Covariables and group in CJS

[GabyTP]

Please if anybody could assist with a question! I am running CJS models for a mark-recapture dataset collected weekly (20 weeks) on weevils with two individual covariables (body size and number of copulations) and sex (coded as a group= male (1) female (0)). I am interested in evaluating survival by sex/body size/copulations. I run successfully all models, but now I would like to estimate survival by sex and body size (and sex and copulations). I found how to estimate survival by size using the linear function, same as for copulations; however my problem is that I do not know how to pull the data by sex. Please can anybody assist? I read the mark book and additional documents, while I understand the linear function I have to built I do not know how to estimate survival for each sex and each covariable.

Another question, the beta parameter for ‘group’ (sex) is -0.5301, is this for males or females? Thanks in advance!

[simone77]

Hi GabyTP,

I am not sure to understand what do you mean when you say “I do not know how to pull the data by sex”.
If you want to get survival estimates for, say, females according to each covariable, you must focus on the beta estimates.
Once you identify to which variable corresponds each beta estimate, you should build the equation for the specific case. For instance, to estimate survival of females according to body size you will need this:

S(fem-body): Beta0(Intercept) + Beta1(sex) + Beta 2(body)

By the way, this is assuming you have only additive effects in your model.
If you have coded your group variable (sex) as 0=females and 1=males, by default the reference level (the intercept) should correspond to the lower number (female) and therefore Beta for sex will correspond to the “male effect” with respect to females. In other words:

S(fem)= Beta0(Intercept)
S(mal)= Beta0(Intercept) + Beta1(sex)

In your case males seem to have lower survival than females.
There is good material explaining this sort of things in the Chapter 6, try to read it again because it is really informative.

Good luck!

[GabyTP]

Hi Simone
Thank you so much for your prompt response, it did help a lot! Kind regards, Gaby