comparison of N(t) given by a Robust Design model

[Globice974]

Hi everyone,

I work on spinner dolphins and I have applied a Robust Design model to my data, with year as primary occasion. The best model gives me an estimated abundance of my population which varies according to year. I would like to compare the resulted abundances in order to highlight a trend for the population, but I don't know how to proceed, given that the annual abundance given by the model does not correspond to a mean from a defined sample. I have performed a chi square test to compare the evolution of the abundance to a theoritical constant distribution across years, but the abundance of each year are not independant.

Thank you for your ideas/responses

[Bill Kendall]

It sounds like you would want to run a variance components analysis, with trend, on the derived estimates of N. This is found by clicking on Output -> specific model output -> variance components -> derived estimates.

It's described in Appendix D of the Cooch and White online book.