Model averaging for beta parameters in occupancy models

[mspinola]

Dear list members,

Is it possible to do model averaging for the beta parameters when using covariates (through the model averaging option in the menu) for occupancy models in MARK?
It seems like is possible to do model averaging for only psi and p.
Thank you very much in advance.
Best,

Manuel Spínola

[cooch]

Dear list members,

Is it possible to do model averaging for the beta parameters when using covariates (through the model averaging option in the menu) for occupancy models in MARK?
It seems like is possible to do model averaging for only psi and p.
Thank you very much in advance.
Best,

Manuel Spínola

Have a look at

http://www.phidot.org/forum/viewtopic.php?t=996

Some the relevant issues are discussed in that topic.

[conroy]

Dear Manuel-
To my knowledge only 'real' or derived parameters can be averaged in MARK, not the design matrix coefficients.

The computations to compute model-averaged estimates and unconditional variances are fairly easy to do in a spreadsheet program. An example is provided as part of a homework assignment for my class; it is in the context of known fate models, but the ideas apply:
http://coopunit.forestry.uga.edu/FORS83 ... wn/assign7

Note that there can be some debate about exactly how to do averaging for coefficients that only appear in some models. My example assumes that the parameter exists in all models. See Burnham and Anderson 2nd ed for discussion on this.

Saludos,

Mike