0

Since I updated R to version 3.2.3 and the package 'nlme' to version 3.1-124 I am unable to use the augPredict (and predict) functions for models including mixed effects. Below you can find an example of my code using the Soybean data that are provided with the nlme package. I get the error: Error in eval(expr, envir, enclos) : object 'Variety' not found Am I doing something wrong (although this code worked fine on my data in previous versions), or is this a bug in the software?

#base model
M1.lis <- nlsList( SSlogis, data=Soybean)
M1.lis

M1.nlme <- nlme( M1.lis )
M1.nlme
summary(M1.nlme)

M1.Fix<-fixef(M1.nlme)

#add fixed effects
M2.nlme <- update( M1.nlme, fixed = Asym + xmid + scal ~ Variety,
                         start = c(M1.Fix[1], 0, M1.Fix[2], 0, M1.Fix[3], 0))

summary(M2.nlme)
plot( augPred(M2.nlme, level = 0:1))
Bhargav Rao
  • 50,140
  • 28
  • 121
  • 140
Tanja Milotić
  • 21
  • 3

1 Answers1

0

It apparently was a bug which has been fixed in 'nlme' version 3.1-126

Tanja Milotić
  • 21
  • 3