Data Dredging?

[EmmaRigby]

Hello,

I'm running a multistate analysis for four groups (adult male, juvenile male, adult female and juvenile female) over six locations (coded A-F) over 10 pooled seasons and am interested in looking at survival, recapture and transition rates between sites.

As there are so many site, group and season factors when I try to open a design matrix for anything other than the very basic models MARK crashes. The only way I have been able to get round this is to keep two of the three parameters (s, p and psi) very basic and then just modify models for the other one i.e. I'm looking at survival, recapture or transition rates separately.

I've read the posts in the thread:
http://www.phidot.org/forum/viewtopic.p ... t=dredg%2A

and know that this is hardly ideal, but as long as I am honest about the limitations in my methodology is it still ok?

Does anyone also have any suggestions of how I would tackle model averaging in order to get final estimates of s, p and psi - would I do it three times (one for each set of models looking at a particular parameter?)

Any feedback would be very gratefully received!
Thanks