2. Uncertainty of Measurement
(UoM)
• A very interesting and important subject. As
practicing analysts, we should understand this.
• For me personally, I have found it gives a deeper
understanding of what we are doing
• Enables us to answer the question, is the method
and result ‘fit for purpose’?
• ISO Guide 17025 states we must be able to provide
UoM data in circumstances where it may be critical.
3. Expressing UoM
UoM usually expressed
as the standard
deviation, or Co-
efficient of Variation.
Sometimes Expanded
UoM. Usually x2, which
for normally distributed
data, corresponds to the
95% Confidence Limit.
4. Determining UoM
• ‘Bottom Up’ approach
- Also called Uncertainty Budget
- ‘Fishbone’ approach
- Considerable amount of work
• ‘Top Down’ approach
- Uses real data, e.g. from QC samples
- Need to understand just what is covered
- Can be estimated from control charts simply
• Combination best but
7. Uncertainty (U %) =
100 ´ (UWL- Mean)
Mean
Top down estimate of uncertainty
8. Summary
• Understanding UoM is important to improve your
service, and to ensure the tests are ‘fit for
purpose’.
• It makes you a better analyst, by having a deeper
understanding of your tests.