# Homework Help: Finding the pdf of the average of n independent random variables

1. Feb 1, 2012

### epr2008

1. The problem statement, all variables and given/known data
The n random variables $X_{1}, X_{2},..., X_{n}$ are mutually independent and distributed with the probability density

$f(x)=\frac{1}{\pi(1+x^{2})}$

i) Find the probability density of the average

$Y=\frac{1}{n}\Sigma^{i=1}_{n}X_{i}$

ii) Explain why it does not converge toward the normal distribution,as would be expected from the central limit theorem.

2. Relevant equations

$F_{Y}(y)=\int\int\ldots\int_{Y=\frac{1}{n}\Sigma^{i=1}_{n}X_{i}}f_{X_{1}}(x_{1})f_{X_{2}}(x_{2}) \ldots f_{X_{n}}(x_{n})dx_{1}dx_{2}\ldots dx_{n}$

3. The attempt at a solution

i)
$F_{Y}(y)=\frac{1}{(2\pi i)^{n}}\int\int\ldots\int_{Y=\frac{1}{n}\Sigma^{i=1}_{n}X_{i}}(\frac{1}{\frac{x_{1}}{n}+i}-\frac{1}{\frac{x_{1}}{n}-i})(\frac{1}{\frac{x_{2}}{n}+i}-\frac{1}{\frac{x_{2}}{n}-i}) \ldots (\frac{1}{\frac{x_{n}}{n}+i}-\frac{1}{\frac{x_{n}}{n}-i})dx_{1}dx_{2}\ldots dx_{n}$

That is as far as I got because I don't know how to integrate over a hyperplane...

ii)
I Haven't gotten to this but I am guessing that the reason that it does not converge is because the variance is negative or something long those lines.

I'd appreciate any help, whether there's a simpler way to do part (i) or I did something wrong in it.

2. Feb 1, 2012

### jbunniii

Do you know the answer to this general question: If you have two independent random variables, say X and Y, then what is the PDF of their sum X + Y in terms of the PDFs of X and Y?

3. Feb 2, 2012

### 80past2

do you recognize what this distribution is? and the properties of this distribution?

4. Feb 2, 2012

### epr2008

jbunniii: Yes I know that it would be the convolution of the pdfs and the integral above can be turned into an convolution integral. What I am saying is that I don't have a clue how to do so for n variables.

80past2: I honestly have no idea what this distribution is. I haven't taken stat since my freshman year and it really never interested me much, so although I got good grades, my knowledge of statistics is not very thorough. See I was just a math major but I doubled up with physics and am in my 4th year of college right now. Right now I am taking a thermal and statistical mechanics class and this is honestly the first time I have been trumped up on the math in a physics class.

5. Feb 2, 2012

### 80past2

try it for two and see what happens

6. Feb 2, 2012

### epr2008

Oh I figured it out it's a cauchy distribution and all I need to do is show that the product of the characteristic equations for x_i/n gives the characteristic equation for y. Thanks for all the help!