Skip to content
Prev 311923 / 398503 Next

Plotting an adjusted kaplan-meier curve

On Nov 25, 2012, at 5:17 PM, Brent Caldwell wrote:

            
Only give it newdata values that have the desired levels. Failing that  
advice being on point, then please provide str(survmat) and  
str(simple.cox.ethnicity$terms) .

(Could it be the case that your use of factor() around the terms means  
that getting hte names of your terms are not what you think they are?  
If the newdata argument is malformed then it gets discarded and the  
original values get used. Maybe you should do the "factorization"  
before you do the fit.)