I think the PIRLS program does what you want. See http://www.jstatsoft.org/v02/i05/pirls/ , for "PIRLS: Poisson Iteratively Reweighted Least Squares Computer Program for Additive, Multiplicative, Power, and Non-linear Models"
[EMAIL PROTECTED] (Iyue Sung) wrote in message news:<[EMAIL PROTECTED]>... > Hello All, > > I've fit a Poisson regression and would like to estimate marginal > incidence rates for each of say, m, treatments, 'adjusting' for a few > factors. > > I'm pretty sure this can be accomplished using an estimate analogous > to a least squares means for 'regular' regression, though I haven't > worked it out yet. Presumably, this will be accomplished via > asymptotic normality of the MLE. . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
