Huggins closed capture estimator within Robust Design

[luciake]

I am trying to estimate N for prairie vole populations kept within outdoor enclosures over 15 week intervals. Initial populations sizes are either 8 individuals or 24 individuals, increasing to at most ~100 individuals. While the sampling methods seem to match the robust design estimator well, I was getting very strange error values and have since switched to the Huggins estimator within the robust design based on previous posts to the forum. Do you still compare Markovian movement vs random movement vs no movement for the gamma values when selecting the best model fit, or is that only for pure robust design models?

[Todd]

You can still compare differences in gamma values using the PIM charts and/or the design matrix and/or fixing parameters (such as gamma'="=0) when using the Huggins version of the model with the derived population estimates.

I believe it has been mentioned that you may be more likely to reach convergence using the Huggins models if you have low sample sizes because there are less parameters being estimated in the likelihood estimation.