We conducted within-family analyses among twin pairs discordant for alcohol diagnosis (DSM-III-R) at age 25, testing for P3 and P2 differences with paired-samples t-tests. Given substantial heritability for symptom counts and high concordance for diagnosis in this selected sample of twins, of whom half satisfied diagnostic criteria, our sample of discordant pairs was small (N = 20), and these analyses yielded no statistically significant results. Pairs were defined discordant when 1 co-twin met criteria for DSM-III-R diagnosis of alcohol dependence and the co-twin was symptom-free, that is, fulfilled none of the alcohol dependence criteria on SSAGA interview at age 25.
