|That friends resemble each other in their behaviors is not surprising. Much (though not all) of this resemblance appears to result from traits that they may have shared before coming into contact with each other. If "birds of a feather flock together," where do the feathers come from? And why do similar birds find each other? This study used the National Longitudinal Study of Adolescent Health (Add Health) to examine two contrasting explanations for the resemblance between adolescents on smoking and drinking behaviors: a family socialization hypothesis and a gene-environment correlation (also known as a gene->environment effect) hypothesis. Covariance matrices were computed on the friends of siblings, i.e., one score of a pair was sibling A's friend's substance use and the other was sibling B's friend's substance use. In a biometric model, parameter estimates yield a heritability of friends' substance use of 64%; the estimated nonshared environmental variation on substance use was zero. We interpret these results as indicating an active gene-environmental correlation, whereby genetic dispositions may lead to a selection of particular experiences, including substance-using peers. These indirect genetic effects may account for an association between children's family
background and their friends' substance use in adolescence.
(Return to Program Resources)