I'm analysing survival of bats over 30 years. The individuals in the dataset belong to six different colonies. The recapture rate heavily varies between the colonies each year. Therefore I need to build the model with an interaction term colony*time for recapture p.

When I'm fitting MCMC models, the model runs if I fit it with colony+time for p. Unfortunately I get the following error message if I fit it as interaction term.

- Code: Select all
`model.parameters=list(Phi=list(formula=~1), p=list(formula=~time*colony))`

MCMCfit=crm(Mnat.NSH.ch.size, model="probitCJS",

model.parameters = model.parameters,

burnin = 10, iter=10)

probitCJS MCMC beginning...

p model = ~time * colony

phi model = ~1

Error in Xy %*% beta.y : non-conformable arguments

In addition: Warning messages:

1: glm.fit: algorithm did not converge

2: glm.fit: fitted probabilities numerically 0 or 1 occurred

Isn't it possible to run MCMC models with interactions?

Thanks a lot

Bianca