Individual covariates with heterogeneity parameter (pi)

[jctroxler]

Hello,
I'm using a closed capture model (Huggins full heterogeneity) to estimate Pi, P, and C, and derive N for a population of black bears. I have 8 week-long sampling sessions, 2 groups (M & F), 2 mixtures, and 1 individual covariate; number of captures during the previous summer.

I tested several candidate models, getting the best results with Pi., time effect in P and C, and roughly equal weight between P+C and P=C models. Adding the covariate to P and C improved these models substantially. Finally, I added the covariate to Pi to get 2 models with 99% of AICc weight; 59 and 40 percent.

Is it appropriate to use an individual covariate to estimate Pi? I was told it may not be since Pi pertains to the population and not individuals.
Thanks

[cooch]

...Is it appropriate to use an individual covariate to estimate Pi? I was told it may not be since Pi pertains to the population and not individuals.
Thanks

That would be correct.

[cooch]

...Is it appropriate to use an individual covariate to estimate Pi? I was told it may not be since Pi pertains to the population and not individuals.
Thanks

That would be correct.

Forgot to mention - the point is moot anyway, since pi is assumed to be constant over all time intervals. If you look at the PIM structure, you'll see this. So, given that it is a constant, then there is no 'temporal covariate' you can apply to it.