Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

A Chi-Square random variable is defined as $Χ^2_{df} = \frac{( n - 1 ) s^2 }{σ^2}$.

I cannot find a source for this anywhere on the internet, and my textbook is of no help.

Does anyone know where this formula came from and why it makes up what we call the Chi-Square?

share|improve this question

1 Answer 1

Chi square distribution with $n$ degrees of freedom is the sum of $n$ independent normal distributions.

If $X_i \sim N(0,1) \ ;i=1,\ldots,n$ then $\sum_{i=1}^n{X_i}^2 \sim \chi_n^2$.

It is mainly used for Testing of Hypothesis and regression statistics.

What you are seeing is another way of writing Chi-square. It was introduced to us in a similar manner in probability.

Consider $X_i \sim N(0,\sigma^2 )$. Then $\frac {X_i^2}{\sigma^2} \sim \Gamma (\frac 12,\frac 12)$. Adding we get, $\sum_{i=1}^n \frac {X_i^2}{\sigma^2} \sim \Gamma (\frac n2,\frac 12)$ whic is precisely Chi square distribution.

For more general result,please check Proof of $\frac{(n-1)S^2}{\sigma^2} \backsim \chi^2_{n-1}$

I don't know how the formula originated, but can give you one of major (probably) reasons behind use of this formula. Suppose you want to find $E(s^2)$, then you can use this formula. This leads to chi square test.

Also I think that it is worth mentioning that chi squared is the distribution of Quadratic forms for standard normal.

share|improve this answer
    
You probably meant to sum over $1...n$, not over $0...n$ - otherwise you get $N+1$ df's. –  gt6989b Nov 15 '12 at 17:39
    
I know, I read that on Wikipedia too. My question was, where's the formula from? –  Imray Nov 15 '12 at 17:55
    
IMO chi squared, Student's t distribution, and F distribution, all are from nowhere, they are names given to some functions of random variables which originated during calculations, like in linear regression, ANOVA, etc.They are to some extent derived distributions as opposed to distributions like normal, etc –  TheJoker Nov 15 '12 at 18:13

Your Answer

 
discard

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.