In article <8gjali$hh1$[EMAIL PROTECTED]>, [EMAIL PROTECTED]
says...
> Hello to all.
>
> I'm just studying Masako Ishii-Kuntz's "Ordinal log-linear models" (SAGE
> 1994 no.97).
> What I wanted to do is to practice fitting by SPSS 8.0 various types of
> models mentioned in the book.
>
> There is no problem to fit both row-effect and column-effect models and also
> uniform association models by SPSS GENLOG procedure. I've got the problem
> with Row-and-Column effects model (RC model), I don't know how to fit it
> using SPSS.
>
[snip]
> I have a feeling that fitting such models (including a term, which is a
> multiplication of estimated parameters) is impossible via SPSS.
> Nevertheless, I hope I'm wrong.
>
You're talking about the log-multiplicative rc II model here. One way of
estimating it is iteratively estimate standard loglinear models, first
treating the scale as given and estimating the effect parameters, then
taking the effect parameter as given and estimating the scale. I can
provide you with a GLIM macro to do this. I also have Sas and Stata
macros that can estimate an RC II model in multinomial logistic
regression. See http://baserv.uci.kun.nl/~johnh/mcl/ for details. This
procedure *could* be done in SPSS with some clever scripting but that
would be a lot of work.
However, I'd recommend using LEM by Jeroen Vermunt for rc II models. It
uses an E-M algorithm, an extension of the IPF algorithm, which makes it
very fast. LEM can estimate a wide array of models including latent
class, log multiplicative, event history. A Win95 version is available at
http://cwis.kub.nl/~fsw_1/mto/mto_snw.htm#software
Hope this helps,
John Hendrickx
===========================================================================
This list is open to everyone. Occasionally, less thoughtful
people send inappropriate messages. Please DO NOT COMPLAIN TO
THE POSTMASTER about these messages because the postmaster has no
way of controlling them, and excessive complaints will result in
termination of the list.
For information about this list, including information about the
problem of inappropriate messages and information about how to
unsubscribe, please see the web page at
http://jse.stat.ncsu.edu/
===========================================================================
