contr.SAS(n, contrasts=TRUE)n rows and k columns, with k=n-1 if
contrasts is TRUE and k=n if contrasts is
FALSE.n levels. The returned value contains the
computed contrasts. If the argument contrasts is FALSE
then a square indicator matrix is returned. These contrasts are similar to those from contr.treatment
except that they represent shifts relative to the last level of the
factor, not the first level. This generates the same parameterization for a
linear model as is used in SAS (see
contr.treatment