The p-value of a hypothesis test is the probability of observing a value of the test statistic that is at least as inconsistent with the null hypothesis as the value of the test statistic actually observed.