I've tried to make a decision tree for the following data set:
....
I don't get a tree...
The rpart routine has decided that the best model it can find is the "intercept only" model, i.e., a tree with no branches at all. Stepwise regression can have the same outcome, by the way, if no variables pass the F-to-enter threshold. The plot and text routines won't plot anything for a null tree, because there is nothing interesting to plot. Terry Therneau