An improved procedure for combining Type A and Type B components of measurement uncertainty

^⋆ Correspondence: robin.willink@gmail.com

Received: 26 October 2012
Accepted: 22 November 2012

Abstract

The Guide to the Expression of Uncertainty in Measurement describes a method of evaluating measurement uncertainty that involves attributing an underlying variance to each component of the total measurement error. Assessment of such a variance by statistical methods is known as Type A evaluation of uncertainty, while assessment by any other means is known as Type B evaluation. The involvement of degrees of freedom with Type A evaluation is in accordance with statistical theory. However, the use of “degrees of freedom” with Type B evaluation is an idea primarily developed for practicality, and might be improved upon. That idea ignores the fact that the underlying distribution of the corresponding error will often be assumed to be blunter than normal, as with a uniform distribution, and it can lead to underestimation of the total error variance. Consequently, an alternative concept for Type B evaluation is presented. The resulting analysis fits easily with the use of degrees of freedom in Type A evaluation and leads to an overall method of evaluation that often permits the proper quotation of shorter uncertainty intervals.