Average Value Uncertainty: How to Compute?

[DoobleD]

I have a set of values I got from measurements, and each of these value has an uncertainty associated to it.

I compute the average of my values. How can I average the uncertainty associated to my average value ? Do I simply take the average of all uncertainties ?

 

[BvU]

No. This way you do injustice to accurate measurements and you let bad measurements have too much weight.
There are simple rules for this. The average isn't the best estimate: that is the http://www.colorado.edu/physics/phys2150/phys2150_sp14/phys2150_lec4.pdf.

 

[DoobleD]

No. This way you do injustice to accurate measurements and you let bad measurements have too much weight.
There are simple rules for this. The average isn't the best estimate: that is the http://www.colorado.edu/physics/phys2150/phys2150_sp14/phys2150_lec4.pdf.

Thank you ! I have just read somewhere that indeed this is more precise and used more often in practice when we have a large set of values for a measurement.

 

[DrClaude]

No. This way you do injustice to accurate measurements and you let bad measurements have too much weight.
There are simple rules for this. The average isn't the best estimate: that is the http://www.colorado.edu/physics/phys2150/phys2150_sp14/phys2150_lec4.pdf.

Loved the Dilbert comic in that link.