Simulations (closed captures): Problem with specification!

[cebert]

Hello!
Now this might be a typical 'beginner-question'... nevertheless I would be very grateful for some advice!
I am trying to do some simulations of closed capture models. Now I have a question the answer for which I did not find in the handbook nor in the helpfiles...
When I do the 'specification of simulations', I do not know how to interpret the 'releases' window. Because there it says 'population size for each group', I first thought this should represent the size of the simulated population. But in the 'beta values'-window, my last beta value also represents N, the population size. At least it does in my PIM-chart... so what I don't understand is which one of those two possibilities really stands for the simulated N - and what the other one then stands for...
Can anybody help me so I can do my simulations right?
Cheers,
Cornelia

[cooch]

Hello!
Now this might be a typical 'beginner-question'... nevertheless I would be very grateful for some advice!
I am trying to do some simulations of closed capture models. Now I have a question the answer for which I did not find in the handbook nor in the helpfiles...
When I do the 'specification of simulations', I do not know how to interpret the 'releases' window. Because there it says 'population size for each group', I first thought this should represent the size of the simulated population. But in the 'beta values'-window, my last beta value also represents N, the population size. At least it does in my PIM-chart... so what I don't understand is which one of those two possibilities really stands for the simulated N - and what the other one then stands for...
Can anybody help me so I can do my simulations right?
Cheers,
Cornelia

The default link for the generating model is the identity link. So, for the final beta, you set it equal to 1.0 (and all the preceding betas to whatever parameter value you want to simulate for p, c), and then set the true population size in the releases window.

[gwhite]

Cornelia:
You can ignore the last beta -- the number specified for Releases or Population Size is what is used.

If you want to be sure you are specifying everything correctly, run a simple model (like M0 or Mt) where we know the estimator works, and make sure the mean of the simulations matches the parameters you specified.

Gary

[cooch]

Cornelia:
You can ignore the last beta -- the number specified for Releases or Population Size is what is used.

If you want to be sure you are specifying everything correctly, run a simple model (like M0 or Mt) where we know the estimator works, and make sure the mean of the simulations matches the parameters you specified.

Gary

Gary -

Just for completeness - won't it depend (somewhat) on whether or not N is included in the likelihood? For the Huggins model(s), the final beta corresponds to the final encounter probability, so you can't simply ignore it...unless I'm missing something.

[gwhite]

Correct -- the last beta only corresponds to N for the closed models with the population size included in the likelihood. For Huggins models, the last beta is the last c.

I just assumed in the previous response that we were talking about the closed models with N in the likelihood.

Gary

[cebert]

Thank you very much for all your replies!!! This helps me a lot...
Cheers,
Cornelia