Sample Size for Known-fate Models With Individual Covariates

questions concerning analysis/theory using program MARK

Sample Size for Known-fate Models With Individual Covariates

Postby tracker » Thu Feb 22, 2018 12:38 pm

What is the effective sample size you are supposed to use for known-fate modeling when coding encounter histories using groups and individual covariates (e.g. body condition)?

I had originally thought that my effective sample size should be according to the number of binary encounters across all of the intervals for all of the individuals (which is what MARK automatically uses unless I specify otherwise), but after looking at information in the help index about saturated models, it says that "for any of the models with individual covariates, the sample size for each encounter history is 1." Should I be correcting MARK by telling it the sample size is actually equal to my number of individual animals for all of the models I am trying (some of which contain individual covariates and others that do not), or should I leave it as is?

Thanks in advance for any advice!
tracker
 
Posts: 5
Joined: Sun Jan 07, 2018 6:03 pm

Re: Sample Size for Known-fate Models With Individual Covari

Postby Bryan Hamilton » Thu Feb 22, 2018 9:17 pm

The general Rule of Thumb in linear regression is 5-10 observations per modeled parameter. Like all statistical rules, its made to be broken. More observations per parameter is always better in that you increase your power and decrease confidence estimates. An individual covariate counts as one parameter. Each group counts as one parameter. So you can see the advantage of using a continuous individual covariate rather than discrete groups in increasing power.

I don't think you need to correct MARK for sample size... Sometimes MARK gets the parameter estimates wrong and those can be corrected but I don't know of a mechanism to adjust sample size....
Bryan Hamilton
 
Posts: 111
Joined: Thu Jun 15, 2006 11:36 am
Location: Great Basin National Park

Re: Sample Size for Known-fate Models With Individual Covari

Postby tlyons4 » Fri Feb 23, 2018 1:17 pm

I think the sample size=1 refers to the coding in the INP file. If I remember right, the known-fate (and others) model wants a column that tells MARK how many individuals have that encounter history. When you start plugging in individual covariates, you should no longer be aggregating encounter histories and set that column to 1.

This is different than the sample size MARK uses for calculating AICc.
tlyons4
 
Posts: 19
Joined: Thu Nov 30, 2017 7:29 pm


Return to analysis help

Who is online

Users browsing this forum: No registered users and 15 guests