At 01:02 AM 1/21/03 +0000, Don wrote:
>
> If the treatment is successful, will it be utilised in future? If so,
> then I don't think you can honestly say that the children in the study are
> the only ones of interest (i.e., the population). They are a sample from
> a population that includes future children, and so the population variance
> is not really known. I'd use t-tests rather than z-tests.
>
> You may also want to set your per-contrast alpha lower than the usual .05
> in order to keep the family-wise alpha at a reasonable level. For a nice
> demonstration of this issue, see Jerry Dallal's Little Handbook of
> Statistical practice (http://www.tufts.edu/~gdallal/LHSP.HTM), and click
> on the link for A Valuable Lesson
> (http://www.tufts.edu/~gdallal/multtest.htm).
I'm currently just concerned with whether the pre-treatment scores of those
in control group, and those in the treatment group are ths same. So for this
exercise, isn't it true to say that I have data for the entire population of
interest, and therefore sigma is known?
Many Thanks,
Don (alias Eircom)
if you are claiming that you have a known sigma (or sigmas) ... because you
have the populations ... then, don't you have known means too? if you have
known means ... then you know the difference ... you don't have to do
inferential tests
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
. http://jse.stat.ncsu.edu/ .
=================================================================
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
. http://jse.stat.ncsu.edu/ .
=================================================================
