[concordDance]

Haven't there been studies untangling causation here? Using random events that hit socioeconomic class as a natural experiment?

Adoption studies might also be informative.