I've looked online but I can't seem to find an answer.
Basically I'm comparing data from two groups, disease and control, and from each induvidual there's two observations. I want to compare the groups to see if one has a larger variance than the other Should I only include one observation from each individual for the test, or both? I think it should be the former but I wanted to be sure.
Thanks.
var.testcertainly assumes independence both within and between groups. I don't think I'd do either of the things you're considering (throwing out half the data or ignoring the obvious dependence). [on other considerations, I wouldn't use the test at all, since it's very sensitive to the normality assumption] – Glen_b Sep 20 '16 at 02:05