and check questions 1 and 4. It takes two minutes and it changes what you believe.
Education, not medical advice.