Calculating measurement uncertainty

[AbsoluteZer0]

Hi,

I've been studying uncertainty in measurement. I'm not sure how to decide if the uncertainty of a given measurement should be ±.01 or ±.02 or ±.03, and so forth. I understand that the number of decimal places in the uncertainty calculation should correspond to the number of decimal places in the measurement, but I am not sure as to when I should decide whether the uncertainty contains a 1, 2, 3, and so forth.

Thanks,

 

[chiro]

Hey AbsoluteZer0.

If it's a statistical measure of error, this usually corresponds to some kind of standard deviation or standard error (both refer to different things: one being a population measure typically and the other being a sample statistic typically), but if it's not statistical it may correspond a maximum error bound.

If it's the latter then typically you will figure out what the maximum error is and use that since all values will lie in-between +- that value.

If it's statistical then this is different because what happens usually in this case is that you have +- so many sigma contains a probabilistic proportion that a fraction of the errors according to some constraint will fall in that region and the rest won't.