Date:         Tue, 23 Sep 1997 13:30:22 -0400
Reply-To:     Sally Vegso <sally.vegso@YALE.EDU>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@UGA.CC.UGA.EDU>
From:         Sally Vegso <sally.vegso@YALE.EDU>
Subject:      Nonorthogonal contrasts in repeated measures ANOVAs
In-Reply-To:  <2.2.32.19970923160138.009c9320@sph.emory.edu>
Content-Type: text/plain; charset="us-ascii"

I have 4 repeated timepoints of a variable.  The overall RMANOVA is
significant (< .05).  I now would like to compare each mean to the baseline
mean to see wherein the differences lie.  I have David Nichols excellent
white paper on "Nonorthogonal Contrasts On WSFACTORS in MANOVA".  My
concern is that when I perform these contrasts on the data, I get the exact
same answers (p values) as when I simply do multiple paired t-tests.  I am
confused about this, because I know it is incorrect to do multiple t-tests
because of the increase chance of type I error.  I guess I expected to get
more conservative answers with the contrasts.  Should I still use
Bonferroni corrections when interpreting these contrasts?  Are there better
ways to address this problem?  Any ideas would be helpful-
 
Sally Vegso
 
 
 
----------------------------------------------------
Sally J. Vegso, M.S.
Yale University Dept. of Psychiatry
sally.vegso@yale.edu
http://info.med.yale.edu/psych/cnru