0
$\begingroup$

Apologies for a silly question - may not make sense. If I have two measurements (assumed to come from normal distribution) X1 = 250 and X2 = 260 where variances are V1 = 5 and V2 = 6 (variance could be thought as the uncertainty in the measurements). Now the average of the two is m = 255 but what is the resulting variance of this average? Is it possible to calculate this?

$\endgroup$
0
$\begingroup$

It's not a silly question, but the concepts should be clear. If we denote your random variables as $X_1$ and $X_2$, the average will be $\bar{X}=\frac{X_1+X_2}{2}$, and we can find the variance of it via a simple calculation noting that the two RVs are independent: $$\operatorname{var}\left(\frac{X_1+X_2}{2}\right)=\frac{\operatorname{var}(X_1)+\operatorname{var}(X_2)}{4}=\frac{5+6}{4}=\frac{11}{4}$$ This calculation has nothing to do with the means of $X_1,X_2$, as well as the measurement values.

Informally, variance of an unknown $X_1$ can be non-zero. However, variance of a known $X_1$ is $0$ because it is known and therefore constant. So, variance of $X_1=250$ will be $0$, and it actually means something like $\operatorname{var}(X_1|X_1=250)$, mathematically, which is $0$. Measurement random variables have non-zero variance, but the measurement itself is fixed and have no variance (e.g. it is 250 always since you measured it).

$\endgroup$

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.