"Invalid link function specified for the true model&quo

questions concerning analysis/theory using program MARK

"Invalid link function specified for the true model&quo

Postby jlr » Thu Oct 19, 2006 1:52 am

I am running a reduced MS model that uses the mlogit link to constrain transitions between six states (areas). The model gives reasonable results, except when I try to run median c-hat. In that case, I get the following error: "Invalid link function specified for the true model". The mlogit link is required for the model to converge and I don't have another good option for GOF. Has anyone experienced this and/or have a suggestion?
jlr
 
Posts: 6
Joined: Tue Mar 21, 2006 9:53 am

Re: "Invalid link function specified for the true model

Postby cooch » Thu Oct 19, 2006 7:00 am

jlr wrote:I am running a reduced MS model that uses the mlogit link to constrain transitions between six states (areas). The model gives reasonable results, except when I try to run median c-hat. In that case, I get the following error: "Invalid link function specified for the true model". The mlogit link is required for the model to converge and I don't have another good option for GOF. Has anyone experienced this and/or have a suggestion?


Interesting problem - not sure there is a simple solution, since at present, the mlogit (which must be user-specified) is not available in the simulation capabilities in MARK.

Hmmm....I suppose the only argument (and its a weak one) is to not use the mlogit model for your general model. Using the mlogit is useful if there is risk that the transition sums will be >1. If that isn't the case, then it may not be necessary to use it at all.
cooch
 
Posts: 1654
Joined: Thu May 15, 2003 4:11 pm
Location: Cornell University

Postby jlr » Thu Oct 19, 2006 7:15 am

Thanks for the response. I have six areas and seem to need the mlogit for the model to converge. I am using it to ensure that movement from area 1 to area 2+3+4+5+6 equals 1, etc. Perhaps there is another way to do this that I am not thinking of?
jlr
 
Posts: 6
Joined: Tue Mar 21, 2006 9:53 am

Postby jlr » Thu Oct 19, 2006 7:51 am

Actually, never mind that last comment. It does converge without the mlogit. I went down the mlogit path earlier in the analysis when I was working with a slightly different (inferior) input file. It didn't occur to me to check this again. Thanks, Evan.
jlr
 
Posts: 6
Joined: Tue Mar 21, 2006 9:53 am


Return to analysis help

Who is online

Users browsing this forum: No registered users and 0 guests