Barker Model question

[Laura Ellison]

I have a dataset with 8 occasions of live, dead, and resighting encounters (so 16 columns of 0, 1, 2s). Why are there 8 columns to the PIMs and what does that 8th interval refer to? I assume the 1st column of the PIM refers to the 1st interval from 1st to the 2nd occasion (year in this case), 2nd column/interval the 2nd to 3rd occasion, etc, but I am confused with the 8th. Am I missing something here? Thank you in advance for any help.

-Laura

[gwhite]

Laura:
There are 8 columns in the PIMs for survival, R, R', and r because the Barker model is like the dead recoveries model with the interval after the last occasion part of the model to estimate survvial. You should only have 7 columns for p.

If the above is not the case, then something is amiss.

Gary

[Laura Ellison]

Thanks, Gary. Yes, there are only 7 columns for p. So, in this case, real estimates of both parameter 7 and 8 in Survival PIM are confounded and inestimable, right?

Thanks.

-Laura

[gwhite]

Laura:
Survival rate 8 (S8) will not be estimable because you evidently have no dead recoveries during this interval, and with a time-specific model, only the product S8 r8 would be estimable.
If you have some live resightings in interval 8 (i.e., you have data to estimate R and R', you could actually unconfound S7 and p8 (but you would have to put some constraints on R8 and R'8 to do this).

I gather you have no data collected in interval 8, so therefore, the simple answer to your question is that S7 and p8 are confounded in the usaul way with the time-specific CJS model.

Gary