The complex nature of unique and shared effects in hierarchical linear regression: Implications for developmental psychology

Hierarchical linear regression and related techniques, such as commonality analysis, path analysis, and linear structural equation models with mediator variables, are often used to determine the extent to which the influence of an exogenous variable on a dependent variable, A, is "unique" to this exogenous variable, or "shared with" another predictor variable, B. The authors formally show that shared and unique effects are related to the partial correlation between A and B controlling for the exogenous variable. We discuss the implications of this property of hierarchical linear regression with a special consideration of the role of chronological age in developmental psychology and warn against the uncritical use of hierarchical linear regression procedures.