Help with categorical covariate!!!!

[Puig]

Hi!

I'm working on some closed capture Huggins models for gray whale abundance estimate. I’m using some covariates, such as Beaufort scale, visibility, etc in this study. I have some trouble with covariate Beaufort scale (it’s a categorical covariate with 4 states) but I’m not sure of how I have to use this covariate in the design matrix in Mark.

I hope you can help me.

Thanks
Héctor

[cooch]

Hi!

I'm working on some closed capture Huggins models for gray whale abundance estimate. I’m using some covariates, such as Beaufort scale, visibility, etc in this study. I have some trouble with covariate Beaufort scale (it’s a categorical covariate with 4 states) but I’m not sure of how I have to use this covariate in the design matrix in Mark.

If the category is ordinal, and you don't mind scale issues, you can simply treat it like a linear covariate. If it isn't ordinal, then they're simply factors of a single classification variable, and get treated the same as any other classification variables (e.g., time-coding treats time as a classification variable).

[Puig]

I treated the Beaufort covariate as a classification variable (states 0, 1, 2, 3), where the reference variable was Beaufort 3 instead of Beaufort 0. Is this correct? Because in my results I had less parameters estimated than I was suppose to (jus’t one!!).

Now I don’t know If the problem is the way I treated the covariate or that I have to treat my covariate as an ordinal value. I believe that the scales issue is not a problem. Can you please give me some advice how to do it in design matrix in mark.

Thanks

Hector.

[cooch]

I treated the Beaufort covariate as a classification variable (states 0, 1, 2, 3), where the reference variable was Beaufort 3 instead of Beaufort 0. Is this correct? Because in my results I had less parameters estimated than I was suppose to (jus’t one!!).

Now I don’t know If the problem is the way I treated the covariate or that I have to treat my covariate as an ordinal value. I believe that the scales issue is not a problem. Can you please give me some advice how to do it in design matrix in mark.

Thanks

Hector.

Read Chapter 7. Continuous covariates (which is what you're assuming if you treat your covariate as ordinal) are covered in detail there.