Date:         Thu, 3 Sep 1998 12:47:32 -0500
Reply-To:     "Nichols, David" <nichols@SPSS.COM>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@UGA.CC.UGA.EDU>
From:         "Nichols, David" <nichols@SPSS.COM>
Subject:      Re: one last nested ANOVA question

Jenny and I have been over her design to a certain extent off line, but I
wanted to clarify for others interested that in the general case, these
kinds of issues are quite data specific, not a general function of
covariates with nested designs.
 
David Nichols
Principal Support Statistician and
Manager of Statistical Support
SPSS Inc.
 
        -----Original Message-----
        From:   JENNY A NOVOTNY [SMTP:jnovotny@du.edu]
        Sent:   Monday, August 31, 1998 12:52 PM
        To:     SPSSX-L@UGA.CC.UGA.EDU
        Subject:        one last nested ANOVA question
 
        One last question for the list regarding nested ANOVAs in SPSS 7.5
 
        Could someone explain how the covariate is handled within a nested
design?
        I noticed from running my analyses both with and without the
covariate
        (and asking only for covariate main effects rather than factor by
        covariate interactions) that
 
        1) no test of the covariate could be done
        2) no parameter for the covariate is estimated
        3) the covariate (as specified for main effect only) seems to affect
only
        the main effect for the factor involving the nesting (and NOT any of
the
        interactions with that factor).
 
        Am I to interpret from these results that
        1) I need to include all factor*covariate interactions
        or
        2) the effect of the covariate is only interpretable for some
factors
 
        I have set up my analysis as a split plot design with 3 factors
(with an
        additional factor nested within 1 factor) and 1 covariate.
 
        Thanks!
 
        Jenny Novotny
        University of Denver (student)