When alcohol group status significantly predicted performance within a neuropsychological domain, post-hoc multiple regression analyses were done on each of the measures within that domain to determine whether AD predicted performance even if lifetime tobacco use, lifetime other substance use, level of education, age at testing, and language of test administration were controlled. The same procedures were followed as with the domain scores. Z-scores were used for all neuropsychological measures, as Kolgorov-Smirnov tests suggested non-normal score distributions.
