I asked this question:
Compare the variances of several groups
The Levene's method only test all groups. How can I identify the groups that have a different variances?
I asked this question:
Compare the variances of several groups
The Levene's method only test all groups. How can I identify the groups that have a different variances?
Instead of using some magic black box tests, do it rigorously. You can find the answer in excelent article from Larry Bretthorst, "On The Difference In Means".
http://bayes.wustl.edu/glb/diff.pdf
Unlike what the title says, the paper is also about the difference and the magnitude of difference between stabdard deviations of two samples.