Simulate trends in Known Fate

questions concerning analysis/theory using program MARK

Simulate trends in Known Fate

Postby sixtystrat » Wed Dec 26, 2012 10:10 am

Is there any way to simulate a linear trend in survival rates in the Known Fate simulation routine? I tried to import a design matrix as my true model and Mark crashes when I do so. Thanks.
Joe
sixtystrat
 
Posts: 117
Joined: Wed Jun 01, 2011 9:19 am

Re: Simulate trends in Known Fate

Postby gwhite » Wed Dec 26, 2012 10:28 am

Joe:
This is one of the "gotchas" that I don't know a good way around. The issue is that the Known Fate simulator does not keep animals in the sample for more than one occasion. So, you may think you are simulating data for 5 years as 1010101010 if the animal were to live all 5 years. However, MARK is actually generating a encounter of 1000000000 unless you have specified new animals to be released on occasions 2 through 5. In other words, even though an animal lives through the interval where it was released, it is not re-released into the next interval. The reason for this behavior is that I don't want to release an animal with a 3-year radio collar but have it survive for more than 3 years.

You can specify the number of animals to release on each occasion and approximately mimic what you want. For example, suppose you have a survival rate of 0.9. Release 100 animals on occasion 1, 90 on occasion 2, and 81 on occasion 3. But, the releases are the expected number of survivors, not the actual number of survivors.

So here is another work around -- use the Burnham joint live-dead data type to simulate and analyze data, but recognize that you can have animals out-living the life of their collar. Set p=1, r=1, and F=1 to make the Burnham joint live-dead equivalent to the known fate data type.

Gary
gwhite
 
Posts: 340
Joined: Fri May 16, 2003 9:05 am

Re: Simulate trends in Known Fate

Postby cooch » Wed Dec 26, 2012 11:29 am

gwhite wrote: So here is another work around -- use the Burnham joint live-dead data type to simulate and analyze data, but recognize that you can have animals out-living the life of their collar. Set p=1, r=1, and F=1 to make the Burnham joint live-dead equivalent to the known fate data type.

Gary


See section 16.6 in 'the book'.
cooch
 
Posts: 1654
Joined: Thu May 15, 2003 4:11 pm
Location: Cornell University

Re: Simulate trends in Known Fate

Postby sixtystrat » Wed Dec 26, 2012 12:10 pm

Ok, thanks. One issue that I am having though is that to create a linear trend (annual), I have to use the design matrix. I set it up as follows for S with animals monitored monthly for 5 years:
1 1
1 1
...
1 1
1 2
1 2
...
1 2
1 3
1 3
...
...
1 5

I save that matrix and when I try to import it into the simulation as the true data type, MARK crashes. I want to simulate an
sixtystrat
 
Posts: 117
Joined: Wed Jun 01, 2011 9:19 am

Re: Simulate trends in Known Fate

Postby sixtystrat » Wed Dec 26, 2012 12:11 pm

Ooops! Disregard that last partial sentence.
sixtystrat
 
Posts: 117
Joined: Wed Jun 01, 2011 9:19 am


Return to analysis help

Who is online

Users browsing this forum: No registered users and 2 guests

cron