This should be fixed in nlme 3.1-112.
However, nlme has little support for formulae such as resp ~ 0, and does
things like p:1 where p is the number of columns in the model matrix.
3.1-112 does better but evidently the design did not consider this
possibility.
On 30/09/2013 13:42, ONKELINX,
Dear all,
predict.lme() throws an error when the fixed part consists of only an intercept
and using newdata. See the reproducible example below. I've tracked the error
down to asOneFormula() which returns in this case NULL instead of a formula.
Changing NULL instead of ~1 in that function (see
