Variance components and model averaging

[hanno]

I did not find any possibility in MARK of obtaining variance components when using model averaging. Is this because i haven't searched hard enough or because it isn't implemented? (Or because wishing to obtain variance components is nonsense for averaged models??)

Of course i may be wrong, but i think it sounds meaningful to distinguish between process variance and sampling (plus model uncertainty?) variance not only for the single best model, but also for averaged survival estimates. Is there any easy way obtaining this, or to i have to calculate it by hand (using the ugly equations given in the help files)?

Hanno

[cooch]

I did not find any possibility in MARK of obtaining variance components when using model averaging. Is this because i haven't searched hard enough or because it isn't implemented? (Or because wishing to obtain variance components is nonsense for averaged models??)

Of course i may be wrong, but i think it sounds meaningful to distinguish between process variance and sampling (plus model uncertainty?) variance not only for the single best model, but also for averaged survival estimates. Is there any easy way obtaining this, or to i have to calculate it by hand (using the ugly equations given in the help files)?

Hanno

Always useful to answer a question that was posted many years ago. (Actually, the question might still be of interest to users even in more 'modern times'). The issue of 'model averaging' and 'variance components' is discussed in some detail in Appendix D.