Combining Results and Errors - A Forum Saviour!

[fasterthanjoao]

the following post probably isn't going to be so cohesive, bear with me anyway: I'm completing a report and need to combine several sets of results for the same experiment and combine their errors to give a best value - can't for the life of me remember the formulae, i know roughly what it is - sum of the squares of the values over the squares of their error, all over the square of their error...

anyway, thanks as usual!

Posting/reading on this forum should become compulsory for all undergrads, its helpful.

 

[Meir Achuz]

For values y_i with errors d_i:
<y>=Sum[y_i/d_i^2]/Sum[1/d_i^2].
The resulting error is 1/D^2=Sum[1/d_i^2].