Re: Confidence Limit

Jo wrote in news:[EMAIL PROTECTED]:

> "Dirk Van de moortel" <[EMAIL PROTECTED]>
> wrote in message news:[EMAIL PROTECTED]
>>
>> from which you can easily algebrate and calculate mu1 and mu2 ;-)
> 
> Thanks for the answer. I've worked out how to do most of it, and I
> have an equation to use.
> 
> The equation is CL (lower) = xbar - z(st.dev/sample size)
> 
> It have applied it fine in case where z has been supplied. What I need
> to know though is how do I get the value of z? 'z' is defined as "the
> (Normal) value such that P(-z < Z < z) = CL.
> 
> I have no clue how to get it - I tried to get it via the Inverse
> General Normal table but I couldn't match the value that was supplied
> to me. 
> 
> e.g. Confidence Level 0.90 then z = 1.645
> 
The usual method: You go to a table of the CDF of normal distribution, 
N(0,1). (I think the CDF is generally easier to find than the Inverse 
table.)  It is generally laid out with values of x and Phi(x) and centered 
on 0 so Phi(0)=0.500 . Phi(x) is the Normal CDF. You find the value of x 
that gives you the needed value of Phi(x), in your case 0.95, since you 
want centered 90% limits. The value of x that does that is 1.645

It should be possible to do it with the Inverse table, though.
 Phi^-1(1.645) should equal 0.95

-- 
David Winsemius