Sex determined only at first capture at age=0

questions concerning analysis/theory using programs M-SURGE, E-SURGE and U-CARE

Sex determined only at first capture at age=0

Postby simone77 » Tue Feb 21, 2012 7:05 am

Hi,

I am dealing with a CMR data set with these characteristics:
- Individuals are marked, and a part of them sexed, by chicks (gender is determined only when they are first captured as chicks) at the breeding colony
- Observations are made at the breeding colony of individuals born there.
- 11 occasions throughout 13 years (two drought years with no breeding)
- Ca. 1/3 of captured individuals have been sexed.

I am interested in two analyses: one for estimating the apparent survival (survival*phylopatry) and the other to estimate the age of first breeding.

To do that I would implement respectively a CJS and a reverse time (Pradel) approach to estimate seniority (and derive probabilities of ages of first breeding).

I consider that gender is an important effect to test both on apparent survival and on seniority. For this reason, I wonder if there is a way to approach this without having to make data-censoring or splitting the data set in three groups (males, females, and unknown). I thought that perhaps I might prepare a data set for estimating apparent survival made of encounter histories with 4 events like these:

00100101000 1;
...
00000020200 1;
...
30003330003 1;

where 1=males, 2=females, 3=unknown.

I might do just the same by reverse encounter histories (for example in U-CARE) and use that file to estimate seniority as it would be a CJS (on reverse-time encounter histories). I might define a number of age classes that make sense according to the number of cohorts (10) and on a biological criterion and go on from there.

Question 1: does it make sense to use E-SURGE in this context? I am not sure at all, but in case it would, this is how I would configure pattern matrices.
Two steps process for Transition, a first one being Survival:
step1 (Survival)
s-*
-s*
--*
and a second one being Transition with parameters forced to be 1 like this:
step2 (Transition)
*--
-*-
--*
and a two steps process for Encounter
step1 (Capture)
*c-
*-c
*--
and:
step2 (State assignment)
*---
-a-*
--a*

Question 2: In case all of this made sense, would it make sense to estimate seniority this way?

Question 3: how should I define event two GEMACO sentences for both analyses if I wanted to consider a time and gender varying probability of state assignment?

The good part of this long post (sorry!) is that the answer might be very very short like answering “no” at question 1.

Thanks in advance for any help

Simone
simone77
 
Posts: 197
Joined: Mon Aug 10, 2009 2:52 pm

Re: Sex determined only at first capture at age=0

Postby CHOQUET » Tue Feb 21, 2012 10:01 am

Replies in bold

Question 1: does it make sense to use E-SURGE in this context?

YES

I am not sure at all, but in case it would, this is how I would configure pattern matrices.
Two steps process for Transition, a first one being Survival:
step1 (Survival)
s-*
-s*
--*
and a second one being Transition with parameters forced to be 1 like this:
step2 (Transition) NOT NEEDED

*--
-*-
--*

and a two steps process for Encounter
step1 (Capture)
*c-
*-c
*--
and:
step2 (State assignment)
*---
-a-*
--a*

Question 2: In case all of this made sense, would it make sense to estimate seniority this way?

Probably yes

Question 3: how should I define event two GEMACO sentences for both analyses if I wanted to consider a time and gender varying probability of state assignment?

t for time
from or to for state effect

For the language, see the paper

Choquet R. (2008). Automatic generation of multistate capture recapture models. The Canadian Journal of Statistics 36(1): 43-57.
CHOQUET
 
Posts: 211
Joined: Thu Nov 24, 2005 4:58 am
Location: CEFE, Montpellier, FRANCE.

Re: Sex determined only at first capture at age=0

Postby simone77 » Wed Feb 22, 2012 7:03 am

Question 2: In case all of this made sense, would it make sense to estimate seniority this way?

Probably yes

How probable is it? :)
My idea would be to prepare a file as specified in Pradel et al. 1997 (Local recruitment in the greater flamingo: a new approach using capture-mark-recapture data), i.e. removing first observation (marking), reversing the capture histories, separating each cohort as if would be a separate group. After that, I would use the same pattern matrices above specified for the CJS analysis (yes I would remove the unnecessary second step transition). My goal is to get estimates more robust mainly from the point of view of possible effects of gender, does it make sense this way? any suggestion?

Question 3: how should I define event two GEMACO sentences for both analyses if I wanted to consider a time and gender varying probability of state assignment?

t for time
from or to for state effect

For the language, see the paper

Choquet R. (2008). Automatic generation of multistate capture recapture models. The Canadian Journal of Statistics 36(1): 43-57.


Thanks for the reference, I had not read it and it is good to me to know it.
Anyway, I was wondering more about how to specify the step 2 of event in terms of ages (in both analyses), I am a bit confused about that because sex assignment is done at first - marking - occasion, and after that each individual has a "fixed" probability (0 or 1) to be correctly sexed.
May you make me an example of how you would write a general GEMACO sentence in this case for this step (also for seniority analysis given that first occasion is removed)?

Again, thanks a lot

Simone
simone77
 
Posts: 197
Joined: Mon Aug 10, 2009 2:52 pm

Re: Sex determined only at first capture at age=0

Postby CHOQUET » Thu Feb 23, 2012 10:41 am

My idea would be to prepare a file as specified in Pradel et al. 1997 (Local recruitment in the greater flamingo: a new approach using capture-mark-recapture data), i.e. removing first observation (marking), reversing the capture histories, separating each cohort as if would be a separate group. After that, I would use the same pattern matrices above specified for the CJS analysis (yes I would remove the unnecessary second step transition). My goal is to get estimates more robust mainly from the point of view of possible effects of gender, does it make sense this way? any suggestion?

Yes, it make sense.

Removing the first observation only if there are chicks
at first capture and separate cohorts to keep track of the age.



Anyway, I was wondering more about how to specify the step 2 of event in terms of ages (in both analyses), I am a bit confused about that because sex assignment is done at first - marking - occasion, and after that each individual has a "fixed" probability (0 or 1) to be correctly sexed.
May you make me an example of how you would write a general GEMACO sentence in this case for this step (also for seniority analysis given that first occasion is removed)?

for capture: used the sentence firste+nexte.[f.t]
for assigment: f.t

f.t can be replaced by any effect, for example f
CHOQUET
 
Posts: 211
Joined: Thu Nov 24, 2005 4:58 am
Location: CEFE, Montpellier, FRANCE.


Return to analysis help

Who is online

Users browsing this forum: No registered users and 8 guests

cron