5
$\begingroup$

I'm trying to prove that the sample variance is an unbiased estimator.

I know that I need to find the expected value of the sample variance estimator $$\sum_i\frac{(M_i - \bar{M})^2}{n-1}$$ but I get stuck finding the expected value of the $M_i\bar{M}$ term. Any clues?

I would also like to calculate the variance of the sample variance. In short I would like to calculate $\mathrm{Var}(M_i - \bar{M})^2$ but again that term rears its ugly head.

$\endgroup$
  • 1
    $\begingroup$ Assuming the $M_i$ are iid, hint: let $\mu = E(M_1)$. Write $\sum (M_i - \bar M)^2 = \sum (M_i - \mu)^2 - n(\bar M - \mu)^2$ by throwing $\pm \mu$ into the difference and expanding. $\endgroup$ – guy Apr 3 '12 at 3:43
  • 2
    $\begingroup$ See the first part of my answer here: math.stackexchange.com/questions/72975/… $\endgroup$ – user940 Apr 3 '12 at 4:01
6
$\begingroup$

I know that during my university time I had similar problems to find a complete proof, which shows exactly step by step why the estimator of the sample variance is unbiased. I took it from http://economictheoryblog.wordpress.com/2012/06/28/latexlatexs2/ but they use a different notation, however I think you can take it from there

$\endgroup$
3
$\begingroup$

The only full and complete proof I could find on the internet can be found under http://economictheoryblog.wordpress.com/2012/06/28/latexlatexs2/

As goes over 70 (!) steps and shows every little formula manipulations I prefer rather not to post it here.

$\endgroup$
0
$\begingroup$

A well explained proof, based on maximum likelihood estimation is explained here: http://www.visiondummy.com/2014/03/divide-variance-n-1/

It first explains how to derive the sample variance formula, and then proofs it is unbiased.

$\endgroup$

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.