I'm using a call to eval to evaluate a linear model, however, I have found that despite calling
options (contrasts=c("contr.sum", "contr.poly"))
prior to evaluation, my model factors are coded using the indicator coding associated with the "contr.treatment" contrast option
As an inelegant work around I am setting the contrast option explicitly in the model, but I'm hoping that somebody could be me some help with how to set options (contrasts=c("contr.sum", "contr.poly"))
in the correct environment so that I don't need my hack. Some insights to what in the enviroments are in the context of eval would also be most helpful.
here is the version of the lm parameters that does what I want.
model next.choice ~ Lag0 , data = the.matrix, contrasts =list( Lag0 =contr.sum )
I don't want to have to add the last parameter contrasts =list( Lag0 =contr.sum ) if possible.
Lag0 is a factor with 4 levels 0,1,2,3
the following is the function that I call to evaluate models
the.model <- function(the.matrix, model)
{
options(contrasts=c("contr.sum", "contr.poly")) # this does not cause factors to be code using contr.sum is subsequent
expr <- paste("lm(", model,")")
rc <- eval(parse(text=expr))
return (rc)
}
_ platform i386-pc-mingw32 arch i386 os mingw32 system i386, mingw32 status major 1 minor 8.1 year 2003 month 11 day 21 language R
Dominic
______________________________________________ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
