Royle Biometrics abundance estimate

[dshealer]

Is it possible that severe overdispersion/zero inflation could result in the Royle Biometrics model returning an abundance estimate that is actually somewhat [i]lower[i] than the naive estimate? The SEs look suspiciously small also. Similar data sets from other years seem to be modeled appropriately, so this one is a head-scratcher.

Thanks,

Dave

[ferny]

Does your data look like it might be zero-inflated (and do you have c-hat>1 indicating overdispersion), or are you just guessing at what might be causing the result you're getting?

Have you tried running a ZIP model rather than Poisson - might that fix it/what did it return, following Wenger and Freeman 2008 or Joseph et al 2009? I'm not aware that PRESENCE can use abundance distributions other than Poisson though so I think that would need to be done in R/WinBUGS unless anyone could advise better - I'm afraid I'm a relative newcomer to the field.

Hope you've figured it out!
Paul