Dear allstatsusers,
I have been working with a loglinear model which it only has one predictive
variable and one response. The model which I obtain is
y=exp(2.76763 + 0.00021x). Because we are in a Poisson model we have that
for each x the lambda of the Poisson is lambda=exp(2.76763 + 0.00021x).
Now I am interested in obtaining a 95% C.I. for the lamda. Does anyone can
give the approppiate formula. Is it correct to calculate the
var(exp(2.76763 + 0.00021x)) and then do lambda +/- 2*var? But in this case
the C.I. will be simetric contrariously of a Poisson distribution.
Any help in Splus/SAS/etc... will be appreciate it.
Kind regards.Pep Puigvert
_________________________________________________________________
Send and receive Hotmail on your mobile device: http://mobile.msn.com
|