Design matrix and individual covariates

questions concerning analysis/theory using program MARK

Design matrix and individual covariates

Postby nesmonde » Wed May 11, 2022 6:30 am

In Chapter 11.8.2 Model averaging with environmental covariates there is an example data set for estimating encounter probability with sex and number of hours of observation

Image

For the past day or two I've been trying to create the last model Phi(t)p(SEX+HOUR+SEX*HOUR) in the design matrix

Image

however the AIC I get is a little higher than the one in the chapter. Could anyone help me resolve this?

Many thanks.
nesmonde
 
Posts: 7
Joined: Tue Nov 02, 2021 4:29 pm

Re: Design matrix and individual covariates

Postby cooch » Wed May 11, 2022 7:46 am

nesmonde wrote:In Chapter 11.8.2 Model averaging with environmental covariates there is an example data set for estimating encounter probability with sex and number of hours of observation Image. For the past day or two I've been trying to create the last model Phi(t)p(SEX+HOUR+SEX*HOUR) in the design matrix Image however the AIC I get is a little higher than the one in the chapter. Could anyone help me resolve this?

Many thanks.


Sure, because your attempted DM isn't even close to correct. Should look something like the following (if you can't see the whole image, right-click the image an open in a new tab - might help). If you don't know why *your* attempted DM is incorrect (whereas the one below is), I'd suggest reviewing Chapter 6, and then the first sections of chapter 11).

Image
cooch
 
Posts: 1628
Joined: Thu May 15, 2003 4:11 pm
Location: Cornell University

Re: Design matrix and individual covariates

Postby nesmonde » Thu May 12, 2022 8:39 am

Ahh ok, I see where I have gone wrong. Thank you very much for your help.
One more query - Phi 6 does not have an interval marked with 1... is that because it is confounded with the first encounter parameter?
nesmonde
 
Posts: 7
Joined: Tue Nov 02, 2021 4:29 pm

Re: Design matrix and individual covariates

Postby cooch » Thu May 12, 2022 9:00 am

nesmonde wrote:Ahh ok, I see where I have gone wrong. Thank you very much for your help.
One more query - Phi 6 does not have an interval marked with 1... is that because it is confounded with the first encounter parameter?


No - it has nothing to do with confounding. The DM for phi is using offset (reference-cell) coding, and the lack of a '1' for the final time interval indicates that interval 6 will be the reference (intercept), and that each individual non-intercept beta is an 'offset' relative to this intercept.

Based on your question, you *really* need to read/study chapter 6, before you do anything else.
cooch
 
Posts: 1628
Joined: Thu May 15, 2003 4:11 pm
Location: Cornell University

Re: Design matrix and individual covariates

Postby nesmonde » Fri May 13, 2022 6:13 am

Thank you. Yes I think I even need to go back to school haha!
nesmonde
 
Posts: 7
Joined: Tue Nov 02, 2021 4:29 pm


Return to analysis help

Who is online

Users browsing this forum: No registered users and 2 guests