Measure the reliability and precision of sample estimates relative to the mean.
Last updated: March 2026
Relative Standard Error (RSE) is a standardized measure of volatility that expresses the standard error as a percentage of the mean. It quantifies the precision of an estimate—how much the sample mean is likely to vary from the true population mean. RSE is particularly useful because it allows comparison of variability across datasets with different scales or units.
The RSE is calculated by dividing the standard error (standard deviation divided by √n) by the absolute value of the mean and multiplying by 100. A lower RSE indicates a more reliable and precise estimate, while a higher RSE suggests greater relative variability. Survey statisticians and quality-control professionals typically use RSE = 25% as a threshold for acceptable estimate reliability.
Unlike absolute standard deviation which depends on the scale of measurement, RSE provides a unitless, comparable measure. This is why government agencies (like the Census Bureau) and statistical organizations publish RSE values to indicate data quality. An RSE of 5% means your estimate has much tighter confidence intervals than an RSE of 30%.
Survey Data: Monthly Household Income Estimates (n=100 sampled households)
Standard deviation measures absolute spread in data units. RSE expresses it as a percentage of the mean, making it unit-free and comparable across different scales. RSE is better for comparing estimate precision.
RSE and confidence intervals measure precision differently. RSE is simpler to calculate and report (single number), while confidence intervals show the range. RSE is faster for quick reliability assessment.
Government statistical agencies typical use RSE ≤ 25% as acceptable. Values 5-15% are good. <5% is excellent. >25% suggests the estimate may be too unreliable for public use, though context matters.
RSE decreases as sample size increases. It's inversely proportional to √n. Doubling sample size reduces RSE by ~29%. This is why larger samples produce more precise estimates.
Yes, that's RSE's main advantage. You can directly compare precision of estimates from different surveys, different time periods, or different subpopulations using RSE percentages.
For RSE calculation, we use the absolute value of the mean. A mean of zero creates problems—RSE would be undefined. Negative means are handled by taking absolute value before division.
Related Tools
Linear relationship strength.
Linear correlation coefficient.
Rank correlation.
Regression fit quality.
Parabolic curve fit.
Third-degree polynomial fit.