Kevin E. answered • 10/02/19
A chemist going back to school for cyber security.
I would say as a first study, you could do two (or possibly three) different studies with stratified samples.
The first group is individuals within similar environments and stratifying based on biological differences. Then measure the differences in the trait you're focusing on. This will test whether there or not there is some sort of biological component, since the environment will be relatively constant.
The second group is individuals with similar biological markers and stratified based on environment. Then measure the differences in the trait you're focusing on. This will test whether there or not there is some sort of environmental component.
You'll likely see that both of these play a role, but this will at the very least tell you which one is stronger.
If you wanted to put in a third group, that one will be something similar to a stratified clustered random sample. (Yes, it is possible to combine two or more methods, but one needs to be exceedingly careful when doing so.) Have different clusters from different environments. Then stratify each cluster based on biology. This will tell you if there is an interaction between biology and environment by comparing differences between environment alone, biology alone, and the third group of both biology and environment. If the third group is greater than (or possibly less than) the sum of the other two, then there is a significant interaction term.
