So I've known $MLE$ for ${\sigma}^2$ is $\hat{{\sigma}^2}=\frac{1}{n}\sum_{i=1}^{n} (X_{i} -\bar{X})^2$, and I'm looking for $MSE$ of $\hat{{\sigma}^2}$. But I'm having trouble to get the result.

What I tried goes like below:

By definition, $MSE$ = $E[(\hat{{\sigma}^2}$ - ${\sigma}^2$)$^2$], which is = $Var(\hat{{\sigma}^2} - {\sigma}^2)+(E(\hat{{\sigma}^2} - {\sigma}^2))^2$ = $Var(\hat{{\sigma}^2})-Var({{\sigma}^2})+(E(\hat{{\sigma}^2} - {\sigma}^2))^2$.

From here, I tried to find $Var(\hat{{\sigma}^2})$, which is = $Var(\frac{1}{n}\sum_{i=1}^{n} (X_{i} -\bar{X})^2$) = $\frac{1}{n^2}Var(\sum_{i=1}^{n} X_{i}^2 -n\bar{X}^2)$ = $\frac{1}{n^2}(\sum_{i=1}^{n} Var (X_{i}^2) -n^2Var(\bar{X}^2))$

But I'm not sure how to get $Var (X_{i}^2)$ and $Var(\bar{X}^2)$. I tried $Var (X_{i}^2)$ = $E(X_i^4) - (E(X_i^2))^2$, But I'm not quite sure what $E(X_i^4)$ would be.

Could anyone help me with this? Am I on the correct path to solve this? Thanks in advance!


With $\displaystyle \widehat{\sigma^2} = \frac 1 n \sum_{i=1}^n \left( X_i - \overline X \right)^2$ you have $\displaystyle\frac{n\widehat{\sigma^2}}{\sigma^2} \sim \chi^2_{n-1},$ so $$ \operatorname{var}\left( \,\widehat{\sigma^2} \, \right) = \frac{\sigma^4}{n^2} \operatorname{var}(\chi^2_{n-1}) = \frac{\sigma^4}{n^2}\cdot 2(n-1). $$


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.