Nest Survival Random Effects Modelling

questions concerning analysis/theory using program MARK

Nest Survival Random Effects Modelling

Postby TaylorBrown » Mon Dec 09, 2019 5:18 pm

Nest survival in MARK - can random effects modelling even be done with nest survival data?
I want to make 'year' into a random effect (I have two years; 2018 and 2019).

I've done this:
- I input year as two separate "groups" of data in the .inp file
- I am not interested in differences in DSR within a season, so in the parameter index matrices I made 2018 all constant at 1 and 2019 all constant at 2.
- I built all my models with year as a fixed effect, using identity matrices.
- I followed MARK book Appendix D for creating fixed effects models - from the output window I selected the model of interest, and then in MARK's menu selected Output > Specific Model Output > Variance Components > Real Parameter Estimates > [Select the two parameter indices, one for 2018 and one for 2019] > “Random Effects Model”
- I then get a popup window which says, "Only 2 parameters were specified for estimating variance components!" and then when I click "OK", another popup saying "Error Code: 50 [ ] / Subsystem: VO-CODE / Error Subcode: 5333 / Argument Number: 2 / etc etc etc" and the only option it gives me is to "Abort". It then closes the entire program.

I'm not sure what I'm doing wrong, perhaps because I'm not strong with statistics.

Any help would be greatly appreciated! :D
TaylorBrown
 
Posts: 2
Joined: Tue Oct 09, 2018 1:18 pm

Re: Nest Survival Random Effects Modelling

Postby cooch » Mon Dec 09, 2019 5:30 pm

Not really -- with only 2 years, you can't reasonably estimate mu and sigma. The general recommendation is at least 10 years, 15 better. [And this recommendation is not dependent on whether you use the 'method of moments' approach, outlined in Appendix D, or the 'cool kids' Bayesian methods -- Appendix E).
cooch
 
Posts: 1628
Joined: Thu May 15, 2003 4:11 pm
Location: Cornell University

Re: Nest Survival Random Effects Modelling

Postby TaylorBrown » Wed Dec 18, 2019 9:52 pm

Thank you for the succinct and reasonable answer!
TaylorBrown
 
Posts: 2
Joined: Tue Oct 09, 2018 1:18 pm


Return to analysis help

Who is online

Users browsing this forum: Bing [Bot] and 13 guests