[R] Issue with predict() for glm models

jrausch@nd.edu jrausch at nd.edu
Thu Sep 23 16:45:14 CEST 2004

Thanks to John Fox, Andy Liaw, and Uwe Ligges for their help with my problem
regarding the use of the predict()  function to obtain predictions for a new
set of predictor values. It appears that the bottom line (at least for my
purposes) is that the names and the setup for the data of the predictors in the
glm and the new data need to be consistent. The safest way that I know to do
this from reading  John, Andy, and Uwe's responses is to label each predictor
separately and place them into the glm model separately. Then, when creating a
new data frame to utilize in the predict() function, ensure to consistently
name the predictors. For illustrative examples, see the reply emails of John,
Andy, and Uwe.  

Thanks again, 


Joe Rausch, M.A. 
Psychology Liaison 
Lab for Social Research 
917 Flanner Hall 
University of Notre Dame 
Notre Dame, IN 46556
(574) 631-3910

"If we knew what it was we were doing, it would not be called research, would
- Albert Einstein

More information about the R-help mailing list