search

bobmuscarella / Panama_analysis

1 stars 0 forks source link

Uneven sample size for BCI vs. other plots #8

Open bobmuscarella opened 9 years ago

bobmuscarella commented 9 years ago

This could be an issue, especially for survival because cocoli only has ~100 dead trees...

I should probably focus on growth as there are more observations.

I wonder if subsampling BCI would be (a) appropriate and (b) a good idea...

bobmuscarella commented 9 years ago

I need to try to run each plot separately in a 2-level model and see what kind of parameter estimates and convergence I get. Then I can consider pooling them into the full 3-level model. I think it is unlikely to be reasonable to pool with survival because of too few deaths in, especially, Cocoli...