Type B Uncertainty
Definition
Uncertainty evaluated by means other than statistical analysis of repeated measurements. Sources include calibration certificates, instrument specifications, reference data, and expert scientific judgement. Assigned a probability distribution (rectangular, normal, or other) and converted to a standard deviation equivalent.
- Evaluation method
- Non-statistical, e.g. calibration certificates, specifications
- Distribution assigned
- Rectangular, normal, or other, per source
- Output
- Converted to a standard-deviation equivalent
- Pairs with
- Type A uncertainty in combined uncertainty budget
Common questions
What is a common example of a Type B uncertainty source in a forensic laboratory?+
A pipette or balance calibration certificate stating a tolerance, or a reference standard's certified purity range, both of which are converted to a standard uncertainty using an assumed distribution rather than by repeating a measurement.
Why does the choice of distribution matter for a Type B component?+
A rectangular distribution divides the stated half-width by the square root of three while a normal distribution divides by the stated coverage factor, so picking the wrong distribution shape over- or understates the contribution to the combined uncertainty.
Related terms
- Combined Standard Uncertainty
- The overall standard uncertainty of a result, obtained by combining all Type A and Type B component uncertainties in quadrature (square root...
- Expanded Uncertainty
- The combined standard uncertainty multiplied by a coverage factor k, chosen to achieve a stated coverage probability. For a normal distribution, k...
- GUM
- Guide to the Expression of Uncertainty in Measurement. The international standard, maintained by eight metrology and standardisation bodies, that defines Type A...
- Type a Uncertainty
- Uncertainty evaluated by statistical analysis of a series of observations. The standard uncertainty is the standard deviation of the mean (standard error)...
- Z-Score (Proficiency Testing)
- A performance statistic in PT defined as z = (x - X) / s_PT, where x is the laboratory's result, X is...