[R-sig-ME] Coefplot2 availability? or How to interpret?

[Ben Bolker]

Sheryn Olson <sheryn.olson at ...> writes:

> 
> Hello all,
> 
> I need to correctly(!) interpret and graphically present a negative
> binomial mixed model generated with glmmADMB.   Is the coefplot2 package
> working as of Jan, 2013?   Is there an alternative if not?   Is there an
> "extraction" package to provide confidence intervals around the
> exponentiated coefficients?   I ran an mcmc approximation to get confidence
> intervals, but I'm not sure how to present them graphically.  Is there
> ggplot code that might work?
> 
> The study system: What are the effects of season and forest stand types on
> snowshoe hare pellet densities.  Do snowshoe hares move more seasonally in
> some forest stands than in others?  Year is a random (nuisance) variable.
>  Each stand type (treatment) has unequal sample sizes that vary by year.
> The count data are constrained at zero, residuals are overdispersed.

The error message is giving you a hint that you should try installing
from the source package:

install.packages("coefplot2",repos="http://www.math.mcmaster.ca/bolker/R",
    type="source")

  That should work even if you don't have the tools installed for compiling
packages from source, because the coefplot2 package doesn't require any
binary compilation (all R has to do is unpack an archive file and put the
bits in the right places, which it shouldn't need any external tools to do).

   Your MCMC results might be more conservative/robust than the default
confidence intervals.

  I haven't otherwise looked at your model/results etc. posted below to see
if they're sensible ...