Mathew, if you use GLM procedure to run ANOVA (or MANOVA, ANCOVA, ETC) and include the emmeans subcommand, it should give you all the comparisons:
/EMMEANS = TABLES(treatment*age) COMPARE (treatment)
/EMMEANS = TABLES(treatment*age) COMPARE (age)
Is this what you were looking for?
----- Original Message -----
From: "Matthew Fuller-Tyszkiewicz" <
[hidden email]>
To:
[hidden email]
Sent: Tuesday, December 15, 2009 7:44:40 PM GMT -08:00 US/Canada Pacific
Subject: Post-hoc comparisons in 2-way ANOVA
Hello all.
�
I’m running a 3 (Treatment) x 3 (Age groups) ANOVA with post-hoc comparisons. I’ve noticed that SPSS provides post-hoc comparisons of levels within each factor separately (e.g., comparing 3 different treatment levels with each other AND separately comparing the 3 different age groups) rather than comparing all 9 groups. I know that I could calculate the comparisons by hand for, say, group 1 (treatment 1, age group 1) v group 3 (treatment 3, age group 1), but I’m wondering if there is a reason why SPSS does not do this calculation automatically? Is there a concept I’m missing?
�
Related to this, is there a reason why comparisons should not be done when covariates are used in ANCOVA or MANCOVA?
�
Thanks in advance,
Matt.
�
=====================
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
