The effects of age, sex, and genotype on self-report drunkenness following a challenge dose of alcohol.

Age is a potential source of variation that contributes to differences between, but not within, twin pairs. In most genetic analyses of twin data, linear and other functions of age are usually removed prior to model fitting. This correction is typically applied only within twin groups of the same sex and zygosity, and no heterogeneity test of age regressions is performed. Here we include age as a variable in the model-fitting procedure and allow for tests of heterogeneity of age regressions across sex and zygosity groups. The LISREL formulation of the approach is illustrated with data collected from Australian twins on subjective impressions of drunkenness following alcohol consumption. The results indicate significant negative covariation of impressions of drunkenness with age. The data support a simple model of additive genetic and unique environmental variation. No evidence was found for sex differences in genetic or environmental components of variation.