Evidence-appraisal glossary

p-value

A p-value is the probability of getting results at least as extreme as those observed if the null hypothesis (no real effect) were true. It measures how surprising the data are under that assumption. It does not give the probability that the hypothesis itself is true or false.

Also called: p.

A p-value quantifies how compatible the observed data are with a specific model, usually one in which there is no real effect. A small p-value means the data would be unusual if that no-effect assumption held; a large one means the data are unremarkable under it. A common threshold is 0.05, but that cutoff is a convention, not a law of nature. When reading a study, ask what null model the p-value tests, and resist reading it as the chance the treatment works or the chance the finding is a fluke. For example, if a trial reports p = 0.03 for a blood-pressure drug, it means results this strong would occur about 3 percent of the time if the drug truly did nothing, not that there is a 97 percent chance the drug works. Also check whether many comparisons were run, since testing many outcomes inflates the odds of a small p-value appearing by chance alone.

This is a plain-language methodology definition for reading research. It is general education, not medical advice.

Back to the glossary