Assess model fit for Royle (2004) in Presence

[Dany]

Hiya all,

I am using Presence to estimate abundance(using the Royle 2004 count-data methodology), using Poisson.

However, not sure how to move on when the model show lack of fit ?
Can anybody suggest any idea please? Have try to use the other type of distribution, but there is no indication of chi-square, so not sure really what is going on .

Thanks

[Eurycea]

I'm not familiar with those models, but in general, it could be that your model just doesn't fit. You may not have enough data. You could try a simpler model, or a different approach to analysis. If lack of fit is due to overdispersion and you've estimated c-hat, you can incorporate that to correct your AIC values to QAIC. But I don't know how to do that with your type of analysis or if you even can. But I do know that with Poisson models, overdispersion is a common problem (i.e. where you variance doesn't actually equal your mean).