Back to Measurement, Uncertainty and Detection Limits

Standard Uncertainty Classes

Standard uncertainties u(xi) are here classified as either Type A or Type B: Type A is a statistical evaluation based on repeated observations.
One typical example of a Type A evaluation involves making a series of independent measurements of a quantity Xi, and calculating the arithmetric mean and the experimental standard deviation of the mean.
The arithmetric mean is used as the input estimate, xi, and the experimental standard deviation of the mean is used as the standard uncertainty u(xi).
Any evaluation of standard uncertainty that is not a Type A evaluation is a Type B evaluation.

Expanded Uncertainty

A measurement may be reported with the combined standard uncertainty uc(y) or it may be multiplied by a coverage factor, k, to produce an expanded uncertainty denoted U, such that the interval y ± U has a specific high probability p of containing the true value of the measurand.
The specific probability p is called the level of confidence or the coverage probability and is generally only an approximation of the true probability of coverage.
The coverage factor often chosen for approximately normal distributions of y is k = 2. If uc(y) represents one standard deviation, U then corresponds to a confidence level of 95 %.

Significant Figures

The number of significant figures of a measured result y depends on the uncertainty of the result. Since there is uncertainty also in the uncertainty estimates, a common convention is to round off the uncertainty to one or two significant figures.
Examples:
Measured Value y
Expanded Uncertainty U=kuc(y)
Reported Result
15.235
0.121
15.24±0.12
15.235
1.213
15.2±1.2
15.235
12.134
15±12
15.235
121.343
20±12
15.235
1213.432
0±1200

Procedure for Estimating Uncertainty