A -value is the probability,
under a specified null hypothesis and its statistical
model, of obtaining a test statistic at least
as extreme as the observed value. The meaning of extreme is fixed by the test and
its alternative hypothesis (Orloff and
Bloom 2022, pp. 1-4). For a random variable
representing the test
statistic, with observed value
, an upper-tailed test uses
|
(1)
|
where denotes probability
calculated under the null hypothesis. A lower-tailed
test uses
|
(2)
|
For a symmetric null distribution centered at 0, a conventional two-sided test uses
|
(3)
|
Thus a -value need not be an upper-tail probability.
It does not give the probability that the null
hypothesis itself is true (Wasserstein and Lazar 2016, p. 131).