www.phidot.org

Jeff helped me work this out today so I am just posting this so it will be archived for future searching, feel free to ignore.

If you are using the multistrata model in RMark and you have been using the pim.type adjustment listed in the RMark guide on page C-48 (current version) to speed up your computations, this does have a downside in that your .ddl for the Psi parameter will not have any age structure associated with it. As such, you will not be able to, for instance, manipulate the .ddl to model age-specific strata to strata transitions without some additional, rather complicated manipulation that may botch other things up. Using the below and the example mstrata data, you can see the difference in the .ddl dependent upon whether or not you specify pim.type="time" in the make.design.data() call.

This will lose the age values in the .ddl

Code: Select all

data(mstrata)
mp=process.data(mstrata,model="Multistrata")
md=make.design.data(mp, parameters=list(Psi=list(pim.type="time")))
head(md$Psi)


Compared to this which will not lose the age values in the .ddl:

Code: Select all

data(mstrata)
mp=process.data(mstrata,model="Multistrata")
md=make.design.data(mp)
head(md$Psi)


Jeff added this to his update list for the RMark guide.

bret