Dear all,
I am testing moderation effects in a logistic regression model. Just wonder
if it is necessary to compare the two nested models (or to see if the model
with interaction terms is significantly different from the model which has
no interaction terms). Is it only if the null hypothesis is not rejected,
meaning the two models are not significantly different, can I accept that
the interaction term are meaningful and should keep it in the final model.
If it's a necessary procedure to do, how can I do it in SPSS?
Thanks
--
Yawen
=====================
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
