Degrees of Freedom for Type B Uncertainty Estimates
For some measurement uncertainty estimates, we may not have a sample of data to work from.  In these cases, we need to estimate uncertainty using our technical experience and knowledge. An uncertainty estimate obtained in this way is called a Type B estimate.  

The built-in Type B Degrees of Freedom Calculator is a tool that helps organize our experience and knowledge in such a way that a Type B uncertainty estimate is obtained in a straightforward and unambiguous manner.  In addition to taking some of the mystery out of obtaining Type B uncertainty estimates, this tool also computes the associated degrees of freedom. This means that Type B uncertainty estimates can be used as statistics, just like Type A estimates.

The Type B Degrees of Freedom Calculator can be accessed from screens and worksheets that deal with the estimation of Type B uncertainties.  The three formats for assembling information are described below. 
 
Containment Limits and Probability

This format reads “Approximately C % ± DC % of measured values have been observed to lie within the limits ± L ± DL.”  The format accommodates cases where the best available information consists of a set of error containment limits (± L) and a containment probability (C %). The 'quality' of the information can be accounted for by indicating 'give-or-take' values for both the containment limits (± DL) and the containment probability (± DC %).
 
Containment Limits and History of Observation
This format reads “Approximately X out of N measured values have been observed to lie within the limits ± L ± DL.”
This format is useful when a variable or the error in a variable can be said to lie within ± containment limits in X out of N cases. The quality of the information can be accounted for by indicating a 'give-or-take' value DL for the containment limits.
 
Containment Limits, Probability and Number of Cases
This format reads “Approximately C % of N measured values have been observed to lie within the limits ± L ± DL.”
This format is useful when a variable or the error in a variable can be said to lie within ± containment limits in C% out of N cases. The quality of the information can be accounted for by indicating a 'give-or-take' value DL for the containment limits.