Factor B has 21 levels?! Is this a typo? What does Factor B represent?
Anyway, I don't use the GLM procedure all that often anymore for such
designs, but I'm guessing that you could test for simple effects
employing the LMATRIX subcommand.
Ryan
On Wed, Apr 13, 2011 at 10:35 AM, Talk Stats <
[hidden email]> wrote:
> Hello,
>
> I'm running a 2 way repeated measures ANOVA (factor A has 2 levels and
> factor B has 21 levels). I have a significant interaction effect (A*B). To
> follow up this interaction, I use the following syntax,
>
> /EMMEANS=TABLES(A*B) COMPARE(A) ADJ(BONFERRONI)
> /EMMEANS=TABLES(B*A) COMPARE(B) ADJ(BONFERRONI)
>
> I get a 'Pairwise Comparisons' table and I get a 'Multivariate Tests'
> table in the output. Why are there no 'Univariate Tests' table ? I'm
> confused about how to proceed, should I,
>
> a) Ignore the Multivariate Tests table and look at Pairwise Comparisons
> table to see where the differences lie ? I say this because the
> Multivariates Tests are a different model from the univariate RM ANOVA
> model.
>
> b) Look at the Multivariate tests table first and only follow up those
> simple effects that are significant ?
>
> Why isn't there a table that just tests the simple effects of a factor at
> each level of the other factor ? I don't want the 'multivariate' simple
> effects.
>
> Cheers,
> TS
>
> This question was posted on Talk Stats Forum.
>
http://www.talkstats.com>
> =====================
> To manage your subscription to SPSSX-L, send a message to
>
[hidden email] (not to SPSSX-L), with no body text except the
> command. To leave the list, send the command
> SIGNOFF SPSSX-L
> For a list of commands to manage subscriptions, send the command
> INFO REFCARD
>
=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
Locked
