www.phidot.org

Dear all,

I have a problem when i want to make model averages.
I get this error:

Parameter Matrix not the correct length in output:p Session1
Returned length=0
Expected length=3

and so on....

I use robust design occupancy models with 22 primary and 86 secondary sessions. My models have the same p (all of them time-varying) and the time intervals are different between primary sessions.

Thank for the help and reply,
Robert

Buco wrote:I use robust design occupancy models with 22 primary and 86 secondary sessions. My models have the same p (all of them time-varying) and the time intervals are different between primary sessions.

Don't have time (or probably enough information) to reply to your main question, but - unequal intervals between primary sessions does pose a problem for RD models. See section 15.9 of the RD chapter (Chapter 15).

...so, for robust design data types (including the multi-state robust designs),
only survival rates are corrected with the time interval, but none of the transition parameters are
corrected. Again, user beware! Think carefully about what unequal time intervals may be doing to
your interpretation of the parameter estimates.

Thank you for the quick reply!

I'll read the section 15.9 of the RD chapter.

Buco wrote:Thank you for the quick reply!

I'll read the section 15.9 of the RD chapter.

The problem referred to (concerning unequal intervals) is subtle, and there is not universal agreement that there is either a real problem, or that it is significant. Gary tends to think so, and so do I. I think at this point the best advice is - think hard about what you're doing, and be cautious. Unequal intervals can (at least in our view) complicate interpretation of some parameters in 'multi-state' models, and as Bill Kendall notes in his chapter (15), a RD model is essentially a multi-state model by any other name.