2
$\begingroup$

This is my first post to this great website :) It seems like an excellent place to learn. I have a question however that is bothering me as I cannot figure it out through my textbook.

The sample variance $$ S^2 = \frac{1}{n-1}\sum\limits_{i=1}^{n}\left(X_i - \bar{X}\right)^2 $$ Of course, we need to assume that $X_i$ is normal but here is where I am very confused. The next step in my notes shows this:

$$X_i - \bar{X} = \sigma [(X_i - \mu) / \sigma - (\bar{X} - \mu) / \sigma]$$

Where did $\sigma$ come from? And can anyone help me with the proof of this? It appears that the distribution of the sample variance is that of a Chi-Squared R.V. with $n-1$ DoF. Thanks so much!

$\endgroup$
2
$\begingroup$

The statement you've got is true for every $\sigma$ that isn't zero, since you multiplied by $\sigma/\sigma=1$ and that doesn't change the value. $\sigma$ is probably going to end up representing some sort of standard-deviation-like object, but for now it doesn't really matter.

To be a bit clearer, you can simplify the right-hand side into the left-hand side by combining the fractions (which cancels the $\mu$) and then multiplying the $\sigma$ through (which also cancels).

$\endgroup$
2
$\begingroup$

Your second display equation has nothing to do with a normality or any other assumption. As long as $0<\sigma<\infty$ it holds. All you're doing is to multiply by $\sigma$ and to divide by $\sigma$ and to add and subtract $\mu$.

You do need the normality assumption, however, to establish the $\chi_{n-1}^2$ distribution. Use the fact that the sum of the squares of $n-1$ independent normals has a $\chi_{n-1}^2$ distribution.

$\endgroup$

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.