Figure 1.

Figure 1a: Cholesky decomposition illustrating the association between Parental Monitoring (PM) and Drug Initiation (DI) in which the sources of covariance between PM and DI are decomposed into shared genetic (a_2,1) and environmental (c_2,1 & e_2,1) effects. This model also models the genetic (a_2,2) and environmental (c_2,2 & e_2,2) effects unique to DI (a_2,2, c_2,2, e_2,2).

A₁, C₁ & E₁ = latent additive genetic, shared, and non-shared environmental risks for PM, A₂, C₂ & E₂ = latent additive genetic, shared, and non-shared environmental risks for DI, PM & DI = common factors as indicated by observed phenotypic symptoms PM_1–3 and DI_1–3 respectively each with their own item specific latent genetic and environmental risk factors, PM₁ = Parent’s knowledge of child’s friends, PM₂ = Parent’s knowledge of how child spent their money, PM₃ = Parent’s knowledge of how child spent their free time, DI₁ = Cannabis initiation, DI₂ = Stimulant initiation, DI₃ = Cocaine initiation.

Figure 1b: Illustrative model of the direction of causation between the latent factors for Parental Monitoring (PM) and Drug Initiation (DI). This model approach predicts the relationship between PM and DI is explained by a reciprocal interaction, or causal parameters (β₂₁ & β₁₂), at the latent factor level.

A₁, C₁ & E₁ = latent additive genetic, shared, and non-shared environmental risks for PM, A₂, C₂ & E₂ = latent additive genetic, shared, and non-shared environmental risks for DI, PM & DI = common factors as indicated by observed phenotypic symptoms PM_1–3 and DI_1–3 respectively each with their own item specific latent genetic and environmental risk factors, β₂₁ = causal pathway from PM to DI, β₁₂ = causal pathway from DI to PM, PM₁ = Parent’s knowledge of child’s friends, PM₂ = Parent’s knowledge of how child spent their money, PM₃ = Parent’s knowledge of how child spent their free time, DI₁ = Cannabis initiation, DI₂ = Stimulant initiation, DI₃ = Cocaine initiation.