RMark not using all capture histories on call to MARK

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

RMark not using all capture histories on call to MARK

Postby Eurycea » Tue May 23, 2023 12:22 pm

I couldn't figure out why RMark and MARK gave me different results using Multistrata models. My best guess is that something is going wrong when RMark passes the data to MARK. The RMark output says

"Number of unique encounter histories read was 321."

But I have 4295 encounter histories, which is correctly reflected when I use the MARK GUI.

Within R, both the input data and processed data report the correct dimensions when using the functions convert.inp and process.data.

The model specification doesn't really seem to matter; I tried a bunch of stuff. Robust design models with similar data seem to run just fine, but once I do the Multistrata type, I get this problem.

I've looked through the help a few times and I just cannot figure out what I am doing wrong, if anything. The example in the appendix of the Gentle Intro runs fine. I've tried changing the coding of my strata (there are only two states). I'm using the same input file as I use in the MARK GUI. Just not sure what is happening here. Please help.


Here's an example of a call to run a model.

Code: Select all
run.mstrata3=function()
{
  ms.dp <- process.data(df,model="Multistrata")
  ms.ddl <- make.design.data(ms.dp)
 
  #create formula
  Psi.stratum = list(formula =~-1+stratum:tostratum,link = "logit")
  p.stratum=list(formula=~stratum)
  S.stratum=list(formula=~stratum)
  model.list=create.model.list("Multistrata")

  mstrata.results=mark.wrapper(model.list,data=ms.dp,ddl=ms.ddl)

  return(mstrata.results)
}
ms.results3 = run.mstrata3()
Eurycea
 
Posts: 99
Joined: Thu Feb 25, 2010 11:21 am

Re: RMark not using all capture histories on call to MARK

Postby Eurycea » Mon Jun 05, 2023 8:18 am

Well we can chalk this one up to "user error" as per usual...

RMark summarizes the CH in the data file, which explains the difference in number of observations.

Jeff helped me figure out that I actually had different time intervals set in each file, and that this impacts the model. I did not know time intervals would affect AIC! Thanks Jeff!
Eurycea
 
Posts: 99
Joined: Thu Feb 25, 2010 11:21 am


Return to RMark

Who is online

Users browsing this forum: Bing [Bot] and 16 guests