z Test vs. F Test - The most appropriate for comparing the Variances

mdurivage

Quite Involved in Discussions
Which of the above tests is the most appropriate for comparing the variances when:

n1 = 2000
n2 = 30-100 (varies during the time period)

One book reccomends when n>100 use the z Test. Does that mean one of the n's or both must be >100.

Thanks

Mark
 
D

Duke Okes

Which of the above tests is the most appropriate for comparing the variances when:
n1 = 2000
n2 = 30-100 (varies during the time period)
One book reccomends when n>100 use the z Test. Does that mean one of the n's or both must be >100.
Thanks
Mark

z test is for comparing means, F test for variances.

You may be choosing between z and t test. z is for n>30, t for n< 30, but both for comparing means, not variances.

And sample size for decision is n1+n2.
 

mdurivage

Quite Involved in Discussions
Duke,

There is a z test for variances as well. It is reccomended for samples <100. The question is do both samples need to be <100 or just one of them.
 
D

Duke Okes

Duke,There is a z test for variances as well. It is reccomended for samples <100. The question is do both samples need to be <100 or just one of them.

I've never seen/heard of it. Can you point me to an online reference?
 
Top Bottom