modelling transients: c-hat estimates

questions concerning analysis/theory using program MARK

modelling transients: c-hat estimates

Postby Eleni Papadatou » Tue Apr 05, 2005 1:24 pm

Thanks Evan, that was useful.

However, I am still concerned with GOF for this global model and c-hat estimates:

- why not use bootstrap c-hat as well as the others?

- I ran median c-hat and the estimate was just below 1 (0.93): if I use this approach, then should I set c-hat=1 and "hold my nose"? :lol:
What if c-hat from UCARE/RELEASE is well above 1 (have not calculate yet, see question below)?

- do I run GOF for this model manually, ie calculate sum of test component results for CJS in UCARE/RELEASE and subtract test 3SR result from it, then see if result is significant? If this is how I go about it, do I subtract female 3SR or both?
Despite going through the manuals, it is still not clear to me how I run GOF for models other than simple CJS in UCARE.

Many many thanks
Eleni :D
Eleni Papadatou
 
Posts: 6
Joined: Mon Apr 04, 2005 11:48 am

Re: modelling transients: c-hat estimates

Postby cooch » Tue Apr 05, 2005 1:35 pm

Actually, this should have been appended to the end of your earlier thread, but no matter...

Also, many of these questions are in fact answered in Chapter 5 - which I'd suggest you read carefully (more carefully than you suggest you did...)

Eleni Papadatou wrote:- why not use bootstrap c-hat as well as the others?


Simple - it doesn't work as well as everyone had hoped. Discussed on p. 25 of Chapter 5.

- I ran median c-hat and the estimate was just below 1 (0.93): if I use this approach, then should I set c-hat=1 and "hold my nose"? :lol:
What if c-hat from UCARE/RELEASE is well above 1 (have not calculate yet, see question below)?


If estimated c-hat is <1, then use 1. Top of p. 3 (grey 'thought box').

- do I run GOF for this model manually, ie calculate sum of test component results for CJS in UCARE/RELEASE and subtract test 3SR result from it, then see if result is significant? If this is how I go about it, do I subtract female 3SR or both?


Use U-CARE - easier than doing things manually.

Despite going through the manuals, it is still not clear to me how I run GOF for models other than simple CJS in UCARE.


Thats a U-CARE documentation issue, not a MARK issue. So, simply use RELEASE, and median c-hat, and take it from there.

But, there is a specific test for transience in U-CARE - it is calculated manually. See image on p. 17 of Chapter 5 for an example.
cooch
 
Posts: 1654
Joined: Thu May 15, 2003 4:11 pm
Location: Cornell University


Return to analysis help

Who is online

Users browsing this forum: No registered users and 2 guests

cron