taking your note literally: including multiple intercepts is impossible
since they are perfectly correlated. but perhaps you mean having one
intercept for each group or state or something. you could have one SET of
intercepts for states and another SET for families. whatever it is you
wanted, the same rules apply as in simple linear regression analysis as in
logit.
Gary
On Sat, 10 Jun 2006, Jacqueline Chattopadhyay wrote:
Hello,
I am running logistic regression for dichotomous dependent variables with a
transition model. I need to obtain output that has three intercept terms.
Does anyone know of a command that tells R to calculate more than one
intercept? I am trying to do all work within Zelig, so if it is a command
that operates within Zelig, that would be optimal. I have tried
unsuccessfully to find information on how to obtain multiple intercepts and
would appreciate any advice.
Thanks,
Jacqueline