Multi-state model and transition probabilities

questions concerning analysis/theory using program MARK

Multi-state model and transition probabilities

Postby RBoys » Thu Sep 14, 2017 11:14 am

I am modelling data in the multi state open robust design, where I have one unobservable and one observable state. However, my survival estimates are confounded with the transition probabilities.

In this project survival estimates are importnat, but transition would be useful but is not necessary. My question is if I set transition probabilities to equal 0, the same as the p, pent parameters for the unobservable state, what am I saying about the population in the model?
Does it just mean that transition is not being modelled? Or am I saying that there is no transition at all? As I need it to be modelled for animals to be emmigrating and immigrating.

Thanks
RBoys
 
Posts: 18
Joined: Fri Jun 17, 2016 9:18 am

Re: Multi-state model and transition probabilities

Postby Dan_Gibson » Sat Sep 23, 2017 10:12 pm

Fixing the transition probability to zero constrains that transition to be not possible. Off the top of my head, I am not certain how MARK handles situations where an impossible transition is observed. I am sure there is warning message involved, but I am not sure if causes the model to terminate.

Instead of fixing the movement parameters to zero, I would suggest try time-constant models for both the movement parameters. If that is not successful, constrain both movement parameters to be identical. If that doesn't work, I think you may have to consider whether you have sufficient data for the model.
Dan_Gibson
 
Posts: 5
Joined: Sun Feb 28, 2016 2:50 pm


Return to analysis help

Who is online

Users browsing this forum: No registered users and 12 guests