turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

Find a Community

- Home
- /
- Analytics
- /
- Stat Procs
- /
- GLIMMIX with binary data

Topic Options

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

01-12-2010 11:13 PM

Dear forum members,

I have a dataset of bird site fidelity. The structure is as follows:

BirdID Year Return Sex Success

--------------------------------------------------------

F0215 2005 1 M 1

F0217 2005 0 F 0

........

"BirdID" is the ring number, "Return" is if the bird return to the nest site of previous year, "Success" is if the bird bred successfully in the previous year. As the dataset was collected over several years, the same bird may occur several times in the dataset. Therefore I think of using random effect in GLIMMIX.

For the question: if previous successful breeders tend to return to the same nest site, the following code was used:

model Return (event='1') = Success / dist=binary link=logit;

random BirdID;

If the reponse variable is numeric, one can see the Type III Test of Fixed Effect in the output to know if the fixed effect is significant. However,as the Type III Test is a F test based on normal distribution, does it apply to such case, in which the responese is binary? I was told that one should use chi-square test to test for significance of explanatory variables in mixed models. How to conduct such a test with GLIMMIX?

Any suggestion is deeply appreciated.

May

I have a dataset of bird site fidelity. The structure is as follows:

BirdID Year Return Sex Success

--------------------------------------------------------

F0215 2005 1 M 1

F0217 2005 0 F 0

........

"BirdID" is the ring number, "Return" is if the bird return to the nest site of previous year, "Success" is if the bird bred successfully in the previous year. As the dataset was collected over several years, the same bird may occur several times in the dataset. Therefore I think of using random effect in GLIMMIX.

For the question: if previous successful breeders tend to return to the same nest site, the following code was used:

model Return (event='1') = Success / dist=binary link=logit;

random BirdID;

If the reponse variable is numeric, one can see the Type III Test of Fixed Effect in the output to know if the fixed effect is significant. However,as the Type III Test is a F test based on normal distribution, does it apply to such case, in which the responese is binary? I was told that one should use chi-square test to test for significance of explanatory variables in mixed models. How to conduct such a test with GLIMMIX?

Any suggestion is deeply appreciated.

May