Impossible states in multi-state model

[Aline]

Is there any way to specify that some states are not possible at some ages? For example, if young individuals always start in a given state (say, nonbreeder) and then later can transition into other states (breeder).

If I don't do anything, parameters for these state combinations are estimated, even if there is no data to support them. So far, I have been fixing the parameters of the impossible states to zero, being careful to specify my models so that they would be estimated separately from other parameters. But this feels a bit wrong, since those parameters don't really "exist." Is there a better way of doing this? I haven't been able to find anything in the literature.

I'm actually running the model in RMark, if that makes a difference.

Thank you.

[cooch]

If I don't do anything, parameters for these state combinations are estimated, even if there is no data to support them.

Those aren't 'estimates', but rather, the starting values MARK uses (typically, 0.45). They should be reported with SE that are not remotely correct.

So far, I have been fixing the parameters of the impossible states to zero, being careful to specify my models so that they would be estimated separately from other parameters. But this feels a bit wrong, since those parameters don't really "exist." Is there a better way of doing this? I haven't been able to find anything in the literature.

Nothing wrong with it at all. Fight the feeling.

Also, remember to set p=0 for those states that are unobservable at some age(s). Both psi and p need to be fixed as logical constraints.