Dot notation and GOF

[PCJones]

I have a data set I am analyzing via CJS models that also has some missing data that I am accounting for by using dot notation in the encounter histories. Is there an appropriate GOF test for this type of model? MARK specifically states that the bootstrap GOF and median c-hat tests do not handle the dot notation. Any suggestions are appreciated.

[abreton]

I suspect U-CARE (Utilities CAture REcapture ... for performing GOF tests on single and multi-state CMR data) would also crash if you attempted to import encounter histories that included the dot notation, nonetheless you might want to scan their documentation for single-state GOF tests,

http://www.cefe.cnrs.fr/biostatistiques ... /logiciels

Alternatively, if you have very few animals with missing data, perhaps dropping those animals to assess fit would be reasonable? Depends on the proportion of the sample missing data. As a reviewer, if it was <5% then I wouldn't be concerned. Not ideal, but an option. And if U-CARE single-state GOF tests on the partial dataset reveal lack of fit (transience, trap-dependence, etc) then it also seems reasonable to assume the small fraction of excluded data have the same issue(s). In MARK you would use the complete dataset ... apply any adjustments (for transience, etc) and then proceed with your analysis. Just some thoughts ...

andre