Royle Biometrics model: problem with convergence?

[jpellet]

Hello everyone,

I am trying to run replicated counts models in PRESENCE 2.4 using red deer counts in multiple populations (no site covariate, 3 sampling covariates).

The following message appears after running the simplest p(.) model:

AIC=4.00
Npar=2

suggesting that something went wrong. This is confirmed by the SE of the estimates = 0.000. Same problem when I modify the design matrix to run model with sampling covariates.

My dataset:

Number of sites = 43
Number of sampling occasions = 34
Number of missing observations = 1282

I first thought that I had too many missing values and reaaranged the original dataset into an equivalent collapsed matrix.

Number of sites = 43
Number of sampling occasions = 12
Number of missing observations = 336

Same problem.

I apologize if this problem is know and already solved in the forum, and I would greatly appreciate any help.

Jerome