Adding photographic effort to CJS model using design matrix

questions concerning analysis/theory using program MARK

Adding photographic effort to CJS model using design matrix

Postby Tamlyn » Tue Feb 21, 2017 6:32 pm

Hi everyone,

I'm having some design matrix issues in my analysis at the moment. I am running CJS models using a long term data set of 20 years for sperm whales. My top model so far is phi(.)p(t) + b - 'b' is a behavioural effect on the first capture occasion - so in my survival pim, for each cohort, the first occasion is specified as "1" and then the subsequent occasions as "2". This is to incorporate a possible behavioural response to initial boat approach on an individual. I am wanting to also add photographic effort to this model, to see the effect. I am adding the raw number of photo-ID appropriate photos taken in each period. The design matrix looks correct to me but I am coming out with exactly the same results as the model without the effort, and the same number of parameters - so the effort parameter is not getting picked up somehow. I am probably just missing something simple but I can't seem to figure out what. I've read the relevant chapters and played around with it quite a bit but maybe some fresh eyes can pick something up! I've also attached my PIMS as well as the design matrix. Thanks in advance!

Image
Image
Image
Tamlyn
 
Posts: 4
Joined: Mon Feb 20, 2017 9:04 pm

Re: Adding photographic effort to CJS model using design mat

Postby cooch » Tue Feb 21, 2017 7:17 pm

Your DM for encounter probability violates the 'first commandment' (pull down the MARKK Help menu, and select 'Ten Commandments')- : Thou shalt not use a temporal covariate with a fully time-varying model..

If you have an unstrctured model for time (i.e.,, p(t)), then that completely explains (models) the variation, such that adding a temporally varying covariate on top does nothing. Simply build two sepeerate models (i) p(t), and (ii) p(effort), and compare the two. If the latter gets more support, then you might take this as evidence that the temporal hetergoeneity is explained to some degree by hours.
cooch
 
Posts: 1628
Joined: Thu May 15, 2003 4:11 pm
Location: Cornell University

Re: Adding photographic effort to CJS model using design mat

Postby Tamlyn » Tue Feb 21, 2017 9:20 pm

Thanks so much for your reply, cooch. Hmmmm, I see... So am I right in understanding that an additive model where both time and the number of photographs each year are included as covariates, but not an interaction term between them, is a violation of this rule? Or have I got multiple modelling techniques confused?
Tamlyn
 
Posts: 4
Joined: Mon Feb 20, 2017 9:04 pm

Re: Adding photographic effort to CJS model using design mat

Postby cooch » Tue Feb 21, 2017 9:24 pm

Tamlyn wrote:Thanks so much for your reply, cooch. Hmmmm, I see... So am I right in understanding that an additive model where both time and the number of photographs each year are included as covariates, but not an interaction term between them, is a violation of this rule? Or have I got multiple modelling techniques confused?


You have a couple of things confused. You can have additive of 'group' (fixed attribute) and time, say, but not a temporally varying covariate additive to temporally varying...time.

So,

p =f(time)

p =f(hours)

but not p=f(time+hours).
cooch
 
Posts: 1628
Joined: Thu May 15, 2003 4:11 pm
Location: Cornell University

Re: Adding photographic effort to CJS model using design mat

Postby Tamlyn » Tue Feb 21, 2017 9:42 pm

Awesome, I've got it. Thanks again for the help!
Tamlyn
 
Posts: 4
Joined: Mon Feb 20, 2017 9:04 pm


Return to analysis help

Who is online

Users browsing this forum: No registered users and 8 guests