www.phidot.org

Hi,
I'm using a Robust design with Closed models for secondary periods with heterogeneity (RDHet) in R to look at survival in a mark-recapture study of rodents. I've figured out how to make the model variables be a function of the animal's sex, primary trapping session, or simply constant, but I'm curious if there's a way to set one of these variables to zero. In particular, I would like to impose the assumption that immigration=emigration=0, so I need to set GammaPrime and GammaDoublePrime to zero.

So something like this (which makes them constant):
##GammaPrime.dots=list(formula=~1,share=TRUE) #Immigration=emigration and is time-invariant

But that makes them equal to zero. I've tried (formula=~0) and (formula=0) but thus far with no luck.

Any input is much appreciated.

Please post questions about RMark in the RMark section.

You should read the section about fixing real parameters. Used fixed=0 in your model specification and it will set them all to 0.

--jeff

jlaake wrote:Please post questions about RMark in the RMark section.

You should read the section about fixing real parameters. Used fixed=0 in your model specification and it will set them all to 0.

--jeff

Agreed - I will move this thread to the RMark sub-forum after this post.

Sorry for the mislabeling and thanks for moving the thread.

So my code right now looks like this and it still isn't setting GammaPrime and GammaDoublePrime to zero - it makes them constant, but above zero.

**************************
CFC1_2005.model<-function() {
CFC1_2005 <- import.chdata("2005_CFC1.txt", header=TRUE)
CFC1_2005.envi.S <- as.data.frame(read.table("2005_CFC1_envi_S.txt",header=TRUE))
CFC1_2005.processed <- process.data(CFC1_2005,begin.time=1,model="RDHet",groups=c("treat","spec"),time.intervals=CFC1_2005.t.intervals)
CFC1_2005.ddl <- make.design.data(CFC1_2005.processed,parameters=list(S=list(pim.type="time")))

CFC1_2005.envi.S$time=CFC1_2005.ddl$S$time[1:6]
CFC1_2005.ddl$S=merge_design.covariates(CFC1_2005.ddl$S, df=CFC1_2005.envi.S, bygroup=F, bytime=T)

S.treat.spec.t.gap=list(formula=~treat*spec*t.gap)
p.treat.session=list(formula=~treat*session) #Survival as a function of treatment and T.interval
GammaPrime.zero=list(fixed=0,share=TRUE) #Immigration=emigration and is time-invariant

cml=create.model.list("RDHet")
return(mark.wrapper(cml,data=CFC1_2005.processed,ddl=CFC1_2005.ddl,adjust=T,realvcv=F,external=T))
}
*************************************

The solution was to include the 'formula=~1' portion to set it constant, and then 'fixed=0' to set it equal to zero:

GammaPrime.zero=list(formula=~1,fixed=0)
GammaDoublePrime.zero=list(formula=~1,fixed=0)

Thanks for the input, everybody.