CJS Model Unequal Time Intervals Coding Correctly

posts related to the RMark library, which may not be of general interest to users of 'classic' MARK

CJS Model Unequal Time Intervals Coding Correctly

Postby cb » Thu Feb 13, 2020 7:04 pm

I am having trouble understanding indexing of parameters with a CJS model with unequal time intervals, and haven't found a clear definition of this.
I am running a CJS model with 4 samples per summer (June-Sept, 1 month apart) over 3 years, so 9 months between last summer survey of year 1 and first summer survey of year 2.
Here is my code
Code: Select all
timint = c(1,1,1,9,1,1,1,9,1,1,1)
fil.proc=process.data(cjscmbW, model="CJS", time.intervals=timint)
fil.dll=make.design.data(fil.proc)


If I look at the structure in the .dll I am unclear of the interpretation. I understand that the estimates are monthly estimates, but I am not sure which of the values reflects the summer versus winter (monthly) estimates. (sorry, that is a mouthful I know :{ )
I would assume time 1 is the survival from June-July in year 1, time 2 is the survival from June-July in year 1, and then is time 4 the survival from Sept year 1 to June year 2? And 16 is the survival from Sept year 2 to June of year 3? And 27 is the survival from Aug year 3 to Sep year 3?
    > fil.dll$Phi$time
    [1] 1 2 3 4 13 14 15 16 25 26 27 2 3 4 13 14 15 16 25 26 27 3 4 13 14 15 16 25 26 27 4
    [32] 13 14 15 16 25 26 27 13 14 15 16 25 26 27 14 15 16 25 26 27 15 16 25 26 27 16 25 26 27 25 26
    [63] 27 26 27 27
    Levels: 1 2 3 4 13 14 15 16 25 26 27

I want to add a seasonal covariate and I am not sure which index to assign this to.
Would the winter monthly estimate be indexes 4 and 16 or 13 and 25?

My time.intervals coding seems intuitively correct, but am I missing something?

Thanks very much!
cb
 
Posts: 13
Joined: Mon Jan 04, 2010 2:29 pm

Re: CJS Model Unequal Time Intervals Coding Correctly

Postby jlaake » Thu Feb 13, 2020 8:28 pm

Please read the workshop notes carefully. The value of time is not an index. It is a time but is a factor variable. The default begin.time is 1 and it adds time.interval values to get the ensuing times. As it states in the documentation, survival is an interval parameter because it relates to the interval between occasions and interval parameters are labelled by the time at the beginning of the interval. The first one is for the interval from 1 to 2 and is labelled with 1. The one with 4 is the beginning time of the interval from 4 to 13.
jlaake
 
Posts: 1417
Joined: Fri May 12, 2006 12:50 pm
Location: Escondido, CA


Return to RMark

Who is online

Users browsing this forum: No registered users and 17 guests