The one-sentence definition
A p-value answers a narrow question: *if this supplement truly did nothing, how often would we see a result this strong (or stronger) by chance alone?* A p-value of 0.04 means that, assuming no real effect, you would expect a result this extreme about 4% of the time [1].
What 'statistically significant' means
Researchers often call a result 'statistically significant' when the p-value falls below 0.05. That 0.05 line is a long-standing convention, not a law of nature. A result at p = 0.049 is not meaningfully different from one at p = 0.051, even though only one clears the bar [1].
Four things a p-value does NOT tell you
- How big the effect is. A tiny, unimportant change can be 'significant' in a large study. For that, look at the effect size and the [confidence interval](/learn/relative-risk-and-confidence-intervals).
- Whether the result matters in real life. Statistical significance is not the same as practical importance.
- The probability your idea is correct. A p-value is calculated *assuming* no effect, so it cannot tell you the odds that the hypothesis is true.
- That the study was well designed. Bias, small samples, or the wrong [study design](/learn/observational-vs-rct) can produce a low p-value that later fails to replicate.
Why this matters for supplements
Marketing often leans on the phrase 'statistically significant' to imply a product is effective. But significance can be manufactured by testing many outcomes and reporting only the ones that cross 0.05 — a practice critics call p-hacking. NCCIH's guidance on evaluating research stresses looking past a single number to the whole study: its size, its design, and whether independent trials agree [2].
How to read a p-value sensibly
Treat a low p-value as one signal among several. Ask how large the effect was, how many people were studied, whether the finding has been repeated, and who funded the work. A trustworthy claim rests on a body of consistent, well-designed human research — not one p-value.