2
$\begingroup$

I have three sets of data sets each 10 to 20 data points. I want to test if the coefficient of variation is the same across all the sets. Please suggest relevant methods.

I came across papers by Cabras(2006) and Amiri(2010). One, they look at comparing only two samples and two, my knowledge of simulation / econometric methods is elementary. Are there packages in [R] or Stata that are available to do this?

Improve this question
$\endgroup$
4

1 Answer 1

Reset to default
2
$\begingroup$

There is an R package on CRAN for the Feltz and Miller Asymptotic Test: cvequality

That package also has a function for the ‘Modified signed-likelihood ratio test (SLRT) for equality of CVs’ (Krishnamoorthy and Lee 2014).

And it has a very nice vignette showing how to use the functions ;) (disclaimer: I'm the package author)

References:

Feltz, C. J., & Miller, G. E. (1996). An asymptotic test for the equality of coefficients of variation from k populations. Statistics in Medicine, 15(6), 647-658. https://www.ncbi.nlm.nih.gov/pubmed/8731006

Krishnamoorthy, K., & Lee, M. (2014). Improved tests for the equality of normal coefficients of variation. Computational Statistics, 29(1-2), 215-232. http://link.springer.com/article/10.1007/s00180-013-0445-2

Improve this answer
$\endgroup$
2
  • $\begingroup$ Thanks, These tests confirm if the COV are different. If I want to test for CV1 > CV2 > CV3, do we have some version of post hoc tests? Or can I look at the values of CV of the samples and directly infer? $\endgroup$
    – Piyush Shah
    Commented Dec 30, 2016 at 16:05
  • $\begingroup$ You're welcome! I'm not currently aware of a post-hoc test. I'd suggest simply plotting the distributions to infer directly, as you can see in the vignette of the cvequality package $\endgroup$
    – Ben
    Commented Dec 31, 2016 at 10:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.