Hello Everyone,
I need some help interpreting a 2-way interaction in a negative binomial model that includes two
categorical predictors (VI - 3 categories, V2 - 3 categories) and two covariates. For example, my
interactions look something like this:
EXP(B) P
V1-2 * V2-2 .55 .003
V1-2 * V2-1 1.33 .003
V1-2 * V2-0 ------ (REDUNDANT PARAMETER)
Based on this, I would expect that the interaction between V1-2*V2-2 results in a lower rate of
the outcome occurring compared to the interaction between V1-2 * V2-0.
When I check the estimated marginal means, though, I see that the EMM for V1-2*V2-2 is
GREATER than the EMM for V1-2*V2-0. Why is this?
Also, is it possible to graph interactions for GLZM? I can't seem to find a syntax example for this,
just for GLM.
Thanks in advance for any help that one can offer!
Courtney
=====================
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
