I want to compare the variability of two log-normal samples. They are being summarized as Coefficient of variations (CVs) and I want to test if one sample's CV is different from the other's.

To compare the geometric means, I performed a t-test on log-transformed data, then applied the antilog to go back to the original scale of the problem.

  1. In the same logic, is it appropriate to apply the F-test for variances on log-transformed data and then revert to the antilog scale?

  2. If not, is there a significance test for CVs or geometric CVs?

  • $\begingroup$ Maybe you should mention that CV here is coefficient of variation and not cross validation. $\endgroup$ – Michael R. Chernick Jan 22 '17 at 15:50
  • $\begingroup$ @MichaelChernick thanks, edited in first occurance and title $\endgroup$ – philsf Jan 22 '17 at 15:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.