and EXT-GR, separately) and age 14 frequency of intoxication, age 17 frequency of intoxication, and age 22 AD. The residual correlation within twin pairs was substantial and stronger in MZ twin pairs compared to DZ pairs; accordingly, regression models were run as hierarchical linear models using PROC MIXED and PROC GENMOD in SAS (SAS Institute, 2008), with twin pairs and individuals within twin pairs being treated as separate levels to account for the non-independence of twin-family data (Kendler et al. 2011). Because there were differences in mean levels of alcohol use, sex was used as a covariate in all analyses that collapsed data across males and females. To test for sibling influences on alcohol use behaviors, we examined the moderating effects of sex of the twin and sex of his/her co-twin. Accordingly, an ANOVA was carried out in SAS 9.3 (SAS Institute, 2008), with twin pairs and individuals within twin pairs being treated as separate levels. Further, differences between observed effects on alcohol use behavior outcomes across development, and by sex, were also examined using an ANOVA in SAS 9.3 (SAS Institute, 2008). We note that only those individuals who had initiated alcohol use were included in analyses of the
