About ANOM for variances

[#1]

Hi,

I'm now studying the ANOM for variances. I don't understand why it needs at least three groups to start the comparison. Does anyone know the reason behind this? 

Thank you!

[#2]

I think this link will answer your questions

Karen Copeland, Ph.D. Boulder Statistics, Boulder, CO 80304https://community.jmp.com › kvoqx44227 › ...

Dr. Copeland says, "

In the case of just two treatments the student’s t-test is often the appropriate method to test for a treatment effect. However, using multiple applications of the t-test to test multiple comparisons within a single study is not appropriate. Rather, one should choose a multiple comparison method suited for the particular question your data is to address."