Hands-on Workshop on Assessing and Reporting Measurement Uncertainty
Purpose This NIST short course covers the propagation of measurement uncertainty using the methods outlined in the JCGM Guide to the Expression of Uncertainty in Measurement from a statistical perspective. The short course will provide participants with a working knowledge of the computational methods needed to assess measurement uncertainty, hands-on experience in the application of these methods, and scientific and statistical insight into the interpretation of the results.
Agenda The Hands-on Workshop on Assessing and Reporting Measurement Uncertainty is a 3-day course that will be held at the Measurement Science Conference in Anaheim, CA. The course consists of lectures, short exercises, and hands-on applications covering many aspects of the propagation of uncertainty using examples from NIST work.
The exercises and hands-on applications will use functions for uncertainty analysis from the software package, metRology, written for the open-source R statistical computing environment. The functions can be accessed directly in R (use of RStudio is recommended), or via an Excel graphical user interface that is available as a free Add-In, metRology for Microsoft Excel. Participants should bring their own laptops, if possible. A laptop for use during the short course can be provided (sharing may be required). If you would like to borrow a laptop, please let one of the instructors know as soon as possible. All software except Microsoft Excel is free.
- Importance of uncertainty analysis
- Different statistical approaches for uncertainty analysis
- Essentials of the GUM approach
- Measurement functions
- Type A and Type B methods for evaluating standard uncertainties
- Degrees of freedom
- Sensitivity coefficients
- Propagation of standard uncertainties
- Effective degrees of freedom
- Expanded uncertainties
- Software for propagation of uncertainty
- Interpretation of results
 Please note: the metRology for Microsoft Excel Add-In is not compatible with Microsoft Excel for Macintosh, but should work on the Mac with full Windows emulation. The metRology package can be used directly in R or RStudio on Windows, MacOS, or Linux.