median chat v release.gof

[megpetrie]

Hi forum,

I have a data set with 5 recapture occasions, several groups (39), several fixed parameters, and a few lines of deleted design data. When I transfer the models I'm testing into MARK, MARK has a hard time with the median chat procedure (i.e., sometimes will not even initiate the process, and sometimes goes through the simulations but then fails after they are over, "every one of your simulations created a chat that is less than the observed chat). The window for the median chat procedure says my observed chat is >250. I've used the release.gof test to get a chat estimate and it is <1. When I test a predefined model in MARK (with no fixed parameters or deleted design data), I get a chat of approx 1.5. So my question is, should I use the RELEASE results and keep my chat at 1 or use the reasonable chat from the simple model? Is there a different model I should try to perform the median chat procedure on?

Thanks so much!

Megan

[jlaake]

Megan resolved this off list. Without fixed parameters she obtained reasonable results with the median chat procedure.