[Rd] Bug in calculation of overdispersion for quasibinomial grouped

[Ulrike Grömping]

A quick addition:

SAS PROC LOGISTIC with option SCALE=PEARSON only calculates what R 
calculates in the grouped version; this is forced by refusing to 
calculate a FI estimate for ungrouped data, unless an AGGREGATE option 
is specified.

Apparently, SAS has decided for one model (not the one I am most 
familiar with), while the R version makes the model decision depend on 
the format of the data. As the family in glm is called quasibinomial, I 
would have expected it to conduct the adhoc adjustment.

Best, Ulrike

-- 
***********************************************
* Ulrike Groemping                            *
* BHT Berlin - University of Applied Sciences *
***********************************************
* +49 (30) 39404863 (Home Office)             *
* +49 (30) 4504 5127 (BHT)                    *
***********************************************
* http://prof.tfh-berlin.de/groemping         *
* groemping at bht-berlin.de                     *