Nest Survival distribution and covariates help

[jlaake]

Fair point as long as all of the models converged. Sometimes even if a model has little weight, its estimate for the std error can screw up model averaged estimates. If you want to limit model averaging, remove the models you don't want from the marklist and then model average. See remove.mark.--jeff

[cooch]

Fair point as long as all of the models converged. Sometimes even if a model has little weight, its estimate for the std error can screw up model averaged estimates.

That would be the

unless there are clear and defensible structural reasons why models 4 -> N are 'invalid' in some fashion...

I referred to. Lack of convergence (wonky SE, from the Latin) being such a 'structural reason'.