Re: Retrieving probit or logit standard errors

[Andreas Christmann <A.Christmann@hrz.uni-dortmund.de>]

Perhaps the following code helps, but it is not elegant.
Assume that y is the response vector (0/1) and
x is the matrix of explanatory variables.

out <- glm(cbind(y,1-y) ~ x , family=binomial)
out2 <- summary(out)$coefficients

Then
out2[,1] is the vector of the ML estimates
and
out2[,2] is the vector of the standard errors.

For each data set i, these vectors can be stored as rows in two matrices,
say betahat and SEhat.

Andreas

---------------------------------------
PD Dr. Andreas Christmann
Universitaet Dortmund
Hochschulrechenzentrum
Wissenschaftliche Anwendungen
D-44221 Dortmund
GERMANY
email  A.Christmann@hrz.uni-dortmund.de
phone  (049)-231-755-2763
fax    (049)-231-755-2731
---------------------------------------