Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Is there an easier way to find the variance of sample average squared $(\bar X^2)$ without using the moment generating function? $X\sim N(\mu, \sigma^2)$

I know that $Var(\bar X^2)= E(\bar X^4) -(E(\bar X^2))^2$. Thus I can find the $E(\bar X^4)$ using the moment generating function, and hence the $Var(\bar X^2)$.

share|cite|improve this question
For the hard part, one can expand $(X_1+\cdots+X_n)^4$ and find the expectation. A bit tedious. – André Nicolas Feb 7 '13 at 3:39
You don't need to expand the fourth power of a sum of $n$ terms, but only of two terms. I'll post an answer. – Michael Hardy Feb 7 '13 at 5:05
up vote 1 down vote accepted

You have $$ \bar X \sim N\left( \mu, \frac{\sigma^2}{n} \right). $$ The density of that distribution is $$ x\mapsto\frac{\sqrt{n}}{\sqrt{2\pi\,{}}\,\sigma} \exp\left( -\frac{(x-\mu)^2}{2\sigma^2/n} \right). $$ Therefore $$ \mathbb E(\bar X^4) = \frac{\sqrt{n}}{\sqrt{2\pi\,{}}\,\sigma} \int_{-\infty}^\infty x^4 \exp\left( -\frac{(x-\mu)^2}{2\sigma^2/n} \right) \, dx. $$ So write \begin{align} y & = \frac{x-\mu}{\sigma/\sqrt{n}}, \\[10pt] dy & = \sqrt{n}\frac{dx}{\sigma}. \end{align} Then the integral becomes $$ \frac{\sqrt{n}}{\sqrt{2\pi\,{}}\,\sigma} \int_{-\infty}^\infty \left(\frac{\sigma}{\sqrt{n}} y + \mu\right)^4 \exp\left( -\frac{y^2}{2} \right) \left( \frac{\sigma}{\sqrt{n}}\,dy\right). $$

When the $4$th power is expanded as the sum of five terms, one gets $$ = \frac{\sigma^4}{n^2\sqrt{2\pi\,{}}} \int_{-\infty}^\infty y^4 e^{-y^2/2} \, dy. $$ Next: $$ \int_{-\infty}^\infty y^3 e^{-y^2/2} \Big( y \, dy \Big) = 2\int_0^\infty \text{ditto} = 2\int_0^\infty u^{3/2} e^{-u} \, du $$ $$ 2\Gamma\left(\frac52\right) = s\cdot\frac32\Gamma\left(\frac12\right) = 2\sqrt{\pi}. $$

One then does the same thing with the other terms in the expansion of the fourth power.

share|cite|improve this answer

Couldn't you use the standard rules for variance calculations? Given that your X:s are IID

enter image description here

We get this many terms

enter image description here

and for Var[xi*xj] we can use Goodman's expression

enter image description here

We put it all together

enter image description here

share|cite|improve this answer
Please use proper syntax and not photos of various sizes. – user88595 May 14 '14 at 15:04

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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