glm.nb, anova.negbin

On Wed, 26 Sep 2001, Jari Oksanen wrote:

juli at ceam.es said:

Error in get(x, envir, mode, inherits) : invalid first argument
In addition: Warning message:
tests made without re-estimating theta in: anova.negbin(t2) >

I looked quickly, and it seems that this error comes actually from
`anova.glm'  and from the following `get':

        if (!is.function(method))
            method <- get(method, mode = "function")

`anova.negbin' sends the  `negbin' `object' to `anova.glm', but the
negbin object does not have a method (`glm' objects have method
"glm.fit"). There seem to be some other differences in `negbin' and
`glm' objects, so that `anova.glm' can't handle them (e.g., glm.control
object is missing in `negbin' object).

I would like to help, but I don't have time now (got to mark exams).

Yes, that's it: R is incompatible with S.  Although the information
is in the saved call, R uses a duplicate copy.  In glm.nb() adding
    fit$method <- method
    fit$control <- control
at the end seems to fix it, but without an example I can reproduce,
I cannot of course be sure.

And that was my main point: if reporting problems *please* report the
results of traceback(), and if possible a small reproducible example.