- just a note on z-tests -

On 3 Nov 2003 07:37:58 -0800, [EMAIL PROTECTED] (Robert J. MacG.
Dawson) wrote:

[ ... ]
> 
>       Ninety-nine times out of 100 when people refer to a "Z test" they do
> not mean the theoretically-justified-but-very-rare test with known
> variance, but a badly-done t test, in which quantiles of Z are used
> where those of t should be, that was once taught as "the right thing to
> do" when n was greater than 30.

Who does that?  
Is it just the old folks went to school  35 years ago?

Here is where I see expect the z-test, and it is appropriate, too.

Consider all those non-parametric statistics, which so often
reduce down to a 1-df  chisquared.  It is 1-df  chisquared, 
because it is Normal at the root.    X2=  Normal-squared.   
You often can describe, if it matters to you, the 'one-tailed'
version  of the test.    Naturally, that will be  Normal.

Unlike the situation with 'normally distributed data,'
the *variance is known*  when you have a binomial, 
or multinomial, or rank-order distribution with no ties.  
That is why the subsequent test can be z, rather than t.

-- 
Rich Ulrich, [EMAIL PROTECTED]
http://www.pitt.edu/~wpilib/index.html
"Taxes are the price we pay for civilization." 
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to