the default conceptualization for psychologists) or (2) as an environment moderating the association between a genotype and an outcome (i.e., the association between the short allele of 5HTTLPR and depression is strongest in individuals experiencing major stressful life events; often the default conceptualization for geneticists). Statistically, these are equivalent and indistinguishable, but they can lead to different interpretations of the same data and different expectations about the likelihood of detecting a gene-environment interaction effect. Conceptualizing GxE as a genetic effect (on an environment-behavior association) may lead one to assume small effect sizes based on the growing GWAS literature demonstrating that genetic effects on complex outcomes generally have very small effect sizes. Conceptualizing GxE as an environmental effect on a gene-behavior association may lead one to assume larger effects sizes. Nevertheless, under either interpretation, the effect size of the candidate gene is critical. Recognizing the modest main effect sizes produced by individual candidate gene polymorphisms, it may be overly optimistic to presume that the effect sizes associated with cGxE will be systematically larger. This is clearly a matter of some debate, as if certain types of cross-over interactions were prevalent, it would be possible. Nevertheless, the lessons we have learned from
