To assess the degree of overlap in genetic and environmental influences between timing of first alcohol use and AD, a bivariate triangular decomposition (also known as a Cholesky decomposition) was fitted. The proportion of variance attributable to common vs. phenotype-specific sources of influence on initiation of alcohol use and AD were estimated under this model. Models were fitted in Mx using Full Information Maximum Likelihood estimation with raw categorical data. As in the univariate cases, thresholds were adjusted for age at the time of interview. The final model was derived by testing sub-models against the full model, as described above.
