Confidence Interval for the Coefficient of Variation (CV)

[#1]

(Please note that this script is a new version that provides an optional By role as of 10Jul2014.)

This script opens a Distribution platform and adds the confidence interval for the coefficient of variation to the report. Simply open the data table with the columns to be analyzed and then open and run the script. Select the columns for analysis and click Y, Response. Select the grouping column and click By, Group if necessary. Specify the significance level and click OK:

The result is presented in a separate outline at the bottom of the platform:

(Please note that some sections are closed merely to keep this image small.)

The computation is based on this reference:

Vangel, Mark G. (1996) Confidence Intervals for a Normal Coefficient of Variation, The American Statistician 15(1) 21-26.

[laura_zambianch]

Very useful for gage studies, thanks!

BR,

Laura

[Ncocaro]

Dear Mark,

Thanks a lot for the script, it's very good one.

However, for the CI of the CV we prefer to use Miller's confidence interval procedure since it has wider coverage, specially when n<10. Would it be possible to add the method to the script?

Thanks in advance

[FrankDG]

Hi Mark,

very useful for asessing analytical chemistry variances vs concentration levels, which are nearly always heteroscedastic.  It would be good to be able to assess significance of the difference in CV for the levels.  Would it be enough to look at the confidence intervals and look for overlaps?  A formal test included in the script would be very appreciated.

thanks!