# Statistic Retour

1. Aug 22, 2009

### Mafer

I haven't touch statistics for year and now I came back to find I am totally lost.
Here is one of the question that I wish to know how to solve it, in terms of steps, so that I can gain back memory about it.

I know it is a bit too much, but I am really very very lost and depressed, help.

2. Aug 22, 2009

### mXSCNT

These types of questions are straightforward, you just have to look up the appropriate technique (and get familiar with the terminology). According to
http://www.fmi.uni-sofia.bg/vesta/Virtual_Labs/interval/interval6.html [Broken] :
let S^2 be the sample variance of normally distributed data
let $$\chi^2_{n,a}$$ be the number x such that if X has a $$\chi^2$$ distribution with n degrees of freedom , then P(X < x) = a.
then
$$\left(\frac{n-1}{\chi^2_{n-1,1-a/2}}S^2,\frac{n-1}{\chi^2_{n-1,a/2}}S^2\right)$$
is a 1-a confidence interval for the distribution variance. You can calculate this interval numerically using a table for the $$\chi^2$$ distribution or some stats software.

In part 2, you need to use a one-sample t-test.

If these terms are unfamiliar to you, you just need to review your stats book and look up what you don't remember.

Last edited by a moderator: May 4, 2017