Examples of gene-environment interaction in human behavioral data are relat
ively rare; those that exist have used simple, dichotomous measures of the
environment. The authors describe a model that allows for the specification
of more continuous, more realistic variations in environments as moderator
s of genetic and environmental influences on behavior. Using data from a po
pulation-based Finnish twin study, the authors document strong moderating e
ffects of socioregional environments on genetic and environmental influence
s on adolescent alcohol use, with nearly a five-fold difference in the magn
itude of genetic effects between environmental extremes. The incorporation
of specific environmental measures into genetically informative designs sho
uld prove to be a powerful method for better understanding the nature of ge
ne-environment interaction and its contribution to the etiology of behavior
al variation.