I have to check some properties of an estimator, but I can't find its distribution.

Let $X_1,...,X_n $ be independent identically distributed exponential variables with parameter $ \theta $, i.e. with pdf $ f(x) = \frac{1}{\theta} \exp(-\frac{x}{\theta})$. What is the distribution of the sample variance $ \overline{X^2} - (\overline X)^2$?

What I found so far:

$\frac{2n}{\theta}\overline X$ ~ $ \Gamma(n,2)$

But how can I find the distribution of the sample variance?

Any help is greatly appreciated!


Your best bet is to use Monte Carlo simulation: repeatedly simulate $n$ draws from a exponential $\theta=k$ distribution, for some $k>0$, calculate the sample variance, repeat. Do this 10k-30k times to get the sampling distribution for the sample variance.

If you have data, then you can choose k to be the MLE for the exponential rate parameter.

| cite | improve this answer | |

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.