1. Confirm that CV is appropriate for the replicates
Use at least two numeric observations measured on the same positive ratio scale, where zero has a meaningful absence interpretation. Preserve the source wells, units, response stage, replicate definition, and every included value.
Choose and state the SD convention before calculation. This guide and calculator use sample SD with an n − 1 denominator; changing to population SD changes the result.
2. Calculate sample SD relative to the mean
Calculate the arithmetic mean from the stored observations, then the sample standard deviation, then divide by the absolute mean and multiply by 100. Keep unrounded intermediate values.
The absolute denominator prevents a negative percentage but does not make a negative or near-zero mean scientifically suitable for CV. The calculator returns no percentage when the mean is exactly zero.
%CV = sample standard deviation ÷ |mean| × 100Calculate 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
- 100.0000
- Sample SD
- 1.8257
- CV
- 1.83%
CV describes spread relative to the mean. Whether that variability is acceptable depends on the assay, response level, replicate design, and approved criterion.
3. Carry 98, 102, 101, and 99 to 1.83%
The mean is 100. The squared deviations sum to 10, so sample variance is 10 ÷ 3 = 3.333333 and sample SD is 1.825742. CV is 1.825742 ÷ 100 × 100 = 1.825742%, displayed as 1.83%.
The calculator opens with these same four observations. A report should show n, mean, sample SD, CV, units, rounding rule, and the individual values or a direct path to them.
| Statistic | Worked value |
|---|---|
| n | 4 |
| Mean | 100.000000 |
| Sample SD | 1.825742 |
| Unrounded CV | 1.825742% |
| Displayed CV | 1.83% |
4. Interpret CV as relative precision only
A lower CV means the selected replicates are closer relative to their mean. It does not prove accuracy, lack of bias, curve comparability, or acceptability of the plate.
When a criterion applies, display the observed value, exact rule, unit, and PASS or FAIL outcome. The threshold must reflect response level, replicate design, assay performance, intended use, and the consequences of variability.
Limits and alternative variability summaries
As the mean approaches zero, small denominator changes produce very large and unstable percentages. Background-subtracted or normalized responses that cross zero often require SD or another method-defined absolute measure instead of CV.
One replicate CV does not separate variation from plates, days, analysts, instruments, or reagent lots; use an appropriate variance-component or robustness design. There is no universal CV threshold that applies to every response and assay.
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.