The below homework question comes from Larsen and Marx, 4th edition.

Is the maximum likelihood estimator for $\sigma^{2}$ in a normal pdf, where both $\mu$ and >$\sigma^{2}$ are unknown, asymptotically unbiased?

I think I understand the notion that an estimator $\hat{\theta_{n}}$ is unbiased if the limit of its expected value as n goes to infinity is $\theta$, but I'm really not sure where to go with trying to answer the above question.

Any idea where to start?

  • $\begingroup$ Please make your question self-contained. $\endgroup$ – Did Jul 23 '12 at 16:44

If $X_1,\ldots,X_n$ are iid $\mathcal{N}(\mu,\sigma^2)$ variables, then the maximum likelihood estimator of $\sigma^2$ is given by $$ \hat{\sigma}_n^2=\frac{1}{n}\sum_{i=1}^n(X_i-\bar{X})^2, $$ where $\bar{X}=\frac{1}{n}\sum_{i=1}^n X_i$ is the average. Recall that we usually do not use this estimator because it is biased. Instead we will often use $$ s^2_n=\frac{1}{n-1}\sum_{i=1}^n (X_i-\bar{X})^2, $$ because $E[s^2_n]=\sigma^2$. Now because $\hat{\sigma}^2_n = \frac{n-1}{n}s^2_n$ it follows that $$ E[\hat{\sigma}_n^2]=\frac{n-1}{n}\sigma^2\to \sigma^2 $$ as $n\to\infty$ and hence the maximum likelihood estimator is asymptotically unbiased.


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.