Shared and Nonshared Environment

In quantitative behavior genetic studies, researchers decompose variability in observed measures, typically referred to as phenotypes, into subcomponents. Why do people differ in how tall they are? Why are some people more outgoing than others? Classical behavior genetic studies attempt to answer these types of questions by partitioning variability into separate sources attributable to genetic, shared environmental, and nonshared environmental influences. This goal is accomplished by comparing the phenotypic similarity of pairs of individuals who have known genetic (e.g., identical and fraternal twins) or environmental (e.g., adoptive siblings) relatedness. The focus of this entry is on the two environmental components, which are much more complicated than they may first appear to be. The shared environment represents between-family effects that make siblings reared in the same home more similar to one another, regardless of their genetic relatedness. The nonshared environment represents within-family effects that make even genetically identical siblings reared in the same home different from each other. As will be seen in what follows, these technical definitions of static variance components hide considerable nuance that emerges once a life-span perspective is taken. This entry highlights several essential definitional features of variance components to keep in mind whenever interpreting estimates of the shared and nonshared environment.

To make the shared and nonshared environment more concrete, consider a hypothetical study design. Researchers randomly assign pairs of genetically unrelated children to parents at birth and allow the parents to rear the children. Presumably, the parents will differ on a wide range of variables, such as parenting style, race/ethnicity, socioeconomic status, religious beliefs, or any other variable that might differ across families. If these sorts of characteristics affect development, then children growing up in the same home should be more similar on some psychological characteristic, such as academic achievement. To test this, researchers examine the children’s high school grade point average and test whether one of the adoptive sibling’s grades can predict the other sibling’s grades. Why would one find such an association? It is because of the shared environment the siblings grew up in. Because the siblings are not genetically related in this example, any similarity must be due to the shared environment. Likewise, one could examine the influence of the nonshared environment by testing the phenotypic similarity of identical twins reared in the same home. Why might these twins receive similar grades at school? They share the same genes and the same home environment. If genes or the shared environment play a role, that would lead to similarity. Now, why might these twins receive dissimilar grades? The twins might have different life experiences, peer groups, or thoughts about the importance of school. The nonshared environment represents all of these types of influences on behavior. Importantly, the nonshared environment also includes idiosyncratic effects like measurement error, which would also lead to sibling dissimilarity.

Over the past century, thousands of quantitative behavior genetic studies have estimated shared and nonshared environmental influences on a wide variety of psychological, social, and medical traits by recruiting many millions of sibling pairs. The consensus from this research is that for most characteristics, genetic influences account for more variance than do shared environmental influences. For example, differences in cognitive [Page 1998]ability in adulthood are primarily due to genetic and nonshared environmental effects, with the shared environment accounting for less than 10% of the variance. At the same time, this line of research found substantial evidence for environmental effects but only of the nonshared variety. For most complex psychological phenotypes, the nonshared environment accounts for as much if not more variance compared to genetic effects.

...