Written by Stephen Fox last updated

[A] p-value is the probability of our observing the sample data we have observed when, in fact, the two population averages are identical. - Anthony Davies, Understanding Statistics (2017)

An Intuitive Explanation

Imagine you are comparing two observations (a la some difference of means test): the mean and standard deviation of employment rate in California and New York, USA.

You can think of the p-value as the probability that the apparent difference in employment rate between California and New York is due to random chance.

Think about this:

  • The higher that probability, the more likely any disparity can be (effectively) written off.
  • The lower that probability, the more significant the difference is and more you should consider investigating further.

Closely related to this idea is statistical significance, where a p-value of 0.05 is typical and 0.01 is used in more stringent trial.

Mathematical explanation

link into a statistics learning path

equation from mean-value test