Used to test the differences in the variances between samples, they allow in particular to ensure the conditions of use of the hypothesis tests..
In most cases, Variance tests are used to determine if the Variances between groups are homogeneous and thus to choose the test that we will implement. They can also detect outliers, which is also detrimental to most tests.
But not only. These tests can also be used to compare Variances.
Choosing the Homoscedaticity Test
There are a multitude of adjustment tests, from the simplest to the most complex. Nevertheless, we will retain 2 who are the best performers :
- The test of Fisher-Snedecor to compare 2 variances only. Attention, it is necessary that our data are Normal to have a good performance of the test.
- The test of Brown Forsythe to compare 2 or X Variances. It’s the most robust that allows you to compare as many groups as you want.
There are other tests, Levene, Bartlett … but these are less robust and in practice less used.