I work with the GUM method to determine measurement uncertainties. As I understand the method, the model created for a particular measurement setup provides an uncertainty for a single measurement.

But normally I repeat such a measurement several times, e.g. I measure the weight of a sample five times and then average the values. With the GUM method, I have an uncertainty for each value:

$$m_1 \pm u_{m,1}$$

$$m_2 \pm u_{m,2}$$

$$m_3 \pm u_{m,3}$$

$$m_4 \pm u_{m,4}$$

$$m_5 \pm u_{m,5}$$

Calculating the mean value is trivial, but how can I take into account the five uncertainties associated with GUM? Here's what I'd like:

$$\bar{m} \pm u_{\bar{m}}$$

I know there are similar questions, but as far as I can see, none in the GUM context.

One of my thoughts is that I simply include random scattering as a normally distributed uncertainty in the budget. However, I could only create the budget after I have taken the measurements. That would also be a kind of cicle conclusion.



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.