Although our sample size of twins is considerable, the models include relatively rare phenotypes (i.e., DA and AUD in females) and therefore have limited statistical power. As the objective is to quantify the magnitude of the variation in the population coming from various sources, we chose to present the full model with A, C and E components rather than try to obtain the best fitting submodel (i.e. an AE model only). This is in line with recommendations based on simulations which show that in such situations, parameter estimates from the full model are typically more accurate than those from submodels even if the latter provide a better model fit (Sullivan and Eaves 2002). We present the log likelihood and AIC values to facilitate comparisons between models. Models were fit in the OpenMx software (Boker et al. 2011).
