roystonteo -

For classical statistical inference regarding the population mean, the t
distribution is appropriate when the sample is used to estimate the unknown
population standard deviation (and you are willing to assume that the
population values are normally distributed).

For that situation, the t distribution can be used for any sample size. For
large sample sizes, the t distribution can be approximated by the normal
distribution. The usual guideline has been "30 or more."

- Mike Middleton
Email
Web http://www.MikeMiddleton.com



"roystonteo via OfficeKB.com" <u48590@uwe wrote in message
news:9135401846fe4@uwe...
Dear Experts,

I will need your kind assistance on this. I am very new towards Excel
Functions, hope you have the patience for me. Thank you in advance.

The question is:

A Researcher wants to carry out a test involving the mean for a sample of
size n = 18. She does not know the true value of population standard
deviation, but is reasonably sure that the underlying population is
approximately normally distributed. Should she use a z-test or at-test in
carrying out the analysis?

--
Message posted via http://www.officekb.com