TSM and additive models

[Pau Lucio]

Dear all,

I am working with a TSM model to deal with transients. I have followed the example in Mark Book 7.4. ‘Time since marking’ – when an age model is NOT an ‘age’model. The data are adult males as they were the only ones showing the presence of transients. Five occasions.

These are the top models:

Model QAICc ΔQAICc QAICc (w) No.Par QDeviance
Φ(.) p(.) 206.37 0.00 0.47 2 16.92
Φ(./.) p(.) 207.71 1.34 0.24 3 16.21
Φ(./time) p(.) 209.48 3.10 0.10 5 13.82
Φ(time/.) p(.) 2011.74 5.36 0.03 6 13.98


I have been asked by a reviewer to use additive models to force transient and resident to have the same difference in survival every year.

What I understand is that I should try to constrain an age model to achieve this. Am I correct? So, I have followed the example of the book 7.2. Constraining an age model: marked as young only on pages 7-14.

The PIM of the Φ(time/time) p(./.) will look as this:

https://drive.google.com/file/d/1wjwZAG3drDtZGoFraVSQ8UhSeJYO06B_/view?usp=drive_link

When I work with the matrix to get a common intercept for both age classes (no interaction terms), the matrix will look like this one:

https://drive.google.com/file/d/12jxkxP62avSZ2TCGIqmwZwYgFFyfkIT_/view?usp=drive_link


The age column 1-4 for the first event (transients and residents) and 5-7 for the residents

I would appreciate any suggestions

Many thanks

[cooch]

If it's in the book, it must be correct.

More seriously, your approach (and DM) are correct.

[Pau Lucio]

Great! Many thanks for your reply