A flexible method based on maximum likelihood theory is introduced for the analysis of binary response data in twins. The method allows for explanatory variables such as age and sex, is free of the untestable distributional assumption of bivariate normality of liability, and makes more efficient use of the data available. The method is illustrated with preliminary data on drinking status in adult twins. Although there is some bias in the ascertainment of male dizygous twins, the results suggest that monozygous twins are more concordant than dizygous twins for drinking status.
