R devel - Feb 2003 - Pearson's residuals in logistic regression (PR#2539)

Full_Name: Bin Nan
Version: 1.3.0
OS: Win32
Submission from: (NULL) (141.211.15.110)


The function resid(fit, "pearson") seems not giving the same
Pearson's residuals
for logistic regression as what Splus does. I found the problem when I fitted
glm with family = binomial using Crowder's seed data.

On Mon, 10 Feb 2003 bnan@umich.edu wrote:
> Full_Name: Bin Nan
> Version: 1.3.0
           ^^^^^

That's way, way old.  This has been changed since.
  
> OS: Win32
> Submission from: (NULL) (141.211.15.110)
> 
> 
> The function resid(fit, "pearson") seems not giving the same
Pearson's residuals
> for logistic regression as what Splus does. I found the problem when I
fitted
> glm with family = binomial using Crowder's seed data.
-- 
Brian D. Ripley,                  ripley@stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

bnan@umich.edu writes:
> Full_Name: Bin Nan
> Version: 1.3.0
> OS: Win32
> Submission from: (NULL) (141.211.15.110)
> 
> 
> The function resid(fit, "pearson") seems not giving the same
Pearson's residuals
> for logistic regression as what Splus does. I found the problem when I
fitted
> glm with family = binomial using Crowder's seed data.
1.3.0 is almost two years old. Some problems of that kind have long
since been resolved. Do try a current version.

-- 
   O__  ---- Peter Dalgaard             Blegdamsvej 3  
  c/ /'_ --- Dept. of Biostatistics     2200 Cph. N   
 (*) \(*) -- University of Copenhagen   Denmark      Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard@biostat.ku.dk)             FAX: (+45) 35327907