Question about logistics regression.

[dinsya]

I have question about logistics regression.
I have 3 year survey data 2005,2006,2007. Risk behavior, let say r1 r2 r3 r4 …etc and outcome (which is categorical data: outcome1 outcome2 outcome3).
I coded year as 1,2,3.
I want to have a polytomous logistic regressiob model and use year as a weight

I use the formula as follows:
proc logistic data=one;
class outcome (ref=' outcome1)/ param=ref ref=first order=internal;
model outcome= r1 r2 r3 r4.../ nodummyprint;
weight year;
run;

was this right?
I am confuse. Can somebody help me, please??

Thank you
dinsya

[Link]

try tying in "link=glogit" after "nodummyprint"

If that still doesn't work, look into this document:
http://www.stat.psu.edu/~jls/stat544/lectures/lec19.pdf

[dinsya]

Thanks for the help...