1. Introduction
For deciding whether two samples can reasonably
be considered to origin from the same distribution
(the null hypothesis of ‘there is no
systematic difference’), a number of statistical
tests can be used. If such a test results in a fairly
large probability that the samples come from one
and the same distribution, the null hypothesis is
not rejected. If the test results in a small probability,
usually 5% or smaller, the null hypothesis is
rejected and the alternative hypothesis (‘the samples
origin from different distributions’) is accepted.
The difference is then said to be
statistically significant at the 5% level.