|
|
Hi all,
I'm trying the GENLIN procedure in SPSS 15 for the first time and have a
couple questions.
1. I am modeling the number of children born as a Poisson regression
with a log link, using (the log of the) age of the woman as an offset
variable. If EMMs are requested with the 'compute means for response'
option selected then are the means the number of children /per year of age/?
2. In another Poisson regression I am looking at ecological data from a
field trial. Data was collected from three sites and in a regular ANOVA
framework site would be considered a random factor. There are no
predictors associated with the site level of the design (ie, not a
multilevel design). I don't see a way to specify fixed/random effects in
GENLIN. Is there a way to account for this variation between sites?
Thanks for any pointers.
Cheers,
Kylie.
|
Locked
