🚀 go-pugleaf

Dear SPSS users,

I have a question/problem that may be of interest to some of you. I'm
using non-liner regression  (NLR) to estimate three parametres of
a growth curve. The function is something like this:

 y= A exp(-B expp(-C x)),

where A, B and C are the parametres.

Let's suppose that I run the analysis for two different populations

(1 and 2). So I obtain A1, B1 and C1 for pop. 1 and A2, B2 and C2
for pop. 2.

The question is whether the growth patterns of the two populations
differ significantly or not.
Some fisheries papers recommend a Hoteling's T2 test based on the
parametres estimates and their variance-covariance matrices. As part
of the output generated by NLR, one gets the Asymptotic S.E. and
Asymptotic correlation matrix of the estimates. So knowing N for
both populations is very simple to obtain the variance-covariance
matrices. The problem is to get this matrices as input of SPSS's
MANOVA. The subcommand READ only is able to read matrix materias
produced by the subcommand WRITE. BTW I'm using SPSS/PC+ 4.0.

Is it any way to solve this. Any commnets, suggestions, hints and
even reassurances will be most appreciated. I would also like to know
if I can make discriminat analyses with matrix materials.

Thanks very much for your attention.

Juan A. Balbuena
Department of Animal Biology, University of Valencia, Spain.
phone: +34 6 386 43 75   fax: +34 6 386 43 72
e-mail: balbuena@vm.ci.uv.es  or  BALBUENA@EVALUN11.BITNET