Date:         Thu, 23 May 2002 00:33:38 +0000
Reply-To:     sward@computonet.com
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Sidne Gail Ward <sward@COMPUTONET.COM>
Organization: ComputoNet
Subject:      Re: Multiple post hoc univariate anovas?

In <dctneukiu4rhn9itqopeftlabq2e5i20vd@4ax.com> Jay Weedon <jweedon@earthlink.net> writes:

>On Wed, 22 May 2002 07:33:21 +0000 (UTC), Sidne Gail Ward
><sward@computonet.com> wrote:

>>I have one two-level independent variable and about 10 continuous
>>dependent variables. I've used manova to get the significance of the
>>model. Now I'd like to see how each individual dependent variable is
>>related to the independent variable.
>>
>>I assume that if I perform a series of univariate anovas, I'll run into a
>>Type I error problem.  How can I control for this?

>PROC MULTTEST seems ideally suited to your problem.

Yes it is. Thank you! I'd never heard of it before (and I can't find it in
my old SAS 6 manuals).  I'm using SAS v8, though, so I managed to find some
documentation on it.

There's an astonishing variety of different tests available.  Now I need
to figure out which to use (Sidak, permutation, bootstrap, step-up,
stepdown, etc.)  Anyone have any pointers to a source to help me select
one (or more)?

Thanks again,
Sidne
--
Sidne Gail Ward
sward@computonet.com