Closed mgoggans closed 4 years ago
Error
is not a part of formula
documentation. It has been implemented as a special function in aov()
(try ?Error
which takes you to aov
), but will give an error, for instance, in lm
where it is not implemented. It is neither implemented in adonis
of vegan. We use formula
as a way of getting model.matrix
for the analysis, and Error()
will also fail in model.matrix
and model.frame
. In this case the error comes from the latter, as you see by issuing traceback()
after having the error.
@jarioksa Thank you for the thorough answer, very much appreciated!
Hello,
I have an experiment with a repeated measures design. I would like to use PERMANOVA to test the entire model - the vignette says adonis can handle a formula notation, but I am having problems.
Repeated measures ANOVA is denoted as such: dependent.variable ~ between.factor * within.factor + Error(id.of.repeat / within.factor)
When I try to run this, I get "Error: cannot find function "Error" "
This "error" notation in the formula is standard for base R linear models. I don't know why it won't work. Is adonis just not made to function with this type of linear model?