Fig. 2.

Structural equation model (SEM) of the full information approach. Answers to three different covariates are displayed for the first ($x_{11},x_{12},x_{13}$) and second twin ($x_{21},x_{22},x_{23}$) of one family. $\mathit{\psi }$ is a latent variable that is estimated for every covariate (${\mathit{\psi }}_1,{\mathit{\psi }}_2,{\mathit{\psi }}_3$) and models covariance within families. The different latent variables for the first (${\mathit{\gamma }}_{11},{\mathit{\gamma }}_{12}$ and ${\mathit{\gamma }}_{13}$) and second (${\mathit{\gamma }}_{21},{\mathit{\gamma }}_{22}$ and ${\mathit{\gamma }}_{23}$) twin model within twin covariance