CV puts standard deviation on a relative scale

For replicate values, calculate the mean and a declared standard deviation convention. Analytical replicate summaries commonly use sample standard deviation, dividing squared deviations by n − 1. This calculator divides sample SD by the absolute mean and multiplies by 100.

Using the absolute mean prevents a negative sign, but it does not make CV scientifically meaningful for negative or near-zero responses. State the convention and review whether the measurement has a meaningful ratio scale.

Formula%CV = sample standard deviation ÷ |mean| × 100
Interactive calculation

Calculate replicate CV

Enter responses separated by commas, spaces, or line breaks. The calculator uses sample standard deviation (n − 1) and divides it by the absolute mean.

n
4
Mean
99.8500
Sample SD
1.1561
CV
1.16%

CV describes spread relative to the mean. Whether that variability is acceptable depends on the assay, response level, replicate design, and approved criterion.

Worked calculation: 98, 102, 101, and 99

The mean is 100. The sum of squared deviations is 10, so the sample variance is 10 ÷ 3 = 3.333. The sample standard deviation is 1.826, and the CV is 1.826 ÷ 100 × 100 = 1.83%.

Retain more precision during calculation and round only the displayed result. A report should show n, mean, SD, CV, units, and individual observations or a path to them.

StatisticValue
n4
Mean100.000
Sample SD1.826
CV1.83%

A near-zero mean makes CV unstable

When the mean approaches zero, small changes in the denominator produce very large CV values. CV is generally meaningful for ratio-scale data with a meaningful zero and positive values. Background-subtracted or normalized responses that cross zero often need an absolute variability measure instead.

CV also does not tell you whether variability comes from within a plate, between plates, days, analysts, instruments, or reagent lots. Variance-component designs are needed to separate those sources.

Avoid a universal CV threshold

A criterion should reflect assay level, endpoint, replicate design, intended use, historical performance, and the consequences of variability. Low-signal controls may need an SD rule while higher signals use CV.

Display both observed CV and configured criterion, and retain the method version that supplied the limit. A PASS/FAIL color is helpful only when the numeric evidence and rule remain readable.

Frequently asked questions

Should CV use sample or population standard deviation?

Declare the method. For a sample of replicate observations, sample SD is common; a fully enumerated population is different. Mixing conventions can change borderline outcomes.

Can I calculate CV when the mean is negative?

The interpretation is usually problematic because CV assumes a meaningful ratio scale. Review the response transformation and consider SD or another assay-appropriate measure.

Does a low CV prove an assay is accurate?

No. CV describes precision relative to the mean. Replicates can agree closely and still be biased or affected by systematic error.

Primary references