Difference in behaviour of model.matrix

[Douglas Bates]

This is an obscure R/S incompatibility but it is tripping up some code
for us in the lme library.  If you specify the contrasts argument in a
call to model.matrix, that seems to take precedence over the
interpretation of the formula.  In S if the formula contains a "- 1",
that will cause the contrasts to be suppressed.

 S> foo <- data.frame(bar = factor(rep(1:3, rep(2,3))))
 S> foo
   bar 
 1   1
 2   1
 3   2
 4   2
 5   3
 6   3
 S> model.matrix(~bar - 1, model.frame(~ bar - 1, data = foo))
   bar1 bar2 bar3 
 1    1    0    0
 2    1    0    0
 3    0    1    0
 4    0    1    0
 5    0    0    1
 6    0    0    1
 S> model.matrix(~bar - 1, model.frame(~ bar - 1, data = foo), 
 +    contr = list(bar = contr.treatment(3)))
   bar1 bar2 bar3 
 1    1    0    0
 2    1    0    0
 3    0    1    0
 4    0    1    0
 5    0    0    1
 6    0    0    1

whereas 

 R> foo <- data.frame(bar = factor(rep(1:3, rep(2,3))))
 R> model.matrix(~bar - 1, model.frame(~ bar - 1, data = foo))
      bar1 bar2 bar3
 [1,]    1    0    0
 [2,]    1    0    0
 [3,]    0    1    0
 [4,]    0    1    0
 [5,]    0    0    1
 [6,]    0    0    1
 attr(,"assign")
 [1] 1 1 1
 R> model.matrix(~bar - 1, model.frame(~ bar - 1, data = foo), 
 +    contr = list(bar=contr.treatment(3)))
      bar2 bar3
 [1,]    0    0
 [2,]    0    0
 [3,]    1    0
 [4,]    1    0
 [5,]    0    1
 [6,]    0    1
 attr(,"assign")
 [1] 1 1

I would be willing to admit that our code should perhaps be redesigned
to avoid this unfortunate construction.
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._