Independent Pathway Model

Abstract

Abstract While traditional multivariate statistical methods, can describe correlational patterns of a set of variables (e.g., psychiatric symptoms or problem behaviors in children), they cannot provide insight into why certain symptoms or behaviors tend to co‐occur in a population. This can be achieved using methods of multivariate genetic analysis . In multivariate genetic models 2 main classes can be distinguished, which differ in the way the common factors are assumed to influence the different observed variables. The way unique factors influence the observed variables is the same in both classes. The first class of models is referred to as common pathway models or psychometric models. Common genetic and environmental factors influence all observed variables via a single psychometric factor, or underlying latent liability. The second class of models is referred to as independent pathway or biometric models. The common genetic and environmental factors influence the observed variables directly, without an intermediate higher order factor. As a result, the common genetic and environmental do not necessarily cause similar groupings. This means that different clusters of variables can be identified for shared genetic and environmental effects. For this reason biometric models can yield different results compared to higher order phenotypic factor models, which implicitly assume identical groupings for genetic and environmental effects.