There are two ways:

1) Used for two sets of samples

CV= square root of variance/mean

2) Used for three groups of samples or more

CV=square root of experimental error/mean

Thanks for your answer.

I referred to the Coefficiente of variation from duplicate measurements, as you can see in the attached file from the help of Medcalc.

Thanks a lot Laila for your answer!

Pleas read this report

MedCalc offers 3 methods:

• Root mean square method
• Logarithmic method
• Within-subject standard deviation method

Details are given on https://www.medcalc.org/manual/cvfromduplicates.php