nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Nonparametric Tests: Kolmogorov Smirnov test for variance testing

robert_reviere

Hello:

I have multiple (large data) distribution that I have already tested the means and determined that they are the same and roughly zero. This is what I expected based on subject expertize. Two of the distribution have larger variations than my other distribution. I see this graphically and based on the summary statistics. I want to be more formal and test for differences in the variances. I was thinking of doing Kolmogorov Smimov testing; however, I could not figure out how to do this for my 12 distributions simultanously. Should I even be using the KS test for this purpose? I do not want to be tied to normal distribution testing. Suggestions are appreciated.

Rob

Accepted answers

Mark_Bailey

Generally speaking, the non-parametric tests help you decide if two or more populations are different with regard to any parameter. They are not specific to a particular parameter, as are tested with the parametric tests.

All comments

txnelson

JMP provides several statistical tests for comparing the variances between multiple groups. Under Fit Y by X (oneway), you can select "Unequal Variances" from the red triangle, and JMP will provide a list of stats that will give you a better option than KS for testing the variances.

Mark_Bailey

robert_reviere

Thanks.