Another mixed question. Suppose I have model like this
Mixed y with time x/fixed time x/
random intercept time | subject(id) covtype(un).
Suppose x is a time varying covariate that stipulated to have an invariant
relationhip with y over ids. Therefore, X has no level 2 variance and it is
not listed in the random statement. I think this is the correct mixed syntax
statement. And if it's not, I'd like to know because I realize I've posted
some mixed questions where the words, syntax and alleged equations didn't
match.
My question is this. How does the procedure distinguish between the above
sort of model and a model where x is simply a level two predictor because x
is constant over time within persons? Is it solely the presence of the
x*time interacton or is it something else?
Thanks, Gene Maguin
=====================
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
