Overview

Relative standard error, RSE, shows how precise a survey estimate is, as a percentage of the estimate. Statistical agencies use it to flag unreliable figures.

Enter the sample standard deviation, the sample size and the sample mean.

Worked examples

InputsResult
Sample standard deviation12Sample size (n)64Sample mean50Relative standard error: 3%Standard error: 1.5
Show 2 more examples
Sample standard deviation5Sample size (n)25Sample mean20Relative standard error: 5%Standard error: 1
Sample standard deviation30Sample size (n)400Sample mean120Relative standard error: 1.25%Standard error: 1.5

Relative standard error by sample mean

With the calculator's starting values, the relative standard error is 3%. The table shows how it changes with the sample mean.

Sample meanRelative standard error
305%
403.75%
50starting value3%
602.5%
702.14%

Other inputs: Sample standard deviation: 12 · Sample size: 64 n. Change them in the calculator above for your own numbers.

How it works

The standard error is the standard deviation divided by the square root of the sample size. RSE divides that by the size of the mean and multiplies by 100. Lower is more precise. Many agencies treat an RSE above about 25 to 30% as unreliable.

The formula

RSE = (s ÷ √n) ÷ |mean| × 100
sSample standard deviation
nSample size

Practical tips

  • s = 12, n = 64 and a mean of 50 give a standard error of 1.5 and an RSE of 3%.
  • Quadrupling the sample halves the RSE.
  • RSE isn’t useful when the mean is close to zero.

Frequently asked questions

Is this the same as residual standard error?

No. In regression, RSE often means residual standard error, √(SSE ÷ df). Our MSE calculator gives that figure as RMSE.

How is RSE different from the coefficient of variation?

CV uses the standard deviation of the data. RSE uses the standard error of the mean, so it shrinks as the sample grows.